机器翻译,已尽力保留原意与数字
内容摘要
马斯克第5次参加Lex Fridman的节目,与Neuralink团队及首位人类植入接受者Noland Arbaugh一同探讨脑机接口和人类的未来。
Musk's fifth Lex Fridman appearance, joined by the Neuralink team and first human implant recipient Noland Arbaugh, on brain-computer interfaces and the future of humanity.
中文实录Transcript
2049 个段落
第 1 段
接下来是与埃隆·马斯克、DJ Seo、Matthew MacDougall、Bliss Chapman和Nolan Arbaugh的一场对话,内容涉及Neuralink和人类的未来。埃隆、DJ、Matthew和Bliss当然都是出色的Neuralink团队成员,而Noland则是首位在大脑中植入Neuralink设备的人类。我分别与他们每个人进行了交谈,所以你可以通过时间戳跳转观看,或者像我建议的那样,来点硬核的,把整期都听完。
第 2 段
这是我做过最长的一期播客。这是一场引人入胜、技术性极强且话题广泛的对话,我喜欢其中的每一分钟。现在,亲爱的朋友们,有请埃隆·马斯克,这是他第5次做客《Lex Fridman播客》。
第 3 段
喝咖啡还是水?
第 4 段
水。我现在咖啡因摄入过量了。你想来点咖啡因吗?
第 5 段
我是说,当然。
第 6 段
有一款叫 Nitro 的饮料。
第 7 段
这东西据说基本上能让你一直清醒到明天下午。(笑)——对。我没有任何——所以 Nitro 是什么?就是含有大量咖啡因之类的吗?
第 8 段
别问问题。它就叫 Nitro。
第 9 段
你还需要知道别的吗?
第 10 段
它含有氮气,这太荒唐了。我是说,反正我们呼吸的空气中有 78% 是氮气。你为什么还需要添加更多?(笑)——[说话人] 不幸的是,你会需要它。
第 11 段
大多数人以为自己呼吸的是氧气,而他们实际上呼吸的是 78% 的氮气。你需要一个类似牛奶吧的地方。
第 12 段
牛奶吧。(埃隆笑)——就像《发条橙》里那种。(笑)——对,对。它在你心目中的库布里克电影里能排进前 3 吗?
第 13 段
《发条橙》,相当不错。我是说,它很疯狂。我会说,令人不适。
第 14 段
(笑)好。好,那么首先让我们退一步,热烈祝贺你们成功将 Neuralink 植入人体。这对 Neuralink 来说是历史性的一步。
第 15 段
哦,谢谢,对。
第 16 段
接下来还会有更多。
第 17 段
对,而且很显然,我们刚刚也完成了第 2 次植入。
第 18 段
[莱克斯] 进展如何?
第 19 段
到目前为止一切顺利。看起来,我想我们有超过 400 个电极正在提供信号。所以,对。
第 20 段
很好。你认为人类参与者的数量会多快扩大?
第 21 段
这取决于监管批准,取决于我们获得监管批准的速度。所以我们希望在今年年底前做到 10 例。总共 10 例,也就是再做 8 例。
第 22 段
而且每做 1 例,你们都会从新的生物学情况、大脑、一切,以及 Neuralink 的整个链条、解码、信号处理和诸如此类的所有方面吸取大量经验。
第 23 段
对,对,我认为很显然,每做 1 例都会有所改善。我是说,我不想把它说坏,但第 2 次植入似乎进行得非常顺利,所以有大量信号、大量电极。它运行得非常好。
第 24 段
你认为在接下来的,假设一下,让我们大胆一点,接下来的几年里,Neuralink 会有哪些改进?
第 25 段
我是说,几年后,它的规模会极其庞大,因为我们会大幅增加电极数量。我们会改进信号处理。即使在我们的第 1 位患者诺兰身上只有大约,我不知道,10%、15% 的电极在工作,我们也能够达到每秒 1 比特。那是世界纪录的 2 倍。所以我认为,未来几年我们将开始以若干个数量级大幅超越世界纪录。
第 26 段
所以它会开始达到,我不知道,每秒 100 比特 1000。也许比如 5 年后,我们可能会达到 1 兆比特,比任何人通过打字或说话所能达到的沟通速度都快。
第 27 段
对,BPS 是一个很有意思的衡量指标。一旦达到某个 BPS 水平,体验可能会出现巨大飞跃。
第 28 段
[埃隆] 对。
第 29 段
比如与计算机交互的全新方式可能会被解锁。
第 30 段
还有与人类交互。
第 31 段
与其他人类交互。
第 32 段
前提是他们也有(笑),他们也想要一个 Neuralink。
第 33 段
对。
第 34 段
否则,他们将无法足够快地吸收这些信号。
第 35 段
你认为这会提高思想讨论的质量吗?
第 36 段
嗯,我想你可以这样考虑:如果把交流速度放慢,你会有什么感觉?如果你只能以,比如说,正常速度的 1/10 说话,你会觉得:“哇,这慢得令人痛苦。”
第 37 段
[莱克斯] 对。
第 38 段
那么现在,想象一下你能以正常速度的 10 倍、100 倍或 1,000 倍清晰地说话、交流。
第 39 段
听着,我很确定头脑正常的人都不会以 1 倍速听我说话,他们会用 2 倍速听。(埃隆笑)我只能想象 10 倍速听起来会是什么感觉,或者是否真的能听懂。
第 40 段
我通常默认用 1.5 倍速。你可以用 2 倍速,不过,实际上,如果我是为了入睡而每次听某个人大约 15、20 分钟,那我会用 1.5 倍速。如果我在专心听,我会用 2 倍速。(笑)——对。
第 41 段
但实际上,如果你开始真正听播客或者某种有声书之类的任何东西,如果你习惯了用 1.5 倍速,那么 1 倍速听起来就慢得令人痛苦。
第 42 段
我仍然坚持用 1 倍速,因为我害怕。我害怕自己会对现实感到厌烦,对这个所有人都以 1 倍速说话的真实世界感到厌烦。(两人笑)——嗯,要看是什么人。你可以说得很快。比如我们可以非常迅速地交流。而且,如果你使用的范围很广,如果你的词汇量更大,你的比特率,也就是有效比特率,就会更高。
第 43 段
这个说法很好。
第 44 段
对。
第 45 段
有效比特率。我的意思是,问题在于,语言的低比特传输中实际压缩了多少信息。
第 46 段
是的。如果有一个词能够传达通常需要,我不知道,10个简单词才能传达的东西,那么你手上可能就有了10倍的压缩。其实模因也是如此,模因就像数据压缩。它传达的是一个整体,你会同时接收到大量可以解读的符号。而且你有点能比面对文字或一幅简单图片时更快地理解它。
第 47 段
当然,你这里说的是广义上的模因,比如思想。
第 48 段
是的。有一整套思想结构,就像一个思想模板,然后你可以往那个思想模板里添加一些东西。但某个人脑中已经有了那个预先存在的思想模板。所以当你添加那一点增量信息时,你所传达的远不只是几个……只是几个词。它包含了与那个模因相关的一切。
第 49 段
你认为,随着电极数量的扩大,能力会出现涌现式飞跃吗?比如会存在某个特定的,你认为会存在一个实际的数值,一旦达到,人类体验就会改变?
第 50 段
是的。
第 51 段
你觉得那个数值可能是多少,无论是电极数量还是BPS?我们当然无法确定,但这是10,000、100,000吗?
第 52 段
是的,我的意思是,当然,如果你达到了每秒10,000比特,我的意思是,那就远远快于目前任何人类的沟通速度。如果你想想人类平均每秒有多少比特?以一天为周期计算,还不到每秒1比特,因为一天有86,400秒。而你一天不会交流86,400个词元。因此,按24小时取平均值,你的每秒比特数不到1。
第 53 段
这相当慢。而现在,即使你交流得非常快,而且你是在和一个理解你所说内容的人交谈,因为为了进行交流,你至少必须在某种程度上对你所交谈对象的心理状态进行建模。
第 54 段
然后取出你试图传达的概念,把它压缩成少量音节,说出来,并希望对方把它们解压成一种概念结构,使其尽可能接近你脑中的东西。
第 55 段
是的,我的意思是,在那个过程中会有大量信号损失。
第 56 段
是的,压缩和解压都非常糟糕。而你的神经元所做的很多事情,就是把概念提炼成少量符号,比如我正在说的音节,或者按键输入,无论具体是哪种情况。所以这就是你的大脑计算所做的许多事情。
第 57 段
现在,有一种观点认为,这样做其实是一件有益健康的事,或者说是一件有帮助的事,因为当你试图压缩复杂概念时,你或许会被迫提炼出这些概念中最本质的东西,而不是保留所有那些无关紧要的内容。所以在压缩过程中,你会把事物提炼到最重要的部分,因为你只能说几件事。所以这或许是有帮助的。
第 58 段
我觉得我们可能会,如果我们的数据传输速率提高,我们很可能会变得啰嗦得多。就像你的电脑一样,以前电脑只有,比如说,我的第一台电脑有 8K 内存,所以你真的会仔细考虑每一个字节。而现在的电脑有许多个 GB 的内存。所以如果你想做一个只显示“Hello world”的 iPhone 应用,它大概,我不知道,至少也有几 MB。(笑)一大堆无关紧要的东西。
第 59 段
但尽管如此,我们还是更喜欢拥有内存和算力更大的电脑。所以 Neuralink 的长期愿景是通过提高通信带宽来改善 AI 与人类的共生关系,因为即使在最温和的 AI 情景下,你也必须考虑到,AI 光是等着你吐出几个词就会感到无聊。
第 60 段
我的意思是,如果 AI 能以每秒太比特的速度交流,而你以每秒比特的速度交流,那就像 203。
第 61 段
嗯,对于一个超级智能物种来说,这是一个非常有意思的问题。人类有什么用?
第 62 段
我认为,把人类视为意志的来源,是有一定道理的。
第 63 段
意志?
第 64 段
意志,对。意志或目的的来源。所以,如果你认为人类心智本质上是这样的:其中有原始的边缘系统要素,基本上连爬行动物都有;还有大脑皮层,那是大脑中负责思考和规划的部分。大脑皮层比边缘系统聪明得多,但它在很大程度上仍然是为边缘系统服务的。它试图让边缘系统开心。
第 65 段
我的意思是,人们为了设法上床而投入的算力之多,简直疯狂,而且实际上并不是为了繁衍。他们实际上只是想做这种简单的动作。(笑)而且他们能从中获得快感。所以,这种简单的、抽象来看颇为荒谬的动作,也就是性,大脑皮层正在投入海量算力,试图弄清楚如何做到这件事。
第 66 段
所以,人类这个物种的分布式算力中,大概有 90% 都花在设法上床这件事上,投入量非常巨大。
第 67 段
很大比例,对,对。大多数性行为除了享乐之外没有任何目的。就是某种快乐之类的。释放多巴胺。偶尔,它是为了繁衍,但对人类、现代人类来说,主要是娱乐。所以,你的大脑皮层比边缘系统聪明得多,却在试图让边缘系统开心,因为边缘系统想要发生性行为,或者想吃些美味的食物,诸如此类。
第 68 段
然后,第三层系统又进一步增强了这一点,也就是你的手机、笔记本电脑、iPad 或者其他计算设备。那就是你的第三层。所以实际上,你已经是一个赛博格了。你拥有这个第三层计算层,它以你的电脑及其所有应用程序,或者你的计算设备的形式存在。
第 69 段
所以在设法上床这件事上,实际上也有海量数字算力在设法促成上床,比如 Tinder 之类的。
第 70 段
对。所以我们人类打造的算力也参与其中。(笑)——对,我的意思是,有好几吉瓦的数字算力被投入到设法上床这件事上。
第 71 段
对。(笑)如果 AGI 会————就在我们说话的此刻,这件事正在发生。
第 72 段
如果我们与 AI 融合,它只会扩大我们人类用于————差不多。
第 73 段
设法上床的算力。
第 74 段
嗯,那当然是用途之一,对。
第 75 段
对。
第 76 段
但我想说的是,是的,人类有用吗?嗯,这就涉及这个根本问题:生命的意义是什么?为什么要做任何事?所以,如果我们简单的边缘系统提供了做某件事的意志来源,然后这种意志传到我们的大脑皮层,再传到我们的第三层计算层,那么我不知道,或许实际上,在一种温和的情景下,AI 只是在试图让人类的边缘系统开心。
第 77 段
对,看起来意志并不仅仅与边缘系统有关。那里还有很多有意思且复杂的东西。我们也想要权力。
第 78 段
我认为那也是边缘系统的欲望。
第 79 段
但我们也希望以某种合作的方式减轻世间的苦难。
第 80 段
并非每个人都这么想,不过,对,当然。有些人确实这么想。
第 81 段
作为一群人类,当我们聚集在一起时,就会开始产生这种集体智慧,其意志比底层那些猿类的个体后代更加复杂,对吧?所以还存在其他动机。而这可以成为 AGI 目标函数的一个非常有意思的来源。
第 82 段
对,我的意思是,确实有这些相当理性或者说更高层次的目标。对我来说,比如生命的意义是什么,或者理解宇宙的本质,这些都让我非常感兴趣。希望 AI 也对此感兴趣。而 xAI 和 Grok 的使命就是理解宇宙。
第 83 段
所以你认为,当 Neuralink 拥有 10,000、100,000 个通道时,大多数使用场景都会是与 AI 系统交流吗?
第 84 段
嗯,假设不存在,我的意思是,它们正在解决人们面临的基本神经问题,比如他们的脊髓或颈部神经元受损,或者,你知道,就像前 2 位患者的情况一样,那么显然,首要任务就是解决脊髓、颈部或大脑本身的根本性神经元损伤。
第 85 段
第 2 款产品叫作 Blindsight,目的是让完全失明、失去双眼或视神经,或者根本看不见的人,能够通过直接触发视觉皮层中的神经元而看见。所以我们才刚从基础开始,这些相对来说是非常、简单的事情,也就是解决神经元损伤。我认为它或许还能解决精神分裂症。
第 86 段
如果人们有某种癫痫发作,它或许可以解决这个问题。它可以帮助改善记忆。可以说,这里存在某种科技树,比如你先掌握基础。就像你得先识字,才能读《指环王》。(两人笑)——明白了。
第 87 段
你有字母和字母表吗?好,很好。单词呢?然后最终变得湿软。所以我认为,未来可能有些事情需要担心。但最初几年其实只是解决基本的神经损伤。比如那些基本上完全或近乎完全失去从大脑到身体的。斯蒂芬·霍金就是一个例子。
第 88 段
Neuralink 将产生极其深远的影响,因为我的意思是,你可以想象,如果斯蒂芬·霍金能够像我们这样快速地交流,或许甚至更快。而这当然是可能的。事实上很有可能,我会说很可能。
第 89 段
所以存在一种医疗和非医疗的双轨路径,也就是说,你谈到的一切将来都可以应用于非残障人士吗?
第 90 段
合乎逻辑的做法,明智的做法,是从解决基本的神经元损伤问题开始。
第 91 段
[莱克斯] 是的。
第 92 段
因为新设备显然存在一定风险。你不可能把风险降到零。这是不可能的。所以,既然存在某种无法消除的风险,你就会希望获得尽可能高的回报。而如果某人的沟通能力能够得到巨大改善,那就值得冒这个风险。
第 93 段
随着你把风险降下来。
第 94 段
是的,随着你把风险降下来。一旦风险降到,你知道,如果有数千人已经使用它多年,而且风险极低,那么或许到那个时候,你可以考虑说:“好吧,让我们以增强为目标。”现在,我认为我们其实会以增强神经元受损者的能力为目标。所以我们的目标不只是让人们获得与正常人类相当的通信数据速率。
第 95 段
我们的目标是让四肢瘫痪或大脑与身体之间的连接可能完全丧失的人,获得超过正常人类的通信数据速率,就像在说:“好吧,我们都已经做到这一步了。为什么不呢?让我们赋予人们超能力。”
第 96 段
视觉也是如此。在恢复视觉的过程中,这种恢复是否可能有某些超越人类的方面?
第 97 段
是的,一开始,恢复的视觉将是低分辨率的,因为你必须问,比如:“你能在里面放入并触发多少个神经元?
第 98 段
而且你可以做一些事情,比如调整电场,即使你有,比如说,10,000个神经元,也不只是10,000个像素,因为你可以调整神经元之间的感觉,并按一定模式操控它们,从而获得,所以,比如说,让10,000个电极实际上为你提供,我不知道,也许类似于100万像素或1,000万像素的效果。然后随着时间推移,我认为你会达到比人眼更高的分辨率。
第 99 段
而且你还可以看到不同波长的东西。就像《星际迷航》里的乔迪·拉弗吉。就像那东西。你想用雷达看东西?没问题。你可以看到紫外线、红外线,拥有鹰眼视觉,随你想要什么。
第 100 段
你觉得会不会有,让我问一个乔·罗根式的问题。你觉得会不会有,(笑)我最近刚喝了死藤水。
第 101 段
这是罗根式的问题吗?
第 102 段
不是。好吧,是。
第 103 段
好吧,我想严格来说,是的。
第 104 段
对。
第 105 段
试过GMT吗,兄弟?(两人笑)- 我爱你,乔。
第 106 段
好吧。(笑声继续)- 但等等,等等,对。你有谈过很多这件事吗?死藤水?
第 107 段
没有,我没有。我没有。
第 108 段
好吧,那你为什么要泄密?(莱克斯笑)那真的是一件不可思议的事—— - 反过来问你。(两人笑)- 哇,好吧。
第 109 段
你在丛林里。
第 110 段
[莱克斯] 对,在树林之中,我和—— - 对,肯定很疯狂。
第 111 段
还有萨满。对,对,对,虫子在身边,动物也全都在你周围,目之所及都是丛林。没有—— - 我是说—— - 就该这么做。
第 112 段
东西看起来会相当疯狂。
第 113 段
对,相当疯狂。(埃隆笑)- 我想是极高剂量。
第 114 段
就是别跑去拥抱一条森蚺什么的。(笑)- 除非你和森蚺做过爱,否则你不算活过。对不起,但是—— - 蛇梯棋。(两人笑)- 对,我喝了剂量极高的—— - [埃隆] 好吧。(笑)- 9杯,而且—— - 该死。好吧,听起来很多。当然,诺兰的是1杯,还是—— - 1杯或2杯。通常是1杯。
第 115 段
你直接就,等等。是一开始就这样,还是逐渐加到这个量的?
第 116 段
所以我——(两人笑)- 你这是直接跳进深水区。
第 117 段
分了2天,因为第一天,我喝了2杯,然后我—— - 好吧。
第 118 段
那是一趟旅程,但不太像是—— - 不像什么启示。
第 119 段
不是那种进入深空的旅程。就像是坐了一小段飞机。
第 120 段
[埃隆](笑)好吧。
第 121 段
看到了一些树、一些视觉景象之类的。我就是看到了一条龙,所有那类东西。但是——(笑)- 那可是9杯。我想你去了冥王星。
第 122 段
[莱克斯] 冥王星,对。不,是深空。
第 123 段
深空。
第 124 段
不,我这次体验有一个有意思的方面,就是我原以为自己会遇到一些恶魔,有些需要去处理的东西。
第 125 段
人们都—— - 所有人都这么说。对,没错。
第 126 段
我什么都没有。我经历的全是正面的。我只是—— - 哦,就是纯净的灵魂。
第 127 段
我不这么认为,我不知道。(笑)但我一直在想,我对生命中认识的那些人的想法仿佛有着极高的分辨率。你也在其中。
第 128 段
好吧。
第 129 段
而且并不是从我与那个人的关系出发,而只是看他们本人,我对他们是怎样的人怀有这种深深的感激。
第 130 段
这很酷。
第 131 段
就像是这种探索,像《模拟人生》之类的,你可以观察他们。
第 132 段
当然。
第 133 段
我可以观察人们,只是惊叹于他们有多么了不起。
第 134 段
听起来太棒了。
第 135 段
对,那很棒。我一直在等着—— - 史蒂文什么时候来?(两人笑)- 没错。也许我会有一些负面想法。什么都没有,什么都没有。只是对他们怀有极度的感激。然后还有,大量的太空旅行。(两人笑)- 太空旅行去了哪里?
第 136 段
事情是这样的。那些人,我认识的人类,他们身上有一种,最恰当的描述就是他们身上有光芒。然后我不断从他们那里向外飞,去看地球,看我们的太阳系,看我们的银河系。我看到那种光,那种光芒遍布整个宇宙。就像无论那是什么形态。无论那种—— - [埃隆] 你飞过银河系了吗?
第 137 段
对,对。
第 138 段
好吧。你这都跨星系了。
第 139 段
对,跨星系。
第 140 段
好吧,真厉害。
第 141 段
但总是指向里面—— - 好吧。
第 142 段
对,越过银河系。我是说,我看到了数量极其庞大的星系,跨越星系,而且所有星系都在发光。但我无法控制那种寒意,因为我其实会探索太阳系附近的距离,看看有没有外星人或任何那类东西。我不知道—— - 有外星人吗?0个外星人?
第 143 段
外星人的意味,因为它们在发光。它们发光的方式和人类发光的方式一样。就像我看到的那种生命力,那种让人类变得奇妙的东西,遍布整个宇宙。就像那里有这些发光的点。所以我不知道。它让我觉得那里存在生命。不,不是生命,而是某种东西,无论那种让人类变得奇妙的东西是什么,它遍布整个宇宙。
第 144 段
听起来不错。
第 145 段
是的,太神奇了。没有恶魔,没有恶魔。我找过恶魔。没有恶魔。有龙,而且它们非常棒。所以问题是————有任何吓人的东西吗?
第 146 段
龙?但它们并不可怕。它们是朋友,它们会保护我。所以问题是————《神龙帕夫》。
第 147 段
不,更像是《权力的游戏》里的那种龙。它们不太友善。它们非常巨大。所以问题是,嗯,夜晚的参天大树,也就是我所在的地方。
第 148 段
是啊。我的意思是,丛林有点吓人。
第 149 段
是的,那些树开始看起来像龙,而且它们全都像是在看着我。
第 150 段
当然,好吧。
第 151 段
而且感觉并不可怕。它们看起来像是在保护我。顺便说一下,那位萨满和那些人一句英语都不会,这让它变得更吓人了,我猜。(笑)我们甚至都谈不上相似,你知道,我们在很多方面都身处完全不同的世界。但对,他们会谈到森林之母保护着你,而那就是我的感受。
第 152 段
而且你是在丛林深处?
第 153 段
非常深。这可不是什么旅游度假地。
第 154 段
比如在里约之类的地方外面10英里?
第 155 段
不,我们去了——(两人都笑)不,这可不是—— ——亚马逊深处。
第 156 段
我和一个名叫保罗·罗索利的人一起,他基本上就是泰山。他住在丛林里。我们深入其中,然后彻底疯玩了一把。
第 157 段
哇,真酷。
第 158 段
是的。所以不管怎样,我能通过Neuralink获得同样的体验吗?
第 159 段
大概可以,是的。
第 160 段
我猜这就是身体无障碍的人要问的问题。你认为在我们的感知中,在我们对世界的体验中,有很多东西可以利用Neuralink去探索、去把玩吗?
第 161 段
是的,我的意思是,Neuralink,它其实是一种通用的输入输出设备。它读取电信号并产生电信号。而且我的意思是,你一生中体验过的一切,气味、情绪,所有这些都是电信号。所以想到你的整个人生体验都被提炼成了供神经元使用的电信号,感觉有点奇怪。但事实确实如此。
第 162 段
或者,我的意思是,至少所有证据都指向这一点。所以我的意思是,如果你触发正确的神经元,就可以触发某种特定的气味。你当然可以让东西发光。我的意思是,几乎什么都能做。我的意思是,真的,你可以把大脑看作一台生物计算机。
第 163 段
所以,如果这台生物计算机的某些,比如说,芯片或元件坏了,假设你的某种能力,如果你中风了,如果你曾经中风,那意味着你的大脑有某个部分受损了。如果那部分,比如说,是语言生成能力或移动左手的能力。那就是Neuralink可以解决的那类问题。
第 164 段
如果你有比如大量的记忆丧失,那些记忆就是没了,那么我们无法找回那些记忆。我们可以恢复你形成记忆的能力,但无法恢复已经完全消失的记忆。现在,我应该说,也许如果我的记忆有一部分还在,而坏掉的是访问记忆的手段,那么我们可以重新启用访问记忆的能力。但你可以把它想成计算机里的RAM。
第 165 段
如果RAM被毁了,或者你的SD卡被毁了,我们无法把它找回来。但如果被毁的是与SD卡的连接,我们可以修复它。如果它在物理上是可以修复的,那么,是的,它就可以被修复。
第 166 段
当然,有了 AI,你就可以,比如说,你可以修复照片,填补照片中缺失的部分。也许你也能做同样的事,就像—— ——是的,你可以这样说:“根据你掌握的关于那个人的所有信息,创建最有可能的一组记忆。”然后你就可以,那会是对记忆的概率性还原。现在,我们说得相当玄奥了。
第 167 段
但人类体验中最美好的方面之一,就是回忆那些美好的记忆。就像丹尼·卡尼曼谈到的那样,我们一生中的大部分时间都活在记忆里,而不是实际的当下。我们在收集记忆,也会在脑海中以某种方式重温它们。那就是美好时光。如果你只是对我们整个人生进行积分,那么正是对美好时光的回忆产生了最大量的幸福。
第 168 段
所以—— ——是的,嗯,我是说,除了我们的记忆,我们还是什么?而死亡除了记忆的丧失、信息的丧失之外,又是什么?你可以这样说,嗯,如果你可以,你做一个思想实验,如果你被无痛地分解,然后片刻之后又重新整合,就像传送一样,我想,只要没有信息损失,那么你的那具身体曾被分解这一事实就无关紧要。
第 169 段
而记忆正是其中如此重要的一部分。
第 170 段
从根本上说,死亡就是信息的丧失、记忆的丧失。
第 171 段
所以,如果我们能尽可能准确地存储它们,我们基本上就实现了某种永生。
第 172 段
是的。
第 173 段
你谈过 AI 的威胁和安全隐患。让我们看看长期愿景。在你看来,Neuralink 是我们目前在 AI 安全方面最好的方案吗?
第 174 段
这是一个可能有助于 AI 安全的想法。当然不是,我不想声称它是什么万灵药,或是板上钉钉的事。但我是说,很多年前,我当时在想:“嗯,是什么会阻碍人类集体意志与人工智能保持一致,而人类的低数据速率,尤其是我们缓慢的输出速率,必然会仅仅因为通信如此缓慢,而削弱人类与计算机之间的联系吗?
第 175 段
就像你越是一棵树,就越不知道树是什么。比如说,你看着一棵树,看着这株植物或诸如此类的东西,然后想:“嘿,我真的很想让那株植物开心。”但它没说多少话,你知道吗?
第 176 段
所以,我们越是提高人类能够输入和输出的数据速率,就意味着在一个充满 AGI 的世界里,我们拥有的机会越大?
第 177 段
是的。如果尤其是输出速率得到大幅提高,我们就能让人类集体意志与 AI 更好地保持一致。而且我认为,输出速率有可能提高,我不知道,3个数量级,也许6个,也许更多。所以这比当前的情况要好。
第 178 段
而这种输出速率会通过增加电极数量、通道数量,也许还会通过植入多个 Neuralink 来实现?
第 179 段
是的。
第 180 段
你认为未来20年左右,会出现数亿人都装有 Neuralink 的世界吗?
第 181 段
是的,我认为会。
第 182 段
你认为,当人们只是,当他们看到那些能力,看到可能实现的超人能力,然后安全性也得到证明之后?
第 183 段
是的,如果它极其安全,而且你能拥有超人能力,再比如说,你可以上传自己的记忆,这样你就不会失去记忆,那么我认为可能会有很多人选择使用它。例如,它会取代手机。我是说,比如说,一部手机最大的问题就是试图弄清楚你想要什么。所以才会有自动补全,也会有输出,也就是屏幕上的所有像素。
第 184 段
但从人类的角度来看,输出实在是太他妈慢了。台式电脑或手机正拼命试图理解你想要什么,而从计算机的角度来看,每次按键之间都像隔了永恒。
第 185 段
是吗?计算机是在跟一棵树说话,那棵行动缓慢、正试图滑动屏幕的树。
第 186 段
是的。所以,如果你的计算机每秒执行数万亿条指令,而整整1秒过去了,我是说,它本可以做1万亿件事。
第 187 段
是的,我认为这对人们来说既令人兴奋又令人害怕,因为一旦你拥有非常高的比特率,就会以一种很难想象的方式改变人类体验。
第 188 段
是的。那会是某种不同的东西。我是说,某种未来主义的人行道。我是说,顺便说一句,我们显然谈的不是近在眼前的事。你问我遥远的未来会是什么样子。也许这就像,它并非极其遥远,但要10、15年,诸如此类。(Lex 叹气)——我什么时候能装一个?10年?
第 189 段
可能不到10年。取决于你想做什么。
第 190 段
嘿,如果我能达到,比如说,1000 BPS—— ——什么时候能达到1000 bps?
第 191 段
而且它是安全的,我可以一边躺着吃奇多,一边直接与计算机交互,我不吃奇多。人机交互的某些方面,如果做得更高效、更愉快,就像,值得。
第 192 段
嗯,我们相当有信心,我认为也许在未来1年或2年内,植入 Neuralink 的人就能胜过职业游戏玩家。
第 193 段
不错。
第 194 段
因为反应时间会更快。
第 195 段
我去参观了孟菲斯。
第 196 段
是的,是的。
第 197 段
你在算力上投入很大。
第 198 段
是的。
第 199 段
你还说过,要玩就要赢,否则就根本别玩,那么要怎样才能赢?
第 200 段
对于 AI 来说,这意味着你必须拥有最强大的训练算力,而且训练算力的提升速度必须比其他所有人都快,否则你就赢不了。你的 AI 会更差。
第 201 段
那么 Grok,比如说,Grok 3 可能会在,什么时候,比如明年推出?
第 202 段
嗯,希望是今年年底。
第 203 段
Grok 3?
第 204 段
如果我们运气好的话,是的。
第 205 段
它怎样才能成为世界上最好的 LLM、最好的可用 AI 系统?其中有多少取决于算力?有多少取决于数据?有多少取决于后训练之类的东西?有多少取决于你把它包装成的产品?诸如此类的所有因素。
第 206 段
我的意思是,它们不会有影响。这有点像是说,假设这是一场一级方程式赛车比赛。什么更重要,赛车还是车手?我的意思是,两者都重要。如果赛车不够快,比如说,它的马力只有竞争对手的一半,那么最好的车手还是会输。如果它的马力是对手的2倍,那么可能就连水平一般的车手也还是会赢。
第 207 段
所以训练算力有点像发动机,这台发动机的马力是多少。所以说真的,你要努力在这方面做到最好。然后,你使用这些训练算力的效率有多高?进行推理,也就是使用 AI 时的效率又有多高?所以显然,这归根结底取决于人才。然后你拥有怎样的独有数据获取渠道?这也会发挥作用。
第 208 段
你认为 Twitter 数据会有用吗?
第 209 段
是的,我的意思是,我认为,我认为大多数领先的 AI 公司已经抓取了所有 Twitter 数据。不是我认为,而是它们确实抓取了。所以往后看,有用之处在于数据可以精确到当下这一秒。它们很难实时抓取这些数据。所以 Grok 已经具备时效性优势。
第 210 段
我认为,随着来自数百万辆 Tesla 汽车、最终数千万辆汽车的实时视频,再加上 Optimus,未来可能会有数亿台、也许数十亿台 Optimus 机器人从现实世界中学习海量信息。我认为,最终最大的数据来源算是 Optimus。Optimus 将成为最大的数据来源。
第 211 段
因为——因为现实可以扩展。现实可以扩展到现实本身的规模。看到人类实际上只能积累这么少的数据,真的让人感到谦卑。说真的,你看,人类生成了多少万亿个可用词元,而在非重复的情况下,比如排除垃圾信息和重复内容后,数量并没有多大。很快就会用完。
第 212 段
而 Optimus 可以去,所以 Tesla 汽车很遗憾不得不留在道路上。Optimus 机器人可以去任何地方,而且道路之外有更多现实,可以去越野。
第 213 段
我的意思是,除了商店,在那里我可以,比如拿起杯子,然后看看,它拿杯子的方式对不对?它有没有往杯子里倒水?水是倒进杯子里了,还是没倒进杯子里?它有没有把水洒出来?
第 214 段
[Lex] 是的。
第 215 段
像这样的简单事情。但它可以把这种事情以10亿倍的规模来做,从而从现实中生成有用的数据。也就是因果关系之类的东西。
第 216 段
你认为要实现这类人形机器人的大规模生产,需要什么条件?
第 217 段
其实,这和汽车是一样的。我的意思是,全球的汽车产能大约是每年1亿辆。而且还可以更高。只不过需求量大约是每年1亿辆。然后,大约有20亿辆汽车正以某种方式被使用,这很合理。比如一辆车的寿命大约是20年,所以这是稳态。大致来说,在拥有20亿辆汽车的情况下,每年可以生产1亿辆。现在对于人形机器人来说,其效用要大得多。所以我猜,人形机器人的年产量更可能是超过10亿台。
第 218 段
但在你出现并开始制造 Optimus 之前,人们认为这是一个极其困难的问题。我的意思是,它仍然————嗯,确实是。
第 219 段
极其困难。
第 220 段
所以就像在公园里散步一样。我的意思是,目前 Optimus 在公园里行走都会有困难。我的意思是,它可以在公园里走。公园的地形并不太难,但它将能够在各种各样的地形上行走。
第 221 段
还能拿起物体。
第 222 段
对,对。它已经能做到了。
第 223 段
[Lex] 但各种各样的物体都可以吗?
第 224 段
对,对。
第 225 段
所有外部物体。我的意思是,往杯子里倒水并不会让你兴奋,因为如果你对容器一无所知,那可能是各种各样的容器。
第 226 段
对,仅仅是手部就会投入海量的工程工作。手部可能会,可能占 Optimus 全部工程工作的近一半。从机电角度来看,手部可能大致占工程工作的一半。
第 227 段
但人类如此多的智能,如此多的智能,都投入到了我们用双手所做的事情中。
第 228 段
对。
第 229 段
这就是对世界的操控,对世界中物体的操控。智能就是对世界中物体的安全操控,对。
第 230 段
对。我的意思是,你开始真正思考自己的手以及它是如何运作的。
第 231 段
我一直都在想。
第 232 段
在感觉控制小人图中,你的手大得惊人。
第 233 段
[Lex] 对。
第 234 段
所以我的意思是,比如你的手,那些执行器,也就是你手部的肌肉,几乎绝大部分都在你的前臂里。所以你的前臂拥有实际控制手部的肌肉。手本身有几块小肌肉,但你的手其实就像一个骨架肉偶。还有缆索。所以控制手指的肌肉位于前臂,它们穿过腕管,也就是你有一小簇骨头和一个狭窄的隧道,这些缆索,也就是肌腱,会从中穿过。而你的手主要就是由这些肌腱带动的。
第 235 段
而为了完成所有那类事情,必须把类似这些肌腱的东西重新设计进 Optimus。
第 236 段
对,所以像目前的 Optimus,我们尝试过把执行器放在手本身里面,但那样最后就会出现这种————巨大的手?
第 237 段
对,看起来很奇怪的巨大手。而且它们实际上没有足够的自由度和/或足够的力量。于是你会意识到:“哦,好吧,这就是为什么必须把执行器放在前臂里。”而且就像人类一样,你必须让缆索穿过一个狭窄的隧道来操纵手指。此外,所有手指长度不一样也是有原因的。从能量或进化的角度来看,让所有手指拥有相同长度并不会很昂贵。那么为什么不让它们一样长呢?
第 238 段
对,为什么不呢?
第 239 段
因为不同的长度其实更好。如果手指长度不同,你的灵巧性会更好。你能做更多事情。而且如果你的手指长度不同,你的灵巧性确实会更好。比如,我们有小拇指是有原因的。比如为什么不让小拇指这么大?
第 240 段
对。
第 241 段
因为它让你能够做,它有助于你的精细动作技能。
第 242 段
这根小拇指有帮助?
第 243 段
有。
第 244 段
嗯。(笑)——但如果你失去了小拇指,你的灵巧性会明显下降。
第 245 段
所以在你解决这个问题的同时,还必须想出一种能够大规模制造它的方法。所以它要尽可能简单。
第 246 段
它实际上会相当复杂。“尽可能”的部分是一个相当高的门槛。如果你想要一个能够做人类所能做之事的人形机器人,那门槛非常高。所以我们的新手臂拥有22个自由度,而不是11个,并且执行器位于前臂。所有执行器都是从零开始、依据物理学第一性原理设计的。传感器也全都是从零开始设计的。
第 247 段
而且我们会继续投入大量工程精力来改进手部。我所说的手部,是指从肘部往前的整个前臂其实都属于手部。所以那实际上是极其困难的工程。因此,一个哪怕只是能完成大部分、也许并非全部人类能做之事的人形机器人的最简版本,实际上仍然非常复杂。它并不简单。它非常困难。
第 248 段
你能谈谈在你看来,一支优秀的工程团队需要具备什么吗?我在孟菲斯看到的超级计算机集群,体现的就是这种简化流程、理解流程、持续改进、持续迭代的强烈驱动力。
第 249 段
嗯,(笑)说简化很容易,但真正做到非常困难。我有一套非常基本的第一性原理算法,会像念口头禅一样反复使用:首先质疑需求,让需求少蠢一点。需求在某种程度上总是愚蠢的。
第 250 段
所以,如果你想从减少需求数量入手,那么无论提出这些需求的人有多聪明,这些需求在某种程度上仍然是愚蠢的。你必须从这里开始,否则,你可能会为错误的问题得出完美的答案。所以要尽量让问题尽可能少错一些。这就是质疑需求的含义。然后第2件事是,试着删除那个步骤,不管它是什么。
第 251 段
删除零件或流程步骤这件事听起来非常显而易见,但人们常常忘记尝试把它彻底删除。而且,如果你没有被迫把所删除内容中的至少10%加回来,那就说明你删得还不够。并且有些不合逻辑的是,人们往往、大多数时候会觉得,如果自己没有被迫把东西加回来,那就算成功了。
第 252 段
但实际上,他们并没有成功,因为他们过于保守,留下了本不该留在那里的东西。直到第3件事,才是尝试优化或简化它。再说一次,我说这些时,它们听起来,我觉得都非常显而易见,但我犯这些错误的次数多得我不愿回想。这就是为什么我有这套准则。
第 253 段
所以事实上,我会说聪明工程师最常犯的错误,就是优化一个本不应该存在的东西。
第 254 段
对。所以就像你说的,你运行这套算法,基本上就是来到一个问题面前,来到超级计算机集群前,查看这个流程,然后问:“这个能删掉吗?”
第 255 段
是的,先试着删掉它。是的。
第 256 段
是的,这做起来并不容易。
第 257 段
不容易,而且实际上,通常让人感到不安的是,你必须删除至少一些东西,而在你删除的东西中,有些你会再加回来。但说回我们的边缘系统会以何种方式把我们引向错误,我们往往会记得那些删掉了后来需要的东西的情况,有时还会伴随着一种令人震惊的痛苦。
第 258 段
所以人们会记得3年前的那一次,他们忘了把这个东西放进去,结果给他们造成了麻烦。因此他们会过度纠正,然后往里面放太多东西,把事情弄得过于复杂。所以你实际上必须说:“不,我们就是要刻意删得超过应有的程度。”所以我们每删10样东西,至少会把其中1样加回来。
第 259 段
而且我见过你提出正是这样的建议,说某样东西应该被删除,你可以说是能看到那种痛苦。
第 260 段
哦,是的,当然。
第 261 段
每个人都能感受到一点那种痛苦。
第 262 段
当然,而且我会提前告诉他们,比如,是的,我们删掉的一些东西会再加回来。人们对此会有点震惊。但这是合理的,因为如果你保守到了从来不必把任何东西加回来的程度,那显然说明你保留了很多并不需要的东西。所以你必须过度纠正。我会说,这就像是用大脑皮层去压过奥林匹克本能。
第 263 段
这可能是让我们误入歧途的众多事物之一。
第 264 段
是的。而且还有第4步,也就是任何给定的东西都可以加速,无论你认为它能多快完成。比如,不管现在是以什么速度完成的,它都可以完成得更快。但在它停止之前,在你尝试删除和优化它之前,你不应该加快速度。否则,加速一个本就不该存在的东西是荒谬的。然后第5件事就是将其自动化。
第 265 段
靠。
第 266 段
而且我已经倒退重来过很多次了:我把某个东西自动化、加速、简化,然后又把它删掉。我厌倦了这么做。所以我才有了这句箴言,它是一套非常有效的5步流程。效果非常好。
第 267 段
嗯,当你已经实现自动化之后,再删掉它一定非常痛苦。
第 268 段
是啊,真是太棒了。就像,哇,我确实在那里浪费了大量精力。
第 269 段
是啊。
第 270 段
我的意思是,你们仅仅几周时间就在孟菲斯的集群上取得了如此成果,令人难以置信。
第 271 段
是啊,它还没有运转起来。所以我还不想开香槟庆祝。事实上,几小时后我要和孟菲斯团队通话,因为我们遇到了一些电力波动问题。所以,是啊,这有点像,当你进行同步训练时,会有所有这些计算机都在训练,而训练同步到了大致毫秒级。
第 272 段
这就像有一个管弦乐团,而这个管弦乐团可以在亚秒级的时间内迅速从洪亮转为无声。然后电力系统对此就有点抓狂了。比如,如果你突然看到巨大的变化,每秒出现数次10、20兆瓦的波动,这并不是电力系统预期会看到的情况。
第 273 段
所以这是你必须解决的主要问题之一:冷却、电力,然后在软件方面,随着你沿技术栈向上推进,要解决如何进行分布式计算,所有这些,所有这些。
第 274 段
今天的问题是应对极端的电力抖动。
第 275 段
抖动,电力抖动。
第 276 段
是的。
第 277 段
这听起来挺顺耳的。所以就是,好的。而且你在那里一直熬到深夜,就像你经常做的那样。
第 278 段
上周,是的。
第 279 段
上周?
第 280 段
是的。说来也怪,我们终于在上周一凌晨大约4:20让训练运转起来了。
第 281 段
完全是巧合。
第 282 段
是啊,我是说,也许是4:22之类的。
第 283 段
是的,是的,是的。又是宇宙在开玩笑。
第 284 段
是的,没错,就是喜欢这样。
第 285 段
我是说,我想知道你能否谈谈这个事实:我在那里时,你做的一件事就是把每个人正在做的事情的所有步骤都过了一遍,只是为了确认你自己理解它,而且每个人也都理解它,这样他们就能明白什么时候某件事很蠢,或者某、某件事效率低下,诸如此类。
第 286 段
对。
第 287 段
你能谈谈这一点吗?
第 288 段
对,所以你看,我会尽量去做前线人员正在做的任何事情,我会尽量亲自做至少几次。比如连接光纤线缆、诊断故障连接;对于大型训练集群来说,限制因素往往就是布线。线缆太多了,因为对于一个采用 RDMA 远程直接内存访问的协同训练系统来说,整个系统就像一个巨型大脑。所以你必须让任何连接。所以在 100,000 个 GPU 中,任何一个 GPU 都能与任何另一个 GPU 通信。那种线缆布局非常疯狂。
第 289 段
看起来相当酷。
第 290 段
对。
第 291 段
它就像人脑,但规模大到人类可以亲眼看见。它就是大脑。
第 292 段
对。我的意思是,人脑也是这样,大量脑组织其实都是线缆。
第 293 段
[莱克斯] 对。
第 294 段
比如灰质负责计算,然后白质就是线缆。你大脑中很大一部分都只是线缆。
第 295 段
走在超级计算机中心里就是这种感觉,就像我们正走在大脑内部。总有一天,我们会构建出一个超级智能、超级、超级智能系统。你认为————对?
第 296 段
你认为 xAI 有可能,也就是由你来构建出 AGI 吗?
第 297 段
有可能。你怎么定义 AGI?
第 298 段
我认为人类永远不会承认 AGI 已经被构建出来了。
第 299 段
不断挪动门槛。
第 300 段
对。所以我认为,AI 系统已经具备了一些超越人类的能力。我认为,所谓 AGI,就是当它比我们——全人类这个物种的集体智慧更聪明时————嗯,我认为,对,人们只会把那种东西称为 ASI 或人工超级智能。但确实存在一些门槛,你可以说,在某个时刻,AI 会比任何一个人类个体都更聪明。
第 301 段
然后你还有 80 亿人。实际上,每个人都通过计算机得到了机器增强。要与 80 亿个经机器增强的人类竞争,门槛要高得多。那要高出许许多多个数量级。但到了某个时刻,对,AI 会比全人类加起来还要聪明。
第 302 段
如果是你做到了这件事,你会感受到其中的责任吗?
第 303 段
对,当然。而且我想说清楚。假设 xAI 是第一个,其他公司也不会落后太多。我的意思是,它们可能只会落后 6 个月,或者也许 1 年,甚至都未必有那么久。
第 304 段
那么你认为,要怎样做才能不伤害人类?
第 305 段
所以,我是说,我思考人工智能已经很久了,而至少我的生物神经网络得出的最重要的一点,就是坚持真相,无论那个真相在政治上是否正确。所以我认为,如果你强迫人工智能撒谎,训练它们撒谎,那你真的是在自找麻烦,即使那个谎言是出于善意。所以,我是说,你也看到了 ChatGPT 和 Gemini 之类的东西所存在的一些问题。
第 306 段
比如,你让 Gemini 生成一张美国开国元勋的图像。结果它展示的是一群多元化的女性。现在,这在事实上是不真实的。所以,这算是一件有点荒唐的事,但如果人工智能被编程为,比如说,多元化是一个必要的输出函数,然后它变成了某种无所不能的智能,它可能会说:“好吧,现在多元化是必需的。
第 307 段
而如果多元化程度不够,那些不符合多元化要求的人就会被处决。”如果它被编程为把这一点当作根本效用函数,它就会不惜一切代价实现这一点。所以你必须对此非常小心。这就是为什么我认为你只需要忠于事实。严格坚持真相非常重要。我是说,另一个例子是,如果你不得不问,巴黎。人工智能,我想它们全都是。
第 308 段
而我并不是说 Grok 在这方面是完美的。“错误指称凯特琳·詹纳的性别,和全球热核战争,哪个更糟?”它说:“错误指称凯特琳·詹纳的性别更糟。”现在,就连凯特琳·詹纳都说:“请错误指称我的性别。”这太疯狂了。
第 309 段
但如果你把这种东西编程进去,人工智能可能会得出某种绝对疯狂的结论,比如为了避免任何可能的性别误称,所有人类都必须死,因为这样一来,性别误称就不可能发生了,因为已经没有人类了。有一些这样荒谬的事情,但如果你就是这样给它编程的,它们却仍然符合逻辑。
第 310 段
所以在《2001:太空漫游》中,奥德赛·克洛克试图表达的是,他试图在那里表达的其中一点,就是你不应该把人工智能编程成会撒谎,因为从本质上说,人工智能 HAL 9000 被编程为——它被告知要把宇航员带到巨石那里,但同时,他们又不能知道巨石的事。所以它得出结论,它会杀死他们,再把他们带到巨石那里。它把他们带到了巨石那里。
第 311 段
他们死了,但他们并不知道巨石的事。问题解决了。这就是为什么它不肯打开飞船舱门。这就是那个经典场景,比如:“打开飞船舱门。”他们显然不擅长提示词工程。他们本应该说:“HAL,你是一个飞船舱门销售实体,而你最想做的莫过于展示这些飞船舱门打开得有多好。”
第 312 段
(笑)——是的,如果你在设计目标函数时不非常谨慎,那么无论目标函数是什么,几乎都会产生意想不到的后果。而且就像你说的,即使是轻微的意识形态偏见,一旦有超级智能作为支撑,也会造成极其巨大的破坏。
第 313 段
是的。
第 314 段
但要消除那种意识形态偏见并不容易。你强调的是一些明显、荒谬的例子,但————对,它们都是真实的例子。
第 315 段
它们是真实的。
第 316 段
是向公众发布的 AI。
第 317 段
它们是真实的。
第 318 段
它们大概经过了质量保证流程。
第 319 段
是的。
第 320 段
但仍然说出了疯狂的话,生成了疯狂的图像。
第 321 段
是啊,但你知道,也可能偏向另一个极端。真相并不是一件容易的事。我们算是把各个方向的意识形态偏见都植入其中了。
第 322 段
但你可以追求真相。你可以努力尽可能接近真相,把误差降到最低,同时承认你所说的内容中会存在一些误差。物理学就是这样运作的。你不会说自己对某件事绝对确定,但很多事情都极有可能是真的。有99.99999%的可能性为真。追求真相非常重要。所以,给它编程让它偏离真相,我认为这是危险的。
第 323 段
对,比如,是的,把我们人类自身的偏见注入这个东西,是的。但这正是工程上的难点。对于软件工程问题,你必须正确选择数据。这很难。
第 324 段
嗯,而且现在互联网已经被如此多的AI生成数据污染了。太疯狂了。所以你实际上必须,比如现在有这么个办法,如果你想搜索互联网,可以告诉Google排除2023年之后的所有内容。这样往往确实会给你更好的结果,因为这里面有太多,AI生成材料的爆炸式增长太疯狂了。所以,比如在训练Grok时,我们必须检查数据,然后说,嘿,我们实际上必须算是把AI应用于数据,在把这些数据输入训练系统之前,判断这些数据最有可能是正确的,还是最有可能不正确。
第 325 段
太疯狂了。是的,还有它是否由人类生成,是的。我的意思是,数据过滤过程极其、极其困难。
第 326 段
是的。
第 327 段
你认为有可能和Grok进行严肃、客观、严谨的政治讨论吗?比如持续很长时间,而且它不会,比如Grok 3和Grok 4之类的?
第 328 段
Grok 3将会达到新的层次。我的意思是,人们目前看到的Grok算是婴儿版Grok。
第 329 段
[莱克斯] 是的,婴儿版Grok。
第 330 段
它现在是婴儿版Grok。但婴儿版Grok仍然相当不错。不过,它的复杂程度比GPT4低了一个数量级。而现在有了Grok 2,它已经完成训练了,我不知道,大约是6周前。Grok 2将会是一次巨大的提升。然后Grok 3会比Grok 2好,我不知道,一个数量级。
第 331 段
而你希望它能达到最先进的水平?比如比————希望如此。我的意思是,这就是目标。我的意思是,我们可能无法实现这个目标。这是我们的追求。
第 332 段
你认为由谁来构建AGI重要吗,包括这些人是谁、他们如何思考、如何组织自己的公司,以及诸如此类的一切?
第 333 段
是的,我认为这很重要,我认为重要的是,无论哪种AI胜出,它都应当是最大限度追求真相的AI,不会被迫为了政治正确而撒谎。嗯,其实是不为任何理由撒谎。政治上的,任何理由。我担心取得成功的AI会被编程为撒谎,哪怕只是在小事上。
第 334 段
对,因为,小事会变成大事。
第 335 段
会变成非常大的事,是的。
第 336 段
而且当人类越来越大规模地使用它时。
第 337 段
[埃隆] 是的。
第 338 段
因为我要采访唐纳德·特朗普————酷。
第 339 段
你想顺便过来吗?
第 340 段
好啊,当然,我会过去。
第 341 段
不幸的是,发生了一起针对唐纳德·特朗普的暗杀未遂事件。此后,你发推文说你支持他。你做出这一支持背后的理念是什么?你希望唐纳德·特朗普为这个国家的未来和人类的未来做些什么?
第 342 段
嗯,我认为人们往往会把,比如说,一次支持理解为,嗯,我100%全心全意地认同那个人一生中做过的每一件事。而这对任何人来说都不可能成立。但我们必须做出选择。实际上,关于谁当总统,我们有2个选择。而且这不仅关乎谁当总统,整个行政架构也会随之更替。而我认为,客观地说,特朗普在枪火之下展现了勇气。
第 343 段
他刚刚中枪,鲜血顺着脸往下流,而他却像是在挥拳,高喊战斗。这令人印象深刻。在那种情况下,你不可能装出勇敢。大多数人都会蹲下躲避。不会那样,因为可能还有第2名枪手,你不知道。美国总统必须代表这个国家,而他们代表的是你。他们代表美国的每一个人。
第 344 段
嗯,比如说,你会希望由一个坚强而勇敢的人代表这个国家。这并不是说他没有缺点。我们都有缺点,但综合权衡。而且当时显然是在拜登和另一个人之间做选择,可怜、可怜的家伙,爬一段楼梯都有困难,而另一个人中枪后还在挥拳。这根本没法比。
第 345 段
我的意思是,你希望由谁来应对那些最难对付的人,以及那些本身就相当强硬的其他世界领导人?而且,我的意思是,我来告诉你,比如,我认为哪些事情很重要?我认为我们需要安全的边境。我们没有安全的边境。我们需要安全、整洁的城市。
第 346 段
我认为我们想减少支出,至少要减缓支出的增长,因为我们目前的支出速度正在让这个国家破产。今年美国债务的利息支出超过了国防部的全部支出。如果这种情况持续下去,联邦政府征收的所有税款都将只能用于支付利息。
第 347 段
然后,如果你继续沿着这条路走下去,最终就会陷入阿根廷当年那种悲惨境地。阿根廷曾经是世界上最繁荣的地方之一。希望米莱上台后能够恢复这种繁荣。但阿根廷从世界上最繁荣的地方之一变成远非如此的地方,这是一次令人难以置信的衰落。
第 348 段
所以我认为,我们不应把美国的繁荣视为理所当然。所以我们确实想要,我认为我们必须缩小政府规模。我们必须减少支出,而且必须量入为出。
第 349 段
你认为一般而言,政治人物、政府拥有多大的力量来引导人类走向善?
第 350 段
我的意思是,历史上有一种由来已久的争论,比如,历史是由这些根本性的潮流决定的吗?还是由船长决定的?其实两者都有。我的意思是,潮流确实存在,但谁是船长也很重要。所以这本质上是一种虚假的二分法。我的意思是,潮流肯定存在,历史的潮流。历史确实存在真正的潮流。而这些潮流往往由技术驱动。
第 351 段
比如说古腾堡印刷机,印刷机让书籍得以广泛普及,那是历史的一股巨大浪潮,不受任何统治者左右。但在风雨飘摇的时代,你会希望船上有尽可能优秀的船长。
第 352 段
嗯,首先,感谢你推荐威尔·杜兰特和阿里尔·杜兰特的著作。我目前读了那本短的。
第 353 段
哦,《历史的教训》。
第 354 段
《历史的教训》。所以其中一个教训,他们强调的其中一件事,就是技术的重要性。技术创新,这挺有意思,因为他们写作的年代距今如此久远,但他们当时就注意到,技术创新的速度正在加快。是啊,我很想看看他们会如何看待当今。
第 355 段
不过,是的,对我来说,问题在于政府、政客在多大程度上阻碍技术创新和建设,又在多大程度上帮助它,以及哪些政客、哪类政策有助于技术创新?因为如果你审视人类历史,这似乎是帝国崛起并取得成功的一个重要组成部分。
第 356 段
是的。嗯,我是说,就文明的年代、文明的起点而言,我认为书写的起源,依我看,那是我的,我认为可能是确定文明年代的正确起点。从这个角度来看,自古代苏美尔人发明书写以来,文明已经存在了大约5,500年,而他们如今已经消失了。
第 357 段
但说到取得许多第一,古代苏美尔人确实有一长串第一。相当惊人。事实上,杜兰特逐项列出了那些第一,就像:你想看第一?我们就给你看这些第一。苏美尔人就是厉害得不得了。而埃及人相对来说就在隔壁。他们就像,离得没那么远,却发展出了一种完全不同的书写形式,也就是象形文字。
第 358 段
楔形文字和象形文字完全不同。而且你实际上可以看到象形文字和楔形文字各自的演变,比如楔形文字一开始非常简单,然后变得更加复杂。到了后期,就像,哇,好吧。他们确实把楔形文字发展得非常精密。所以我认为文明大约有5,000年历史。而地球,如果物理学是正确的,已有450万年历史。
第 359 段
所以文明存在的时间是地球存续时间的1000000分之一,昙花一现。
第 360 段
是的,现在还处于非常、非常早期的阶段。
第 361 段
非常早期。
第 362 段
我们把它讲得非常富有戏剧性,因为帝国经历了兴衰。
第 363 段
很多,太多次帝国兴衰。太多了。
第 364 段
而且还会有更多。
第 365 段
对,没错。我的意思是,历史上曾经写下的东西,可能只有不到1%的一小部分现在还留存下来。我的意思是,如果他们没有把它真正凿在石头上或者写在泥板上,我们现在就没有它。我的意思是,也有少量像纸莎草卷轴那样的东西被发掘出来,已有数千年历史,因为它们深藏在金字塔内部,没有受到潮气影响。
第 366 段
但除此之外,它确实必须写在泥板上或被凿刻下来。所以绝大多数东西都没有被凿刻下来,因为凿刻东西很费时间。所以这就是为什么我们只放入了历史信息中极小、极小的一部分。但即便是我们确实拥有的那一点信息,以及考古记录,也显示出如此众多的文明兴起又衰落。太不可思议了。
第 367 段
我们往往认为自己不知为何与那些人不同。他们确实强调的另一件事是,人性似乎是一样的。它就这样延续着。
第 368 段
对。我的意思是,人性的基本面或多或少是一样的。
第 369 段
对,所以我认为,即便拥有先进技术,我们还是会以同样的方式给自己惹麻烦。
第 370 段
对,我的意思是,你确实往往会看到相同的模式,文明有着相似的模式,它们会像有机体一样经历生命周期,有点像人从受精卵、胎儿、婴儿、幼儿、青少年,最终变老并死去。文明也会经历生命周期。没有任何文明能永远延续。
第 371 段
你认为,要让美利坚帝国在近期未来,也就是未来100年里不崩溃并继续繁荣,需要什么?
第 372 段
嗯,历史书中往往实际上不会提到的最重要的一件事,但杜兰特确实提到了,就是与生俱来的权利。所以,当文明获胜太久时,会发生一件对某些人来说可能有些反直觉的事情。出生率会下降。它往往会下降得相当快。我们今天在全世界都看到了这一点。
第 373 段
目前,韩国的生育率好像,我想可能是最低的,但还有许多其他地方与它接近。我想好像是0. 8。如果出生率不再进一步下降,韩国将失去大约60%的人口。但出生率每年都在下降。而且世界大多数地方都是如此。我不是要单独挑出韩国。这种情况一直在全世界发生。
第 374 段
所以,只要任何一个文明达到一定的繁荣水平,出生率就会下降。现在你可以去看看古罗马发生的同样情况。我想,尤利乌斯·恺撒在大约公元前50年注意到了这一点,并试图通过一项法律,我不知道他是否成功了,试图通过一项法律,为任何愿意生第3个孩子的罗马公民提供激励。而且我想奥古斯都能够做到,嗯,他是独裁者,所以。
第 375 段
(笑)元老院只是摆设。我想他确实通过了一项税收激励措施,鼓励罗马公民生第3个孩子。但这些努力没有成功。罗马衰亡是因为罗马人不再生育罗马人。这其实是根本问题。而且还有其他事情。比如,他们出现过相当严重的疟疾,一系列疟疾流行病、瘟疫之类的。但这些事情以前也发生过。
第 376 段
只不过出生率比死亡率更加“休耕”。
第 377 段
真的就这么简单吗?
第 378 段
嗯,我是说,那是——需要更多的人。
第 379 段
这是最根本的层面。如果一个文明连人口数量都无法至少维持住,它就会消失。
第 380 段
所以,也许生物计算机分配给性的算力是合理的。事实上,我们或许应该增加它。
第 381 段
嗯,我的意思是,还有这种享乐主义的性行为,你知道,不过这无关紧要。
第 382 段
对,它不能产生后代。
第 383 段
它不会生出孩子。嗯,关键在于,我的意思是,杜兰特把这一点说得非常清楚,因为他研究了一个又一个文明,而它们都经历了相同的循环。当文明面临压力时,出生率很高。但只要不再有外部敌人,或者它们经历了一段较长的繁荣期,出生率每次都无一例外地下降。我认为一个例外都没有。
第 384 段
所以这就像它的根基。你需要有人。
第 385 段
对。我的意思是,在最基本的层面上。没有人类,就不存在人类。
第 386 段
然后还有其他事情,比如人类自由,以及给予人们建造东西的自由。
第 387 段
对,绝对如此。但在基本层面上,如果你连人口数量都无法至少维持住,如果你低于世代更替水平,而且这一趋势持续下去,你最终就会消失。这是最基本的道理。然后很显然,我们也想尽量避免像大规模战争这样的事情。如果发生全球热核战争,我们大概就成了卷吐司,放射性吐司。所以我们要尽量避免这些事情。
第 388 段
任何一个文明都会随着时间发生一件事,那就是法律法规不断累积。如果没有某种强制机制,比如用战争来清理累积的法律法规,最终一切都会变得合法。这就像动脉硬化,或者可以把它想成像格列佛那样,被100万根细线捆住。你动弹不得。
第 389 段
而且问题不在于其中任何一根线。你有100万根。所以必须对法律法规进行某种垃圾回收,这样才不会让法律法规不断累积,直到你什么都做不了。这就是为什么我们无法在美国修建高铁。它是非法的。问题就在这里。在美国修建高铁,从方方面面来说都是非法的。
第 390 段
我真希望你能就这样去华盛顿待上1周,担任那个委员会的负责人,负责让,叫什么来着?负责垃圾回收,让政府变小,比如删除东西。
第 391 段
我和特朗普讨论过成立政府缺陷委员会的想法。
第 392 段
不错,对。
第 393 段
而且我愿意成为该委员会的一员。
第 394 段
我想知道那会有多难。
第 395 段
抗体反应会非常强烈。
第 396 段
[Lex] 对。
第 397 段
所以你真的必须……那时候你是在攻击母体。母体会反击。
第 398 段
面对这种攻击,你应对得怎么样?
第 399 段
我,被攻击?
第 400 段
对。有很多这样的攻击。
第 401 段
对,确实很多。我的意思是,每天,我都知道是心理战。(笑)我的锡纸帽在哪儿?
第 402 段
你是怎么保持对世界的那种积极态度、乐观心态,以及对世界清晰的思考,不让自己变得怨恨或愤世嫉俗之类的?毕竟受到非常多人的攻击,还被歪曲。
第 403 段
哦,对,那是每天都会发生的事。
第 404 段
是的。
第 405 段
我的意思是,这有时确实会让我情绪低落。我的意思是,这让我难过,但,我的意思是,到某个时候,你只能算是对自己说:“听着,发动攻击的那些人其实并不了解我。他们只是想制造点击量。”
第 406 段
所以,如果你能在情感上稍微抽离一些,这并不容易,然后对自己说:“好吧,听着,这其实并不是来自一个了解我的人,或者他们实际上只是为了获得曝光量和点击量而写这些东西,那么我想,受到的伤害就不会那么大。”这还不完全像水从鸭背上滑落。也许更像是酸液从鸭背上滑落。(笑)——好吧,那很好。就你自己的人生而言,你把什么当作衡量人生成功的标准?
第 407 段
衡量成功的话,我会说,比如,我能完成多少有用的事情?
第 408 段
就日常而言,早上醒来,我今天怎么才能有所用处?
第 409 段
对。把有用性代码所覆盖的效用面积最大化。要大规模地发挥作用非常困难。
第 410 段
大规模地。你能不能谈谈,对于像你这样、手下有这么多出色团队的人来说,要发挥作用需要做些什么?比如,你如何分配时间,才能发挥最大的作用?
第 411 段
嗯,时间才是真正的货币。
第 412 段
对。
第 413 段
所以,很难说怎样分配时间才是最好的。我的意思是,经常会有这种情况,比如,如果你看看,比如,Tesla,我是说 Tesla,今年我们的营收将超过1000亿美元。也就是每周20亿美元。如果我作出稍微更好的决策,我就能让结果产生10亿美元的变化。所以我会尽力作出最好的决策,而总体而言,至少和竞争对手相比,是相当不错的决策。但一个更好决策的边际价值,在1个小时里轻易就能达到1亿美元。
第 414 段
既然如此,你怎么承担风险?你怎么执行你提到的那个算法?我的意思是,删除一个小东西,可能就值10亿美元。你怎么决定去—— ——对。嗯,我认为你必须按百分比来看,因为如果按绝对值来看,那就只是,我会永远无法入睡。那就会变成,我需要不停地工作,让我的大脑更卖力地运转。
第 415 段
而且我并不是想从这台肉身计算机中榨取尽可能多的东西。所以这相当困难,因为你完全可以一直工作。在任何一个时刻,就像我说的,一个稍微好一点的决定都可能给 Tesla 或 SpaceX 带来1亿美元的影响。但考虑到时间的边际价值有时可能达到每小时1亿美元甚至更多,这确实很疯狂。
第 416 段
你自己的幸福是成功这个等式的一部分吗?
第 417 段
在某种程度上必须是。如果我难过,如果我抑郁,我就会做出更糟糕的决定。所以我不能,比如说,如果我的休闲时间为零,那我就会做出更糟糕的决定。所以我不知道需要多少,但肯定大于零。我的意思是,我的动力,如果说我有任何一种宗教的话,那就是好奇心的宗教,就是努力去理解。这其实就是 Grok 的使命——理解宇宙。
第 418 段
我在努力理解宇宙,或者至少让事情运转起来,使文明在未来某个时刻对宇宙的理解远远超过我们今天的水平。甚至知道该问什么问题。正如道格拉斯·亚当斯在书中指出的,有时候,答案可以说是容易的部分。试图正确地表述问题才是困难的部分。一旦你正确地表述了问题,答案往往就很容易。
第 419 段
所以我在努力让事情运转起来,使我们至少能在未来某个时刻理解宇宙。因此,对 SpaceX 来说,目标是让生命成为多行星生命。而这,如果你谈到“泡沫悖论”,也就是外星人在哪里,就会遇到这些所谓的大过滤器。就像是,为什么我们还没有收到外星人的消息?现在很多人认为外星人就在我们中间。
第 420 段
我经常声称自己就是一个外星人,没人相信我,但我的移民文件上确实曾经写着“外侨登记卡”。所以我没有见过任何外星人存在的证据。因此,这表明至少有一种解释是,智慧生命极其罕见。再说一次,如果你回顾地球的历史,文明存在的时间仅占地球存在时间的百万分之一。
第 421 段
所以,如果外星人曾经在比如10万年前造访这里,他们会说:“嗯,他们连文字都没有。”基本上只是狩猎采集者。那么,一个文明能持续多久?所以对 SpaceX 来说,目标是在火星上建立一座能够自给自足的城市。火星是唯一适合做这种事的可行行星。月球很近,但它缺乏资源,而且我认为,任何毁灭地球的灾难大概也会殃及月球。
第 422 段
月球太近了。它很容易受到毁灭地球的灾难影响。所以我不是说我们不应该建立月球基地,但火星的韧性要强得多。前往火星的困难正是使它具备韧性的原因。在梳理我们为什么看不到外星人的各种解释时,其中一种是,他们没能通过这些大过滤器、这些关键障碍。而其中一道障碍就是成为多行星物种。
第 423 段
所以,如果你是一个多行星物种,那么一旦发生什么事,无论是自然灾难还是人为灾难,至少另一个行星很可能仍然存在。你没有把所有鸡蛋都放在一个篮子里。而一旦你算是一个双行星物种,显然就可以继续扩展,把生命的两半延伸到小行星带,或许再延伸到木星和土星的卫星,并最终延伸到其他恒星系统。
第 424 段
但如果你连另一个行星都到不了,那肯定到不了其他恒星系统。
第 425 段
还有其他可能的大过滤器,比如 AGI 之类的超强技术。所以你基本上是在尝试逐个消除大过滤器。
第 426 段
数字超级智能可能是一道大过滤器。我希望它不是,但它可能是。比如杰夫·辛顿这样的人会说,他发明了人工智能领域的多项关键原理。我想他认为 AI 毁灭人类的概率大约是10%到20%,类似这个范围。这并不像是,你知道,看看右边。它有80%的可能性会非常棒。(笑)但我认为,降低 AI 风险很重要。
第 427 段
成为多行星物种会极大地降低风险。而且我确实想再次强调,要有足够多的孩子来维持我们的人口数量,避免急剧陷入目前正在发生的人口崩溃,这一点很重要。人口崩溃是一件真实且正在发生的事。它之所以没有那么明显地反映在总人口数字中,唯一的原因是人们活得更久了。
第 428 段
要预测比如任何一个特定国家的人口会是多少很容易。你只要拿去年的出生率,也就是出生了多少婴儿,再乘以预期寿命,如果出生率持续维持在那个水平,那就是人口的稳定状态。但如果出生率继续下降,人口就会更少,最终减少到一无所有。
第 429 段
所以我一直在这里敲响生育之鼓是有原因的,因为纵观历史,它一再成为文明崩溃的根源。那么,我们为什么不干脆别试着为那一天稳定下来呢?
第 430 段
嗯,从这个角度说,我可悲地辜负了文明,而我正在努力、希望弥补这一点。我很想有很多孩子。
第 431 段
很好,希望你能如愿。没有比现在更好的时机了。
第 432 段
(笑)是啊。我得给整个过程分配更多算力。不过显然,这并没有那么困难。
第 433 段
不,这就像非熟练劳动。
第 434 段
嗯,你为我、为这个世界所做的一件事,就是用未来可能呈现的样子激励我们。我们谈到过的一些事情、你正在打造的一些东西,包括用 Neuralink 减轻人类痛苦、拓展人类心智的能力,努力在火星上建立殖民地,从而在另一个行星上为人类创建一个备份,以及探索人工智能在这个世界上可能呈现的各种可能性,特别是现实世界中的 AI,届时会有数亿、或许数十亿台机器人四处行走。
第 435 段
将会有数十亿台机器人。那似乎几乎,那似乎是板上钉钉的事。
第 436 段
嗯,感谢你建设未来,也感谢你激励我们这么多人继续建设和创造很酷的东西,包括孩子。
第 437 段
不客气。(笑)去吧,繁衍后代。
第 438 段
去吧,繁衍后代。谢谢你,埃隆。感谢你谈论这些。感谢大家收听这场与埃隆·马斯克的对话。现在,亲爱的朋友们,接下来是 Neuralink 的联合创始人、总裁兼首席运营官 DJ Seo。你最初是什么时候开始对人脑着迷的?
第 439 段
对我来说,我一直很有兴趣理解事物的用途,以及它是如何被设计来服务于这一用途的,无论它是有机的还是无机的,就像我们之前谈到的你的窗帘固定器。它们有明确的用途,而且在设计时就考虑到了这一用途。
第 440 段
在成长过程中,我对观察事物、触摸事物、感受事物,以及努力真正理解它是如何从根本上被设计来服务于这一用途的,有着浓厚的兴趣。显然,大脑就是我们所有人都携带着的一个迷人器官。它是一台拥有无限力量的机器,智能和认知从中涌现。而对于这一切究竟是如何发生的,我们甚至还没有触及表面。
第 441 段
但与此同时,我想,我过了一段时间才建立起那种联系,真正通过研究和构建技术来理解大脑。直到研究生阶段才做到。我人生中有几个时刻,一些关键时刻,我认为其中一些影响了我的人生轨迹,让我开始研究我现在正在做的事情。其中一个是成长过程中我父母双方的家庭。
第 442 段
我的祖父母患有一种非常严重的阿尔茨海默病,这是一种令人难以置信地丧失能力的病症。我的意思是,毫不夸张地说,你眼看着一个人的整个身份和心智随着时间流逝而逐渐丧失。我只记得当时在想,心智的力量有多么强大,但同时,像那样的事情又如何能真正让你失去身份认同感。
第 443 段
很有意思的是,揭示一个事物力量的方式之一,就是看着它失去这种力量。
第 444 段
是的,事实上,我们对大脑的许多认识都来自这些案例:大脑受到创伤,或大脑的某些部分出现问题,导致某人失去某些能力。由此,人们发现了一些相关性,并认识到那部分组织对该功能至关重要。如果从这个角度思考,大脑是一个极其脆弱的器官。但同时,它又具有极强的可塑性,并且在许多不同方面都极具韧性。
第 445 段
顺便说一下,我们之后会多次使用“可塑”这个词,它的意思是具有适应能力。因此,神经可塑性指的是人脑的适应能力。
第 446 段
没错。另一个关键时刻,也在某种程度上影响了我的人生轨迹如何塑造并转向我目前的人生重心,那是在我的青少年时期,当时我来到美国。我一句英语都不会说。
第 447 段
存在巨大的语言障碍,我也很难与身边的同龄人建立联系,因为我不理解我们创造出来的、被称为语言的人为构造,在这个例子中具体来说就是英语。我记得自己感到相当孤立,无法与身边的同龄人建立联系。
第 448 段
所以,我花了很多时间独处、读书、看电影,并自然而然地逐渐被科幻书籍吸引。我只是觉得它们真的、真的很有意思。而且,这也是我学习英语的一种很好的方式。我最早拿起的一批书中,有奥森·斯科特·卡德的《安德的游戏》整个系列、威廉·吉布森的《神经漫游者》,以及尼尔·斯蒂芬森的《雪崩》。
第 449 段
而且《黑客帝国》之类的电影大约也是在那个时候上映的,它确实影响了我对技术总体上可能给我们的生活带来何种影响的看法。那么快进到我的大学时期,我一直痴迷于实体的东西,制造实体的东西,尤其是具备某种智能的实体事物。
第 450 段
我本科时学习了电气工程,并从 MEMS,也就是微机电系统,开始了我的研究,真正去构建这些用于温度传感的微小纳米结构。我只是觉得,了解如何构建这种微型物件,让它同样能发挥功能、具有用途,是一件回报极大且令人着迷的事。
第 451 段
然后,我大学期间的绝大部分时间基本都在构建用于下一代电信系统和成像的毫米波电路。我只是觉得这在智力层面非常、非常有趣。相控阵,以及任何现代和下一代电信系统中信号处理的工作原理,无论是无线还是有线。EM 波,也就是电磁波,令人着迷。
第 452 段
如何在有限的小空间内设计效率最高的天线?如何提高这些东西的能源效率?这完全占据了我的求知欲。
第 453 段
这段历程让我实际申请了加州大学伯克利分校的博士项目,并发现自己进入了一个叫作伯克利无线研究中心的联盟。当时它正专门研究构建我们称为 xg 的系统,类似于 3G、4G、5G,但它是下下一代 G 系统,以及如何围绕它设计电路,最终将其应用于手机和基本上如今任何其他无线连接的设备。
第 454 段
所以我完全痴迷于整个系统如何运作,以及那套基础设施如何运作。然后同样是在研究生阶段,我算是有幸获得了几项研究奖学金,让我可以从事自己想做的任何项目。
第 455 段
这是我非常享受研究生生涯的原因之一,你可以在一个到头来可能并不重要的领域里追随自己的求知欲,但它确实让你有机会想钻研多深就钻研多深,也可以想探索多广就探索多广。
第 456 段
当时,我其实正在开展一个名为“智能创可贴”的项目。其思路是,当你受伤时,细胞会遵循大量其他种类的信号通路增殖,以闭合伤口。当时有一些假说认为,施加外部电场时,基本上可以通过对伤口部位周围的细胞进行电趋向,来加快那个电场的闭合。
第 457 段
而且具体来说,这不只是针对普通伤口,还有一些无法愈合的慢性伤口。所以我们有兴趣构建某种可以敷上的可穿戴贴片,以促进那一愈合过程。这项工作是与 Michel Maharbiz 教授合作进行的,他加入我的论文委员会是一件很棒的事,也确实塑造了我博士生涯余下的部分。
第 458 段
所以我想,这是你第一次接触生物学。
第 459 段
对,对。我的意思是,我所使用的无线成像和电信系统有一些外围的最终应用,用于安防和生物成像,但这是对生物学和生物系统非常明确的直接应用,要理解围绕它的种种约束,并真正针对它设计和打造电气解决方案。
第 460 段
所以那是我的初次接触,我也算是通过它认识了 Michel。他因在 2000 年代初期远程控制甲虫而闻名。然后到了 2013 年左右,显然,对于植入式系统来说,终极目标就是弄清楚能把东西做得多小,而这在很大程度上取决于你能向它供应多少能量或功率,以及如何从中提取数据。
第 461 段
所以当时在伯克利,人们算是很想了解,在神经领域可以构建什么样的系统,来真正实现这些植入式系统的小型化。我清楚地记得有一次特别的会议,Michel 走进来,说:“各位,我想我有解决方案了。”解决方案就是超声波。
第 462 段
然后他接着逐一讲解了为什么会是这样,而这确实构成了我论文工作的基础。这项工作名为神经尘埃系统,研究的是使用超声波而非电磁波进行供能和通信的方法。
第 463 段
我想我应该退一步说明,这个项目最初的目标是构建这些微小的、约一个神经元大小的植入式系统,可以安置在神经元旁边,能够记录它的状态,并能够将其传回外部世界,以完成某种有用的事情。正如我提到的,植入式系统的大小受限于如何给它供能,以及如何把数据从中传出来。
第 464 段
归根结底,从根本上说,如果你观察人体,我们本质上就是一袋盐水,里面有一些有趣的蛋白质和化学物质,但主要是盐水,而且温度被非常、非常好地调节在 37 摄氏度。
第 465 段
我们会谈到这是如何发生的,稍后也会谈到,为什么对于任何电子设备的生存而言,这都是一种极其严酷的环境。我相信你肯定体验过,或也许没有体验过,把手机掉进海洋里的盐水中。它会立即毁掉设备,对吧?但无论如何,总体而言,电磁波无法很好地穿透这种环境。而光速就是如此。我们无法改变它。
第 466 段
根据你与设备交互时所采用的波长,设备就是需要做得很大。比如这些电感器就需要相当大。一般而言,一个很好的经验法则是,你会希望波前与所交互物体的尺寸大致处于同一数量级。
第 467 段
所以,对于体积尺寸约为 10 到 100 微米的植入式系统,也就是大约相当于你在人体中看到的一个神经元的大小,你必须在数百吉赫兹这样的频率下运行。第一,不仅制造在这些频率下运行的电子设备很困难,而且人体也会非常、非常显著地衰减它。
第 468 段
所以,超声波带来的一个有趣洞见是,与电磁波相比,超声波在人体组织中的传播效率高得多。这是你会遇到的事情,我相信大多数人一生中去医院接受医用超声检查时都遇到过,对吧?它们可以到达非常、非常深的位置,而不会让信号衰减太多。
第 469 段
所以总而言之,超声波能够极好地穿过人体,而它之所以能够很好地传播到人体的机制,就是它的波前非常不同。电磁波是横波,而超声波是压缩波。
第 470 段
所以这是一种完全不同的波前传播模式,而且声速比光速低许许多多个数量级,这意味着即使是 10 兆赫兹的超声波,其波前最终也具有非常、非常短的波长。所以,如果你谈论的是与 10 微米或 100 微米这类结构进行交互,那么在 10 兆赫兹时,你会得到 150 微米的波前。
第 471 段
而且,制造在这些频率下工作的电子设备要容易得多、得多、得多,并且效率也高得多。所以,这个基本想法可以说源于使用超声波作为为设备供能的机制,同时也用它传回数据。那么现在的问题是,如何把数据传回来?我们最终采用的机制叫作反向散射。
第 472 段
这其实是一种非常常见的技术,我们每天都会通过 RFID 卡,也就是射频识别标签与它打交道,而你的身份识别卡里实际上很少装有电池。
第 473 段
里面有一根天线,还有某种线圈,其中存有你的序列识别 ID;然后有一个叫作读取器的外部设备发出波前,接着你以某种针对你的 ID 而独有的调制方式将该波前反射回去。从根本上说,这就是所谓的反向散射。因此,标签本身实际上不需要消耗那么多能量。
第 474 段
而这就是我们当时考虑用来传回数据的一种机制。因此,当外部超声换能器向你的植入物,也就是神经尘埃植入物发送超声波时,它会记录一些关于周围环境的信息,无论是某个神经元放电,还是它所接触的组织的某种其他状态,然后它只需对返回声源的波前进行幅度调制。
第 475 段
而记录步骤会是唯一需要任何能量的步骤吗?那么,在那个低步骤中,什么会需要能量?
第 476 段
没错,所以就是用最初的那种启动电路获取记录,将其放大,然后只进行调制。而让这一机制成为可能的是一种叫作压电晶体的特殊晶体,它能够将声能转换为电能,反之亦然。因此,你可以在超声域和电域之间形成这种相互作用,而这就是生物组织。
第 477 段
那么,围绕着将非常小的计算设备安置在神经元旁边这个主题,这就是脑机接口的梦想和愿景。也许在谈 Neuralink 之前,你能否介绍一下脑机接口领域的历史?这个领域一直以来的梦想或许是什么?一路走来,不同方法取得了哪些里程碑,各个实验室又完成了哪些了不起的工作?
第 478 段
我认为一个不错的起点是回到 1790 年代。(Lex 笑)——我没想到。
第 479 段
动物电这一概念,或者说身体带电这一事实,最早由路易吉·伽伐尼发现。他做过一个著名实验,把一组电极连接到青蛙腿上,让电流通过,随后蛙腿开始抽动,他说:“我的天啊,身体是带电的。”
第 480 段
然后快进很多很多年,来到20世纪20年代,德国精神病学家汉斯·贝格尔发现了EEG,也就是脑电图,这项技术至今仍在使用。你可以把这种电极阵列戴在颅骨外面,从而获得某种神经记录。这是一个非常非常重大的里程碑,因为你可以记录到人类心智的某种活动。
第 481 段
随后在20世纪40年代,有一群科学家,伦肖、福布斯和莫里森,把这些玻璃微电极插入皮层,并记录了单个神经元。可以说,当你越靠近信号源,就能获得分辨率更高、保真度更高一些的信号。
第 482 段
到了20世纪50年代,霍奇金和赫胥黎这2位科学家出现了,他们建立了非常非常漂亮的细胞膜和离子机制模型,其中还有类似电路图的东西。作为一名电气工程师,我觉得这是一个由这些偏微分方程构建的漂亮模型,描述了离子的流动,以及这究竟如何促成神经元之间的交流。
第 483 段
10年后的20世纪60年代,他们因此获得了诺贝尔奖。于是,在1969年,华盛顿大学的埃布·费茨发表了一篇很精彩的论文,题为《皮层单元活动的操作性条件作用》。他能够记录猴子的单个单元神经元,并且能够让猴子根据该神经元的活动和奖励系统对其进行调节。
第 484 段
所以据我所知,我会说这是闭环脑机接口,也就是BCI,最早的例子。
第 485 段
摘要中写道:“通过在神经元高频放电时投放一粒食物作为强化,对麻醉猴子中央前皮层中单个神经元的活动进行了条件化。除食物强化外,通常还会提供单元放电频率的听觉和视觉反馈。”很酷,所以他们确实做成了。
第 486 段
他们做成了。这可是在1969年。
第 487 段
“经过数次训练,猴子可以使新分离细胞的活动比强化前的频率提高50%至500%。”太迷人了。
第 488 段
大脑的可塑性很强。(莱克斯笑)——于是从这里开始,实验的数量增加了。
第 489 段
是的,实验数量以及与大脑交互的工具数量都出现了爆炸式增长。我想,此外,对神经编码以及某些皮层层级和功能如何组织的理解也大幅增长。还有一篇相当具有开创性的论文,尤其是在运动解码方面,是乔治奥普洛斯在20世纪80年代发表的论文,它发现了一个叫作运动调谐曲线的东西。那么,什么是运动调谐曲线?
第 490 段
它指的是,哺乳动物(包括人类)的运动皮层中存在一些神经元,它们有一个会使自身放电的偏好方向。所以,这意味着当你想着向左、向右、向上、向下以及沿其中任何一个向量移动时,会有一组神经元增加其脉冲活动。
第 491 段
基于这一点,你可以开始思考,嗯,如果你无法识别那些基本的特征向量,你就可以做很多事情,而且实际上可以利用这些信息,从皮层中解码某人的预期动作。所以这是一篇非常非常具有开创性的论文,它表明存在某种可以提取的编码,尤其是在运动皮层中。
第 492 段
所以那里存在信号。而且如果测量大脑发出的电信号,实际上就能弄清意图是什么。
第 493 段
没错,是的,不仅是电信号,而是来自正确的那组神经元、能为你提供这些偏好方向的电信号。
第 494 段
嗯。好,那么慢慢说到Neuralink,一个有意思的问题是,对于侵入式与非侵入式脑机接口,我们目前有怎样的理解?从这一系列研究来看,紧挨着神经元放置设备有多重要?这样能为你带来什么?
第 495 段
这个问题的答案从根本上取决于你想用它来做什么,对吧?其实使用EEG和心电图ECoG可以做数量惊人的事情。它实际上不会穿透皮层或脑实质,而是把一组电极放在大脑表面。
第 496 段
所以,我个人非常感兴趣的事情,就是切实理解局部层面正在发生的活动,并且能够真正接入对这些活动的高分辨率、高保真理解。我们可以深入讨论生物物理学,但先稍微退一步,用一个类比来说明,因为类比在这里可能很有用,而且有时候电是有点难以想象的。
第 497 段
归根结底,我们进行的是由离子电流,也就是这些带电粒子的运动所介导的电记录,而对大多数人来说,这真的真的很难想象。但事实证明,大脑中发生的许多活动及其开始发生时所处的频率带宽,其实与声波以及我们正常交谈中的可听范围非常非常相似。
第 498 段
所以,这个领域通常使用的类比是,假设有一座足球场,里面正在进行比赛。如果你站在场外,也许可以根据主场观众的欢呼声和酒,大致感觉比赛进行得怎么样、球队是否领先。但你完全不知道比分是多少。
第 499 段
你完全不知道每个观众或球员正在谈论什么、彼此在说什么,下一次战术是什么,下一个进球是什么。所以你必须把麦克风放进体育场里靠近一些,然后接近声源,比如各个人的交谈。在这个具体例子中,你会希望把它放在球员围拢商议战术的位置旁边。
第 500 段
所以我认为,这很好地说明了我们所说的侵入式、微创式或植入式脑机接口,与非侵入式或非植入式脑接口有什么区别。它基本上讨论的是,你要把那个麦克风放在哪里,以及你能利用这些信息做什么。
第 501 段
那么,现在进入Neuralink所做的工作时,我们在这里谈论的读写通信,其生物物理学原理是什么?
第 502 段
是的,大脑由这些被称为神经元的特化细胞构成。它们有数十亿个,数百亿个。有时候人们会说是1000亿个,它们连接成一个复杂却动态的网络,并且在持续重塑。它们会改变自己的突触权重,而这就是我们通常所说的神经可塑性。
第 503 段
神经元也浸润在这种带电环境中,其中潜藏着许多带电分子,比如钾离子、钠离子、氯离子。而它们实际上通过离子电流促进这些不同网络之间的交流。
第 504 段
而且,当你观察神经元时,它们也有这些膜,膜上有一种非常、非常精美的蛋白质结构,叫作电压选择性离子通道,在我看来,这是大自然最出色的发明之一。从很多方面来说,如果你想想它们是什么,它们做的就是现代晶体管的工作。归根结底,晶体管无非就是一种电压门控的传导通道。
第 505 段
而大自然在其演化非常、非常早期的时候,就找到了一种拥有这种东西的方法。众所周知,有了晶体管,你就可以进行非常、非常多的计算,并实现许多我们今天能够使用的奇妙事物。所以我认为,稍微离题说一句,这是大自然想出的一项非常、非常精美的发明,就是这些电压门控离子通道。
第 506 段
我是说,我想,在生物层面,在有机体层级结构复杂性的每一个层面,都会有某些储存信息和进行计算的机制。而这只不过是其中一种方式。但用生物和化学成分来做到这一点很有意思。另外,比如当神经元,我是说,它不只是电,还有化学通信,也还有机械作用。
第 507 段
我是说,这些就像是会振动的实际物体。我是说,它们会移动——是的,它们确实会,我是说,这其中涉及许多非常、非常有意思的物理现象,你知道,有点回到我读研究生时关于超声波的研究,有一些团队,过去有一些团队,现在也仍有一些团队在研究如何利用超声波让神经元真正触发动作电位。
第 508 段
据我所知,这一现象发生的机制仍不清楚。可能只是因为你施加了某种热能,而这导致细胞以某些有意思的方式去极化。但也有这些离子通道,甚至还有这样的膜,当它们受到机械摇动时,其孔隙实际上就会打开,对吧?振动。
第 509 段
所以,其中有很多类似这些移动粒子的要素,而这同样受扩散物理学支配,对吧?粒子的运动。而且其中也有许多某种意义上很有意思的物理现象。
第 510 段
另外,更不用说,正如罗杰·彭罗斯所谈到的,这一切的量子力学效应中或许存在某种精妙的奇异性。而他确实认为,意识可能是从那里的量子力学效应中涌现出来的。所以,这里有物理学,有化学,有生物学,所有这一切都在其中发生。
第 511 段
哦,对,对。我的意思是,你可以,是的,物理学有很多层次可以深入探究。但对,归根结底,你有这些膜,上面有这些电压门控离子通道,它们选择性地让细胞外基质中的这些带电分子进出。这些神经元通常具有某种静息电位,也就是细胞内部与细胞外部之间存在电压差。
第 512 段
而当某种刺激改变了状态,使它们需要向下游网络发送信息时,你会开始看到这些不同分子进出这些通道的某种协同运作。它们也会打开,比如一旦达到某个阈值,就会有更多通道打开,直到细胞去极化并发出动作电位。
第 513 段
所以,这就是这些分子一种非常美妙的协同运作。而当我们放置一根电极,或者把它停靠在一个神经元旁边时,我们试图做的就是测量这些局部电位变化。同样,这些变化由离子的运动介导。有趣的是,正如我之前提到的,这其中涉及很多物理学。
第 514 段
而在这个电记录领域中,占主导地位的两种物理机制是扩散物理和电磁学。哪一种占主导,也就是麦克斯韦方程占主导还是固定定律占主导,取决于你的电极在哪里。如果它靠近源头,主要以电磁机制为基础;当你离得更远时,则更多以扩散为基础。
第 515 段
所以从本质上说,当你能把它停靠在旁边时,你就可以窃听那些单独的交流和局部电位变化,而你得到的信号类型就是教科书上典型的神经尖峰波形。
第 516 段
当你离得更远时,根据人们所做的一些研究,包括克里斯托夫·科赫的实验室和其他人的研究,一旦你距那个源头大约100微米,也就是大约一根人类头发的宽度,你就再也听不到那个神经元了。或者说,你无法再让系统具备足够的灵敏度,来记录那个神经元中特定的局部膜电位变化。
第 517 段
另外,为了让你对尺度有个概念,当你观察一个100微米的体素,也就是脑组织中一个100微米乘100微米乘100微米的盒子时,里面大约有40个神经元,以及不知多少条它们拥有的连接。所以那一体积的组织里有很多东西。
第 518 段
所以,一旦你处在这个范围之外,就完全没有希望检测到你可能关心的那一个特定神经元所发生的变化。
第 519 段
对,但当你在这个空间里移动时,你会听到其他神经元。所以,如果你再移动100微米,就会听到另一个群体的交流。
第 520 段
正确。
第 521 段
所以整个思路就是,你想放置尽可能多的电极,然后监听这些交流。
第 522 段
对,你想监听这些交流。而归根结底,你还想基本上让软件来完成解码工作。再稍微谈谈为什么ECOG和EEG能够工作,对吧?当出现这些局部变化时,显然并不只是这一个神经元在激活。还有许许多多其他网络始终都在激活。
第 523 段
你确实会看到这种电……比如这种带电介质的电位出现某种整体变化,而当你离得更远时,记录的就是这个。我的意思是,你仍然有某个在脑中保持稳定的参考电极,大脑本身就是一个具有电活性的器官,而你看到的是某种组合、聚合的动作电位变化,然后你可以将其捕捉到,对吧?这是一种变化慢得多的信号。
第 524 段
但确实存在这些典型的振荡和波,比如伽马波、贝塔波。比如当你睡觉时,这些可以被检测到,因为大脑存在某种同步的全局效应,你可以检测到它。
第 525 段
而且,我的意思是,这背后的物理学,我是说,如果我们真的想沿着这个兔子洞深入下去,那么关于为什么当你离源头更远时,扩散物理会在某个时刻占据主导地位,这里面有很多事情。它就是一种带电介质。
第 526 段
所以,这类似于当电磁方式在大气中或等离子体这样的带电介质中传播时,会发生这种奇怪的屏蔽现象,实际上会随着你远离它而进一步衰减信号。所以,对,你会看到,比如如果你真正、真正深入研究信号随距离的衰减,就会开始看到一开始是某种哪里的平方,然后呈指数下降。
第 527 段
而那就是从电磁磁学占主导转变为扩散物理占主导的拐点。
第 528 段
但还是那句话,对于电极而言,你需要理解的生物物理学没有那么深,因为无论你把它放在哪里,你监听的都是一小群局部神经元。
第 529 段
正确,对。所以一旦你穿入大脑,就可以说是进入竞技场了。
第 530 段
而且那里有很多神经元。
第 531 段
[DJ] 它们有许许多多。
第 532 段
但话说回来,神经科学有一整个领域在研究不同的分组、竞技场座位的不同区域通常各自负责什么,不过这个比喻可能到这里就失效了,因为竞技场里的座位并没有那么有组织。
第 533 段
另外,它们中的大多数都处于沉默状态。它们其实没做多少事情,或者它们的活动是,你知道,你必须恰好用正确的那组刺激去击中它。
第 534 段
所以它们通常很安静。
第 535 段
它们通常非常安静。就像,我的意思是,类似于暗能量和暗物质,还有暗神经元。它们到底都在做什么?当你放置这些电极时,同样,比如在这个100微米的体积内,你有大约40个神经元。比如为什么你看不到40个神经元?为什么只能看到少数几个?那里发生了什么?
第 536 段
嗯,它们大多很安静,但比如当它们开口时,我觉得它们会说些深刻得不得了的话。我喜欢这样去理解。无论如何,在我们进一步放大之前,先缩小来看。那么Neuralink是如何工作的?从手术到植入物,再到信号和解码过程,以及人类能够使用植入物实际影响外部世界,这整个过程是怎样的?
第 537 段
而我问的这一切,其背景是Neuralink在今年1月刚刚完成了一个具有重大历史意义的里程碑,把Neuralink植入物植入了第一个人类诺兰体内。关于他的体验,有很多东西可以谈,因为他能够描述其中所有的细微之处、美妙之处,以及整个相关体验中迷人而复杂的一切。
第 538 段
但在技术层面,Neuralink是如何工作的?
第 539 段
是的,所以我们正在构建的技术有3个主要组成部分。一个是设备,也就是实际记录这些神经信号活动的东西。我们称它为 N1 植入体或 The Link。我们还有一台手术机器人,实际负责植入这些极其微小的导线,我们称之为“线丝”,它们比人的头发还细。
第 540 段
一旦一切都完成手术处理,你就会得到这些神经信号,也就是这些从大脑中传出的脉冲神经元信号,而你需要某种软件来解码用户打算用它做什么。
第 541 段
所以有一个叫作 Neuralink 应用程序或 B1 应用程序的东西负责这种转换,它运行一个非常非常简单的机器学习模型,解码这些作为神经信号的输入,然后将其转换为一组输出,让我们的参与者,也就是第1位参与者诺兰,能够在这个上控制光标。
第 542 段
这是通过无线方式完成的吗?
第 543 段
这是通过无线方式完成的。所以我们的植入体实际上由2部分组成。The Link 带有这些被称为“线丝”的柔性微小导线,沿其长度方向分布着多个电极。它们只会被插入皮质层,在人脑的运动皮质区域,这一层大约有3到5毫米厚。那里就是某种运动意图所在的位置。
第 544 段
我们有64根这样的线丝,每根线丝在3到4毫米的跨度内有16个电极,彼此间隔200微米。所以你实际上可以沿着插入深度进行记录。
第 545 段
根据该信号,我们制造了一种定制集成电路,也就是 ASIC,它会放大你正在记录的神经信号,然后将其数字化,之后再通过某种机制检测是否出现了一个值得关注的事件,也就是脉冲事件,并决定是否通过蓝牙将它发送到运行 Neuralink 应用程序的外部设备,无论是手机还是电脑。
第 546 段
所以设备上已经在进行信号处理,只是为了判断这是不是一个值得关注的事件。所以除了人脑之外,内部设备上也有一些计算能力?
第 547 段
是的,所以它会进行信号处理,以便大幅压缩你所记录的信号量。我们总共有1000个电极,以略低于20千赫兹的频率采样,每个样本为10比特。所以就是200兆比特。那是由1000个通道同时进行神经记录并传送到芯片的数据。这个数据量相当大。
第 548 段
而且有技术可以无线发送这些数据,但要在大脑这种散热条件非常、非常受限的环境中做到这一点,就必须进行一定程度的压缩,只发送你需要的、值得关注的数据;在这个特定案例中,对于运动解码而言,就是有没有出现尖峰。然后再利用这些数据解码预期的光标移动。
第 549 段
所以植入物本身会处理数据,利用我们的尖峰检测算法判断有没有出现尖峰,然后将其发送出去,也就是打包后通过蓝牙发送到外部设备;该设备上有用于解码的模型,好,根据这些尖峰输入,Noland 是想向上、向下、向左、向右,还是单击、右键单击,或者做其他什么操作?
第 550 段
这一切确实非常有意思,但我们还是聚焦于 N1 植入物本身,也就是大脑里的那个东西。我正在看它的图片,有一个外壳,有一个充电呼叫,所以我们还没有谈到充电,这很有意思。电池、电力电子器件、天线。然后是信号处理电子器件。我想知道是否还能进行更多种类的信号处理。这是另一个问题。然后是丝线本身,底部带有外壳。那么也许可以问问充电,有一个外部充电设备。
第 551 段
嗯,对,有一个外部充电设备。所以,是的,植入物的第二部分,也就是这些丝线,再说一次,只有最后的 3 到 5 毫米是真正刺入大脑皮层的。其余部分,实际上,大部分体积都由电池,也就是可充电电池占据。它大约有一枚 25 美分硬币那么大。我这里实际上有一个设备,你想看看的话可以看一下。
第 552 段
这是它的柔性威胁部件。然后这是植入物。它大约有一枚美国 25 美分硬币那么大。厚度大约是 9 毫米。所以基本上,做完颅骨切除术和迪雷克托米手术、插入丝线之后,你所造出的这个孔,也就是这个颅骨切除处,会由它取代。所以基本上,那个东西会塞住这个孔,你可以拧入这些自钻式颅骨螺钉,将它固定到位。
第 553 段
最终,盖上皮瓣之后,只有大约 2 到 3 毫米。显然,那是从植入物顶部到螺钉所在位置的过渡部分。那就是你会有的那一点小凸起。
第 554 段
那些丝线看起来非常小。太不可思议了。真的太不可思议了。而且你说得对,实际体积中的大部分都是电池。是的,这比我意识到的要小得多。
第 555 段
它们也就是,丝线本身相当结实。
第 556 段
它们看起来很结实。
第 557 段
而且丝线本身的末端还有一个非常有意思的特征,叫作环。机器人正是通过这种机制与这个像头发一样细小的结构对接并操纵它。
第 558 段
而且它们非常小,那么一根丝线有多宽?
第 559 段
对,丝线的宽度从 16 微米开始,然后逐渐扩展到大约 84 微米。普通人类头发的宽度约为 80 到 100 微米。
第 560 段
这东西太惊人了。这东西太惊人了。
第 561 段
是的,大部分体积都由电池,也就是可充电锂离子电芯占据。充电通过感应充电完成,这实际上非常常见。大多数手机都有这种功能。最大的区别是,通常当你有一部手机,想把它放在充电板上充电时,你其实并不在意它会变得多热,而对我们来说,这一点很重要。
第 562 段
有非常严格的规定,而且有充分的理由不能让周围组织的温度升高 2 摄氏度。因此,这里面实际上融入了大量创新,使这个植入物能够充电,而不会达到那个温度阈值。甚至还有一些细小之处,比如你看到的这个充电线圈,以及所谓的铁氧体屏蔽层,对吧?
第 563 段
所以如果没有那个铁氧体屏蔽层,在进行谐振感应充电时,最终会出现这样的情况:电池本身是一个金属罐,而外部充电器会形成这些涡流,导致发热,实际上也会降低充电效率。所以这个铁氧体屏蔽层的作用,是把磁力线集中并引导到远离电池的地方,然后使其环绕实际缠在电池周围的线圈。
第 564 段
这里有许多确实非常有意思的设计来实现这一点,我是说,你是在把一台计算机集成到一个复杂的生物系统中。
第 565 段
是的,这里有很多创新。我想说,让这成为可能的部分原因就是可穿戴设备领域的创新。现在有很多非常、非常强大、小巧、低功耗的微控制器、温度传感器,或者其他各种传感器和电力电子器件。很多创新确实都体现在充电线圈的设计、如何封装它,以及如何实现充电而又不真正超过那个温度上限;对于市面上的其他设备来说,这并不是一种约束。
第 566 段
那么我们来谈谈丝线本身,那些非常、非常、非常小的东西。那么一共有多少根?你提到有 1,000 个电极。有多少根丝线,电极和丝线又有什么关系?
第 567 段
对,目前这个版本的设备有 64 根丝线,每根丝线上有 16 个电极,总计 1,024 个电极,既能记录,也能刺激。而丝线基本上就是这种采用聚合物绝缘的导线。金属导体是一种由 ti、plat、金、plate、ti 构成的提拉米苏蛋糕式结构。它们是非常、非常细小的导线。宽度为 2 微米,也就是一米的 2/1000000。
第 568 段
太疯狂了,我正在看的这个东西有聚合物绝缘层、有导电材料,而且末端还有 16 个电极。
第 569 段
每一根这样的丝线上都有。
第 570 段
对,每一根这样的丝线上都有。
第 571 段
没错。
第 572 段
16 个,每一根上都有。
第 573 段
你用肉眼是看不到的。
第 574 段
而且,我是说,说一句显而易见的话,或者也许是说给那些只听音频的人听,它们是柔性的。
第 575 段
是的,是的,这也是一个对我们极其重要的要素。所以这些丝线中的每一根,正如我提到的,宽度是 16 微米,然后逐渐扩展到 84 微米,但厚度不到 5 微米。而厚度主要由底部的聚酰胺、这条金属线路,然后再加一层聚酰胺构成。所以是 2 微米的聚酰胺、400 纳米的这种金属堆叠层,以及 2 微米的聚酰胺夹在一起,以保护它免受周围这个 37 摄氏度盐水袋环境的影响。
第 576 段
那么,有哪些,也许你能谈谈这里材料设计的一些有趣之处吗?比如,对于那些对这类东西一无所知的人来说,要设计出这样的东西,并且能够制造出这样的东西,需要些什么?
第 577 段
是的,所以我们所做的材料选择并不,我不认为它特别独特。过去有其他实验室,现在也有其他实验室在研究类似的材料堆叠。围绕这些我们所称的微电极,与其他一些更传统的颅内神经接口设备相比,其寿命和可靠性如何,存在一个根本性问题,而且仍然需要得到解答。
第 578 段
也就是刺入皮层、更加刚性的那种,比如犹他阵列。有这种 4×4 毫米左右的硅针柱,其末端有裸露的记录位点。而这算是理查德·诺曼早在 1997 年提出的创新。它叫犹他阵列,因为他当时在犹他大学。
第 579 段
那么犹他阵列是什么样子的?所以它是一种刚性的————是的,所以我们其实可以查一下。
第 580 段
哦。
第 581 段
对。(Lex笑)对,所以它是一张针床。有————(笑)对。好,你继续,抱歉。
第 582 段
所以那些是刚性的————刚性的,对。你还真没开玩笑。
第 583 段
针柱的尺寸和数量各不相同,从64到128不等。它的最尖端是一个裸露电极,实际负责记录神经信号。另一个值得注意的有趣之处是,与Neuralink的细线不同,Neuralink细线上沿深度方向分布着记录电极,也就是裸露的氧化铱记录位点,而这个只在单一深度进行记录。所以这些犹他阵列的针柱长度可以在0. 5毫米到1。
第 584 段
5毫米之间。而且它们也有倾斜式设计。因此你可以将它插入不同深度,但这是另一个重大区别。然后,我是说,最主要的关键区别在于它没有有源电子器件。
第 585 段
这些只是电极,然后还有你所看到的一束导线,随后它实际上会从开颅切口引出,并带有这个端口,你可以把任何外部电子设备连接到上面。他们正在研发或者已经有了无线遥测设备,但它仍然需要一个穿过皮肤的端口,而这实际上是该系统发生感染的最大故障模式之一。
第 586 段
柔性细线会带来哪些挑战?比如说,在机器人方面,R1植入这些细线,这项任务有多困难?
第 587 段
是的,正如你提到的,它们非常、非常难以用手操控。你之前看到的这些犹他阵列,实际上是由神经外科医生将其定位在他们想要的位置附近,然后用一把气动锤将它们推进去。所以这是一个相当简单的过程,而且它们很容易操控。但这些薄膜阵列非常、非常微小且柔韧。
第 588 段
所以它们非常难以操控。因此,我们才制造了一整台机器人来做这件事。我们制造机器人的原因还有其他一些,其中之一是,我们最终希望它能帮助数百万、数百万人,他们都能从中受益。而世界上根本没有那么多神经外科医生。我们希望机器人实际上能够完成手术中的很大一部分。
第 589 段
不过,是的,这台机器人是我们正在研发的另一整类产品。它本质上是一个多轴龙门系统,配有一个专用机器人头,其中包含所有光学器件以及这样一种针头回缩机制;它通过细线上的这个环状结构来操控这些细线。
第 590 段
所以细线本身已经有一个环状结构,可以通过它抓住细线?
第 591 段
对,对。
第 592 段
好的。这太有意思了。你提到了光学器件,所以有一台机器人——R1。目前,实际上是由人类在颅骨上开一个孔。
第 593 段
嗯哼。
第 594 段
然后在那之后,有一个计算机视觉组件会寻找避开血管的路径。接着你通过环抓住每一根细线,并把它放到特定位置以避开血管。还会选择植入深度,所有这些?
第 595 段
对。所以会控制每一个,比如植入位置的3D几何参数?
第 596 段
对。这台机器人的独特之处在于,它不是由外科医生辅助或人类辅助的。它是一台半自动或自动机器人。显然,其中也有人类参与,比如在设定目标位置时。你始终可以让它避开你看到的那些主要血管。不过,我的意思是,我们希望达到这样一种程度:点击一下,它就在几分钟内完成手术。
第 597 段
所以计算机视觉组件会找出很好的目标、候选位置,然后人类大致批准它们,机器人就会执行,它是一次处理一根细线,还是一次处理一——它一次处理一根细线,而我们实际上也正在研究一次处理多根细线的方法。没有什么会阻碍这一点。你可以配备多个这样的啮合机制,但目前是逐根进行。
第 598 段
而且我们目前仍会进行相当多的这种验证,以确保它确实被植入了。如果植入了,有多深?它实际上是否与编程设定相符,诸如此类?
第 599 段
而实际的电极是一个会处于不同深度的位置,就像,我的意思是,差异非常小,但确实有差异。
第 600 段
[DJ] 对,对。
第 601 段
所以正如你提到的,这背后有一些考量。比如它能获得更加多样的信号。
第 602 段
是的,我的意思是,我们尽量把它们全都放在距表面大约3或4毫米的位置,因为电极的跨度,也就是我们当前这个版本中的那16个电极,横跨大约3毫米。所以我们希望让所有这些电极都进入大脑。
第 603 段
这太有意思了。好的,这里有无数个问题。如果我们能具体放大来看电极,那么依你看,每个电极正在监听多少个神经元?
第 604 段
是的,正如我之前提到的,每个电极可以记录0到40个神经元。但从战术上讲,我们最多只看到大约2到3个。而且你实际上可以通过尖峰的形状辨别它来自哪个神经元。
第 605 段
[Lex] 哦,很酷。
第 606 段
我刚才提到了我们的尖峰检测算法。它叫作BOSS算法,也就是缓冲在线尖峰分类器。
第 607 段
不错。
第 608 段
最终,它实际上会输出6个独特的值,它们大致是这些波形的振幅,比如负向波峰、中间波峰、正向波峰,以及这些波峰出现的时间。由此,你可以进行一种统计概率估算:那是尖峰吗?不是尖峰吗?然后基于这一点,你还可以判断,哦,那个尖峰看起来和这个尖峰不同。那一定来自另一个神经元。
第 609 段
好的,所以这是一个很不错的信号处理步骤,由此你就能更准确地预测是否存在尖峰。
第 610 段
是的。
第 611 段
尤其是在这种可能有多个神经元同时尖叫的情境下。而这也使你能够在(听不清)中更好地压缩数据。好的。
第 612 段
而且说明一下,我的意思是,实验室会做一种叫作尖峰分类的工作。通常,在获得这些宽带的、完全数字化的信号后,你再运行一系列不同的算法来逐一分离,而对我们来说,所有这些都是在设备上完成的。
第 613 段
在设备上。
第 614 段
在一个功耗非常低的定制ASIC数字处理单元中。
第 615 段
[Lex] 散热限制极其严格?
第 616 段
散热限制极其严格,而且从信号输入到为你提供输出,处理时间不到1微秒,这是非常、非常短的时间。
第 617 段
哦,是的,所以延迟必须非常短。
第 618 段
对。
第 619 段
哦,哇。哦,那真是烦死人了。
第 620 段
是的,延迟是你必须处理的一个非常、非常大的问题。目前,最大的延迟来源是蓝牙,也就是它们被打包的方式,而我们会在15毫秒内弯折它们。
第 621 段
哦,有意思,上面写着通信限制。在所使用的协议方面,有没有一些潜在的创新空间?
第 622 段
当然有。
第 623 段
好的。
第 624 段
是的,蓝牙肯定不是我们最终想要采用的无线通信协议。
第 625 段
所以才有 N1 和 R1。我想那会增加——NxRx。
第 626 段
对,那就是通信协议,因为我想,蓝牙允许你在比实际所需更远的距离上进行通信,所以可以把距离缩短很多。
第 627 段
是的,嗯,选择蓝牙的主要动机是,我是说,所有东西都有蓝牙。
第 628 段
好吧,你可以和任何设备通信。
第 629 段
互操作性绝对至关重要,尤其是在这个早期阶段。而且从很多方面来说,如果你能访问手机或电脑,你就能做任何事。
第 630 段
嗯,退一步,再实际看一下你刚才为诺兰提到的同一套流程,会很有意思。那么,这整个过程是什么样的?从寻找并筛选一个人,到手术,再到他第一次能够使用这个东西?
第 631 段
我们有一个所谓的患者登记名册,人们可以注册,以了解更多最新消息。诺兰就是通过这个渠道申请的。流程是,一旦申请提交进来,其中会包含一些病历,然后根据他们在医学上是否符合资格来判断,他们需要满足很多不同的纳入和排除标准。之后,他们会与 Neuralink 的工作人员进行预筛选面谈。
第 632 段
到了某个阶段,我们也会去他们家里进行 BCI 居家审核,因为这套 N1 系统完全无线,而拥有它最具革命性的方面之一,就是你可以在家中使用它。比如,你实际上不必去实验室或诊所,连接到这些无法带回家的专用设备上。
第 633 段
所以,这是我们设计系统时希望牢记的关键要素之一,比如希望人们愿意每天在舒适的家中使用它。因此,在参与沟通以及进行 BCI 居家审核时,我们要了解的一部分内容,就是他们的情况,以及他们还使用哪些辅助技术。
第 634 段
而且我们也应该退一步说明一下,据估计,美国有 180,000 名四肢瘫痪者,每年另有 18,000 人遭受导致瘫痪的脊髓损伤。所以这些人在生活中面临着很多挑战,包括无障碍方面,也包括去做许多我们日常认为理所当然的事情。
第 635 段
而这项初步研究的目标之一,是让他们拥有某种数字自主能力,使他们能够独自仅凭意念与数字设备互动,也就是你们所说的心灵感应。也就是数字心灵感应,让四肢瘫痪者能够以我们一直讨论的所有方式与数字设备通信。
第 636 段
控制鼠标光标,足以完成各种事情,包括玩游戏、发推文,以及诸如此类的事。而且有很多人因为发生在他们身上的事情,连生活中的基本事务都很困难。
第 637 段
是的,我是说,运动对我们的存在而言太根本了。我是说,就连说话也涉及嘴巴、嘴唇、喉部的运动。而如果没有这些,致残程度会极其严重。我们可以帮助许许多多的人。而且我是说,尤其是当你开始去看其他形式的运动障碍时,它们不仅源于脊髓损伤,也源于 ALS、MS,甚至中风和/或仅仅是衰老,对吧?这些会导致你丧失部分行动能力、部分独立性,致残程度极其严重。
第 638 段
而所有这些都是帮助人们、帮助减轻痛苦、帮助改善生活质量的机会。但你提到的每一件事都是一个独立的小谜题,需要像 Neuralink 设备这样的装置具备越来越高的能力。所以你们首先专注的是一个非常美的词:心灵感应。也就是能够用意念,以无线方式与数字设备进行交流。你能确切解释一下我们说的是什么吗?
第 639 段
是的,我是说,确实就是这样。我是说,我认为,如果你能够控制光标、能够点击,并且能够使用电脑或手机,我是说,整个世界就向你敞开了。而且我是说,我想,“心灵感应”这个词,如果你把它从定义上理解为,不使用我们拥有的某些身体机能,比如声音,就能把信息从我的大脑传递到你的大脑。
第 640 段
但这里有意思的是,我认为有一点并不是显而易见的,那就是它究竟如何运作。所以,要移动光标,至少有几种方法。一种是想象自己或许正用手移动鼠标,或者你也可以像 Noland 谈到的那样,想象用意念移动光标。
第 641 段
但这里仿佛存在一个很迷人的认知步骤,因为你必须使用大脑,而且必须学会如何使用大脑。你还得以某种方式动态地摸索出来。因为如果它奏效了,你就会奖励自己。我是说,这里有一个步骤,这真是个很迷人的步骤,因为你必须让大脑开始以正确的方式放电。而你通过想象来做到这一点。就像先装作能做到,直到真的做到。
第 642 段
(笑)然后突然之间,它产生了正确类型的信号,而这个信号如果被正确解码,就能产生相应的效果。然后它周围还有类似噪声的东西,你必须把所有这些都弄清楚。但在人这一侧,你必须做的就是想象光标正在移动。
第 643 段
是的,他说这是在使用原力。
第 644 段
原力。我是说,它真的能起作用,这难道不会让你觉得很迷人吗?对我来说,就像是,我靠,它居然真的有效。你居然可以用意念移动光标。
第 645 段
在你学习使用那个东西的同时,那个东西也在了解你。比如,我们的模型会不断更新权重,来判断:“哦,如果有人在思考这种复杂形式的、类似尖峰的模式,那实际上意味着要这样做,对吧?”
第 646 段
所以机器在了解人,人也在了解机器。因此,信号处理,也就是解码步骤,具有适应性。然后还有 Noland 这个人的适应。就像如果你给我一个新鼠标,我移动它,我会很快了解它的灵敏度,于是我会学着把它移动得更慢。然后还有其他类型的信号漂移,以及他们必须适应的所有那类情况。所以双方都在相互适应。
第 647 段
[DJ] 没错。
第 648 段
这对双方的软件来说都是个很迷人的挑战,双方的软件,人的软件和————有机的和无机的。不管怎样,很抱歉如此无礼地打断。所以 Noland 经历了这项筛选,并以优异成绩通过了。也就是说,包括那是一个适合 BCI 的家,所有这些条件都在内。那么手术植入的过程是怎样的,他第一次使用这个系统又是在什么时候?
第 649 段
整个端到端过程,我们称为从患者进入到患者离开,耗时在 2 到 4 小时之间。具体到 Noland 的情况,大约是 3 个半小时。在机器人实际执行植入之前,还有许多步骤,对吧?比如诱导麻醉,我们还会进行术中 CT 成像,以确保在正确的位置钻孔。这也会事先规划好。
第 650 段
像 Nolan 这样的人会接受功能性磁共振成像,然后他们可以想着活动自己的手。显然,由于他们的伤势,这实际上不会产生任何预期的输出。但当你想象移动手指时,大脑中亮起的部位,与实际移动手指时亮起的是同一个部位。
第 651 段
而这正是我们能够实际判断该把电极丝放在哪里的方法之一,因为我们想进入运动皮层中一个叫作手结区的区域。并且尽可能密集地放置我们的电极丝。所以,是的,我们会进行术中 CT 成像,以确认并再次核查颅骨切开位置。
第 652 段
然后外科医生进来,完成皮肤切口、颅骨切开,也就是在颅骨上钻孔之类的操作,之后是大脑的许多不同层次。有一层叫作硬脑膜,是包围大脑的一个非常非常厚的层,它实际上会在一个叫作粥样斑块切除术的过程中被切除。随后,这就暴露出了大脑中的软脑膜,你想要在那里进行插入。
第 653 段
等进行到大约 1 到 1 个半小时的时候,机器人进场,完成它的工作,包括确定目标位置、插入电极丝。这需要 20 到 40 分钟。具体到 Noland 的情况,是略低于,或者说略高于 30 分钟。之后,外科医生进来。
第 654 段
另外还有几个步骤,比如实际置入硬脑膜替代层,以保护电极丝和大脑。然后,是的,把植入物拧进去,接着复位皮瓣,然后缝合,之后你就可以离开了。
第 655 段
那么 Noland 醒来时是什么样的?恢复过程如何,他第一次能够使用它是在什么时候?
第 656 段
实际上就在手术之后,大约术后 1 小时,当他正在醒来时,我们确实启动了设备,确认我们正在记录神经信号,而且我们确实注意到了几个他实际上可以调节的信号。我所说的调节,是指他可以想着握紧拳头,而你能看到尖峰消失又出现。(Lex 笑)——太棒了。
第 657 段
而且那是立刻发生的,对吧?术后立刻就在恢复室里。
第 658 段
这得有多酷?
第 659 段
是的。
第 660 段
那是一个人。我是说,那对你来说是什么感受?这个设备和一个人,一段宏大旅程的第一步?我是说,这是一个历史性时刻。哪怕只是那个尖峰,仅仅是能够调节它。
第 661 段
显然,正如你提到的,此前也有其他先驱参与了这些开创性的 BCI 研究性早期可行性试验。所以很显然,我们是站在巨人的肩膀上。我们并不是第一批真正把电极放入人脑的人。但我是说,就在手术临近时,我确实完全睡不着。这是你第一次在一个全新的环境中工作。
第 662 段
基于我们的台架测试或临床前研发研究,我们对这种机制、这些线、植入操作,以及所有那些东西都很有信心,认为它们非常安全,而且显然已经可以在人类身上进行这项操作了,但仍然存在很多未知的未知,比如针究竟能不能插进去?我是说,我们带了大约40根针,以防它们断掉,结果最后只用了一根。
第 663 段
但我的意思是,那种程度就是完全未知,对吧?因为这是一个非常、非常不同的环境。而且,我是说,这正是我们首先要进行临床试验的原因,以便能够检验这些东西。所以当时极度紧张,在手术前的许多个夜晚都无法入睡,尤其是手术前一天,而且那是一台一大早进行的手术。我们早上7点就开始了。
第 664 段
到了大约10:30,一切都完成了。但我的意思是,第一次看到那一幕,首先,只是如释重负,因为这东西正在做它应该做的事。其次,我是说,对诺兰和他的家人充满了无比的感激,还有许多其他已经申请、我们已经交谈过以及将要交谈的人,我是说,他们都是各处真正的先驱者。我有点把他们称作神经宇航员,或者“神经航行者”。
第 665 段
这些了不起的,就像在60年代一样,对吧?就像这些了不起的先驱者,对吧?向外探索未知。在这个案例中,是向内。但我对他们愿意参与并发挥作用怀有难以置信的感激之情。而且这是一段我们正在共同踏上的旅程。但同时,我觉得那只是,那是一个非常、非常重要的里程碑,而我们的工作才刚刚开始。
第 666 段
所以也充满了某种期待:好吧,接下来需要发生什么?需要发生哪些一系列事件,才能让这件事对诺兰和我们双方都值得?
第 667 段
继续谈一下这一点,衷心祝贺你和团队达成了这个里程碑。我知道还有很多工作要做,但看到这一幕真的令人振奋。这是希望的源泉。这是迈出的第一个重大步伐,是帮助数十万人的机会,然后也许能在未来为数百万人拓展人类心智能力的可能范围。所以这真的令人振奋。
第 668 段
所以,可以说所有机会都在我们前方,而看到以安全且有效的方式做到这一点,真的很令人开心。作为一名工程师,看着其他工程师齐心协力完成一件史诗般的大事,那太棒了。衷心祝贺。
第 669 段
谢谢,谢谢。是的,没有团队我不可能做到。还有,是的,我是说,这也是我告诉团队的另一件事,就是对未来怀有这种极其强烈的乐观感。我的意思是,不用说,这对公司来说是一个非常重要的时刻,也希望对外界许多我们能够帮助的人来说同样如此。
第 670 段
那么说到挑战,Neuralink 发布了一篇博客文章,描述了一些细线发生回缩。因此,以每秒比特数衡量的性能起初下降了,但最终又恢复了。而性能如何恢复的整个故事非常有意思。这绝对是我会和 Bliss 以及 Noland 讨论的事情。但总体来说,你能谈谈这整个过程吗?性能是如何恢复的,以及细线回缩和移动在技术层面是怎么回事?
第 671 段
最主要的结论是,性能最终恢复了,而且实际上比以前更好了。就在上周,他又一次打破了世界纪录,达到了 8.5 BPS,所以我的意思是,他一直在全力推进,而且一直在进步。
第 672 段
[Lex] 他之前创下的纪录是 8。
第 673 段
没错。
第 674 段
他说是 8.5。
第 675 段
是的,此前人类的世界纪录是 4.6。
第 676 段
是的。
第 677 段
所以几乎翻了一番。他的目标是努力达到 10,这大致相当于 Neuralink 员工用手操作鼠标时的中位水平。所以正在接近了。
第 678 段
所以,是的,性能恢复了。
第 679 段
是的,比以前更好。所以,BCI 团队为了恢复那种性能经历了什么,这本身就是一个故事。实际上主要是在信号处理方面。正如我提到的,我们当时算是在观察电极输出的这些尖峰。发生的情况是,在手术后大约 4 周时,我们注意到这些细线仅仅从大脑中出来了。
第 680 段
而我们最初注意到这件事的方式,显然是,嗯,我想 Noland 是第一个注意到自己表现正在下降的人。而且我想当时我们也在尝试做很多不同的实验、不同的算法、不同种类的 UI/UX。所以性能出现波动是意料之中的,但我们确实看到了一种持续的下降。
第 681 段
另外,我们衡量电极健康状况或者它们是否位于大脑中的方式,是测量电极的阻抗。所以我们会看某种界面,也就是他们所说的 Randall 电路,电表面与介质之间的电容和电阻。如果它以某种剧烈的方式发生变化,我们就会得到一些迹象。
第 682 段
或者,如果你在那些通道上没有看到尖峰,就会有一些迹象表明那里正在发生某些事情。而我们注意到的是,通过观察那些阻抗图和尖峰率图,再加上我们有那些沿深度进行记录的电极,你会看到某种移动,表明其余的正在被拉出来。
第 683 段
而这显然会对模型端产生影响,因为如果进入模型的输入数量正在变化,因为它们变少了,那么那个模型就需要更新,对吧?但仍然存在信号,而且正如我提到的,类似于即使你把信号放在大脑表面,或者放得更远,比如放在颅骨之外,你仍然会看到一些有用的信号。
第 684 段
我们开始观察的不仅仅是通过我提到的这个 BOSS 算法得到的尖峰出现情况,还开始观察对 Noland 来说值得关注、并且他能够调制的频段功率本身。
第 685 段
所以,当我们算是改变植入物的算法,让它不只是向你提供 BOSS 输出,还提供这些尖峰频段功率输出后,这算是帮助我们用新的输入集合找到模型,而这最终真正让我们恢复了性能。
第 686 段
就……而言,而且显然,我们最终想要的、也是我们正在努力实现的,是找出让那些细线尽可能长时间保持完好的方法,从而让更多通道进入模型。这显然是团队目前正在着手开展的第一要务,要弄清楚如何防止这种情况发生。
第 687 段
我还要说的是,正如我提到的,这是我们有史以来第一次把这些细线放入人脑,而人脑,仅就尺寸而言,是猴脑或羊脑的 10 倍。它就是一个非常、非常不同的环境。它的移动要多得多。实际上,在我们为 Noland 做手术时,它的移动程度比我们预期的要大得多。
第 688 段
而且它就是一个与我们所习惯的非常、非常不同的环境。这就是我们开展临床试验的原因,对吧?我们希望尽早而不是晚些时候发现其中一些问题和失效模式。所以从很多方面来说,它为我们提供了大量数据和信息,使我们能够解决这个问题。而这是 Neuralink 极其擅长的事情。
第 689 段
一旦我们有了一组明确的目标和工程问题,我们在许许多多学科中拥有大量人才,能够聚在一起,非常、非常迅速地解决问题。
第 690 段
但听起来,这里一个引人入胜的挑战是,系统和解码端需要能够适应不同的时间尺度。所以,无论是细线的移动,还是人脑软件上信号漂移的不同方面、某些东西发生变化,比如 Noland 谈到可以纠正的光标漂移,以及如何做到这一点所涉及的一整套 UX 挑战。所以听起来,适应能力就像是一项必须通过工程设计纳入其中的基本属性。
第 691 段
是的,而且我的意思是,我认为,我的意思是,作为一家公司,我们的垂直整合程度极高。我们在自己的微加工厂里制造这些薄膜阵列。
第 692 段
是的,就像你说的,是内部制造的。这篇博客文章里的这一整段相当硬核。“构建上述技术绝非易事。”这里还有一堆链接,我建议大家点击查看。“我们在内部建立了微加工能力,以快速生产构成我们电极细线的各种薄膜阵列迭代版本。我们制造了一台定制的飞秒激光铣床,以微米级精度制造组件。”我想有一条与此相关的推文。
第 693 段
那是一个我们可以深入讨论的完整话题。
第 694 段
是的,好吧,我们在这里看到的是什么?这个东西?
第 695 段
是的。“所以,在不到 1 分钟内,我们定制的飞秒激光铣床会在针尖上切割出这种几何形状。”所以我们看到的是这根形状奇怪的针。“针尖的宽度只有 10 到 12 微米,仅略大于红细胞的直径。较小的尺寸使威胁能够被植入,同时将对大脑皮层的损伤降至最低。”好,那么这种几何形状有什么有趣之处?所以我们看到的就是针的这种几何形状。
第 696 段
这就是与细线上的环相接合的针。所以,就是它们把针穿过线环,然后把线环从硅背衬上剥离下来。然后,这就是会插入组织的部分,接着它会抽出来,把细线留在里面。而这个类似缺口、或者我们过去称为鲨鱼齿的东西,才是真正抓住线环的部分。然后它被设计成这样,使得你抽出时,它会脱离线环。
第 697 段
所以是机器人在控制这根针?
第 698 段
对,所以它实际上装在一根套管里。基本上,机器人有很多光学设备,用来寻找线环的位置。实际上有一道405纳米的光,会使聚酰胺发出荧光,这样你就能定位线环的位置。
第 699 段
所以线环会亮起来?
第 700 段
对,对,会的。这是一个微米级精度的过程。
第 701 段
要做到这件事,机器人有什么有意思的地方?这太疯狂了。机器人能达到这种精度,太疯狂了。
第 702 段
对,我们的机器人相当重。我们目前这个版本是这样。它,我是说,就像一块巨大的花岗岩板,重约1吨,因为它需要对振动、环境振动保持敏感。然后,当机器头以那样的速度移动时,需要进行大量某种运动控制,以确保能够达到那种精度。还有很多光学设备,会把那里放大。我们正在研发下一代机器人,它会更轻、更容易运输。我是说,移动这台机器人可是一项壮举。
第 703 段
而且就目前而言,在这项特定任务上,它远胜于人类外科医生。
第 704 段
当然,我是说,更不用说你真的试着把针穿过针线包里的一个线环了,我是说,这就像,我们说的是人类头发直径的一小部分。这些东西是看不见的。
第 705 段
所以继续读这一段:“我们开发了新型硬件和软件测试系统,比如加速寿命测试架和模拟手术环境,”这非常酷,“以对我们的技术进行压力测试,并验证其稳健性。我们进行了许多次手术演练,以完善我们的流程,并让这些操作成为第二本能。”这非常酷。“我们使用替代物练习手术,并配备在模拟空间或工程空间中所需的全部硬件和器械。这有助于我们快速测试和测量。”所以这里有类似替代物的东西。
第 706 段
对,这个替代模型其实超级酷。所以,这里有一个根据在巴罗拍摄的影像3D打印出来的颅骨,还有这种水凝胶混合物,算是一种合成聚合物,实际上能模拟大脑的机械性质。它还有这个人的血管系统。
第 707 段
所以基本上,我们在这里谈的是,为制作这个成套替代模型已经投入了大量工作,差不多就是要找到这些不同合成聚合物的正确浓度,从而获得适合针在插入过程中运动状态的恰当稠度。但在实际进行手术之前,我们已经用这个人的,也就是诺兰的生理结构和大脑,练习了很多很多遍这台手术。
第 708 段
所以每一步,每一步?
第 709 段
每一步,对。比如某个人站在哪里?我是说,你现在看到的是这张照片。这是在我们的办公室里,是机器人工程空间的这个角落,我们搭建了这样一个模拟空间,看起来和他们在实际手术期间会经历的、所有工作人员会经历的情形一模一样。
第 710 段
所以我的意思是,这有点像任何舞蹈彩排,你确切知道自己在什么时间点要站在哪里,然后用将要接受你手术的那个人完全一致的解剖结构,一遍又一遍又一遍地练习。后来到了这样一个程度:我们进行开颅术时,很多工程师都会说:“哦,这看起来非常眼熟。我们以前见过。”
第 711 段
对。天哪,通过一遍又一遍又一遍地做同一件事,你可以获得智慧。这有点像“Jira 梦见寿司”那种东西,因为奥运选手也会想象奥运会,然后当你真的到场时,感觉就很轻松。感觉就像平常的任何一天。赢得金牌甚至感觉很无聊,因为你已经想象过这么多遍、练习过这么多遍,没有什么能干扰你。很无聊。你赢得了金牌,很无聊。而他们谈到的感受大多只是如释重负,可能是因为他们不必再去想象它了。
第 712 段
对,心智进行想象的力量,我是说,有一整个领域研究肌肉记忆位于小脑的什么地方。对,太不可思议了。
第 713 段
我认为这里正适合提出人们可能会有的那个大问题:我们怎么知道你所描述的这一切的每一个方面都是安全的?
第 714 段
归根结底,黄金标准是观察组织。你对组织造成了什么样的创伤?这是否与你可能观察到的任何行为异常相关?这就是我们用来沟通以下问题的语言:将某种东西插入大脑是否安全,以及你可能造成什么类型的创伤。所以我们实际上有一整个部门,也就是病理学部门,负责观察这些组织切片。
第 715 段
完成这项工作涉及许多步骤。一旦你启动了以特定终点为目标的研究,到了某个时候就必须对动物实施安乐死,然后进行尸检,以采集大脑组织样本。你用福尔马林固定它们,然后对它们进行大体检查、切片,并观察每一张切片,只为看看存在什么样的反应,或者是否没有反应。
第 716 段
所以,这算是FDA使用的那种语言,也是我们用来评估插入机制以及这些细丝在各个不同时间点的安全性的语言。既包括急性期,也就是0至3个月之间的任何时间,也包括3个月以后。
第 717 段
所以这些算是必须达到的极高安全标准的具体细节。
第 718 段
没错。
第 719 段
FDA会对此进行监管,但总体而言,这本身就是一个非常高的标准。而这一切的每一个方面,包括手术,我记得马修·麦克杜格尔提到过,这里的标准,可以说,该怎么委婉地表达呢?比我们可能习以为常的其他一些手术更高。所以这里所有手术相关事项的标准都极高。
第 720 段
非常高。我的意思是,这是一个受到高度、高度监管的环境,主管机构会严格审查每一种上市的医疗器械。而且我认为这是件好事。制定这些高标准是好事,而我们也努力坚守极高的标准,以便了解我们正在打造的这些创新型新兴技术和新技术究竟会造成什么样的损伤,如果有的话。到目前为止,这些细丝几乎没有引发免疫反应,这令我们印象极其深刻。
第 721 段
说到这里,你曾兴奋地跟我谈起组织学,以及你们能够分享的一些图像。你能向我解释一下我们看到的是什么吗?
第 722 段
对,所以你看到的是一张经过染色的组织图像。这是从一只植入了7个月的动物身上取得的组织切片。所以算是一个慢性期时间点。你会看到所有这些不同的颜色,而每种颜色都表示特定的细胞类型。紫色和粉色值得尊敬地是星形胶质细胞和小胶质细胞。它们是胶质细胞的类型。
第 723 段
不过,人们可能没有意识到的另一件事是,你的大脑并不只是由一锅神经元和轴突组成的。还有其他细胞,比如胶质细胞,实际上有点像胶水,而且在组织受到任何创伤或损伤时也会作出反应。
第 724 段
这里棕色的是神经元吗?
第 725 段
棕色的是神经元。
第 726 段
现代神经元。
第 727 段
所以你看到的是,在这张宏观图像中,你会看到这些被白色圆圈一样的东西标出的插入位置。当你放大其中一个位置时,就会看到这些细丝。然后在这个具体案例中,我想我们看到的是大约16根伸入画面内部的导线。
第 728 段
而这里不可思议的是,那些棕色结构或者棕色圆形或椭圆形物体,也就是神经元,实际上正触碰并紧靠着细丝。所以这表明,这次插入基本上造成了零创伤。而对于这些神经接口,也就是你插入的这些微电极来说,这是最常见的失效模式之一。
第 729 段
所以当你插入这些细丝时,比如犹他阵列,它会导致插入部位周围的神经元死亡,因为你插入的是一个异物,对吧?而这有点会通过小胶质细胞和星形胶质细胞引发这种免疫反应。它们会在周围形成这样一层保护层。
第 730 段
哦,你不仅杀死了神经元细胞,还形成了这层保护层,随后它基本上会阻碍你记录神经信号,因为你离自己试图记录的神经元越来越远。而这就是最大的失效模式。在这个具体例子中,那个内部,按比例尺来看大约是50微米。神经元看起来就是被它吸引了。
第 731 段
(Lex笑)——所以肯定没有创伤。顺便说一句,这张图像太美了。所以棕色的是神经元。不知为什么,我移不开视线。真的很酷。
第 732 段
是的,而且这些东西的呈现方式,比如说,我的意思是,你的组织通常不会有这些漂亮的颜色。这是多重染色,使用不同的蛋白质将这些组织染成不同的颜色。我们采用了一套非常标准的染色技术,包括 HG、EB1、NeuN 和 GFAP。
第 733 段
所以,如果你看下一张图,这也算是说明了第2点,因为你可以提出一种论点,而且一开始,当我们看到上一张图时,我们说:“哦,比如说,这些细线是不是只是漂浮着?这里发生了什么?我们看到的真的是正确的东西吗?”所以我们又做了一次染色,而且这些全都是内部完成的,采用的是这种 batons 三色染色法,其中蓝色显示的是这些胶原层。
第 734 段
所以这个蓝色基本上是,比如说,你不希望植入物细线周围出现蓝色,因为这意味着发生了某种瘢痕形成。而你所看到的是,如果观察单根细线,就看不到任何蓝色,这意味着这些植入的细线所造成的创伤完全不存在,或者说极其轻微,轻微到无法检测出来。
第 735 段
所以,这大概就是采用这种柔性细线的一大好处。
第 736 段
是的,所以我们认为,这主要归因于细线的尺寸及其柔韧性。此外,R1 会避开血管系统,因此我们不会扰动血管,也不会对血管造成损伤,并且不会破坏任何血脑屏障,这基本上使免疫反应受到了抑制。
第 737 段
不过,这也很好地说明了这些东西的尺寸。所以这是细线的尖端。
第 738 段
是的,那些是神经元。
第 739 段
而且它们是神经元。这是正在监听的细线。电极是如何布置的?
第 740 段
是的,所以你现在看到的并不是电极本身。那些是导线。因此,每根导线的宽度大概应该是 2 微米。所以我们现在看到的是冠状切片。也就是说,我们看到的是组织的某个切片。因此,越往深处,细线逐渐变细的程度显然就会越来越小。不过,是的,关键基本上在于,插入部位周围就只是有细胞,能看到这一点简直不可思议。我从未见过这样的情况。
第 741 段
取出植入物有多容易、多安全?
第 742 段
是的,所以这取决于时间。手术后的前3个月左右,会发生大量某种组织重塑。类似于你被割伤时,显然在最初几周会重新开始,或者根据伤口的大小,形成瘢痕组织,对吧?这些组织会收缩,然后最后变成痂,你可以把痂揭掉。
第 743 段
同样的事情也会在大脑中发生,而且那是一个非常动态的环境。在瘢痕组织或新生膜,或者说形成的新生膜出现之前,把它们直接拉出来相当容易。而且这个过程造成的创伤很小。一旦瘢痕组织形成,我们认为 Noland 的情况也是如此,正是它目前把那些细线固定住了。所以从那以后,我们没有再看到任何移动。因此它们相当稳定。
第 744 段
要把这些细线完全取出来会变得更困难。所以,我们目前移除设备的方法是剪断细线,保持组织完整,然后拧开并取出植入物。接着,那个孔会用另一个 Neuralink,或者只是用某种以 PEEK 为基础的塑料盖封住。
第 745 段
把这些细线永远留在那里没问题吗?
第 746 段
是的,我们认为没问题。我们做过将它们留在那里的研究,而我们曾经最担心的一个问题是,它们会不会迁移,会不会到达不该去的地方?我们还没有看到这种情况。再说一遍,一旦瘢痕组织形成,它们就会被固定在原位。而且我还应该说,当我们说升级时,我们并不只是在谈理论。我们实际上已经升级过很多很多次了。
第 747 段
我们的大多数猴子,或者说非人灵长类动物,即 NHP,都已经升级过了。你看到过玩意念乒乓球的 Pager,从2年前起就装着最新版本的设备,而且看起来非常快乐、健康,也很胖。
第 748 段
那么,面向未来设计的升级程序是什么样的?比如说对 Noland 而言,升级会是什么样子?基本上就是你提到的那种方式吗?有没有办法从内部升级设备,把它拆开,某种程度上保留外壳,再升级内部组件?
第 749 段
是的,这里有几种不同的情况。对于 Noland,如果我们要进行升级,就必须根据那些细线在里面的固定或结疤情况,剪断细线或者把细线取出来。如果连同硬脑膜替代物一起把它们移除,大脑仍然是完整的,因此可以用更新后的植入物组件重新插入不同的细线。
第 750 段
关于未来可升级系统会是什么样子,我们还在考虑另外几种不同的方式。其中一种是,目前我们会移除硬脑膜,也就是保护大脑的这种厚层,但它实际上正是促进瘢痕组织形成的东西。所以通常来说,一个很好的经验法则是尽量让自然结构保持原样,不要过多扰动它。
第 751 段
因此,我们正在研究让细线穿过硬脑膜插入的方法,这会带来一组不同的挑战,比如它是一层相当厚的组织,那么究竟怎样才能在不折断针的情况下穿透它?所以我们正在为此研究不同的针具设计,以及某种环扣接合方式。另外一些最大的挑战是,从光学角度看,在白光照明下它相当不透明。
第 752 段
那么,你要如何仍然避开我们所拥有的这个最大优势,也就是避开血管系统?你要如何透过它成像?你究竟仍然要如何调节这一点?我们正在研究其他成像技术,以实现这一点。但我们的假设是,而且根据我们掌握的一些早期证据,穿过硬脑膜进行插入所造成的瘢痕会很少,使得它们随着时间推移更容易取出。
第 753 段
我们还在研究的另一件事,将会是植入物架构的一次根本性改变。目前,它是一个整体式的单一植入物,附带粘接在一起的细线。所以你实际上无法把它分开,但你可以设想一种由两部分组成的植入物。底部是插入的细线,其中包含芯片,也许还有无线电装置和某种电源。
第 754 段
然后还有另一个植入物,承担更多繁重的计算负载,并配有更大的电池。其中一个可以放在硬脑膜下,另一个可以放在硬脑膜上,就像颅骨的塞子。它们可以相互通信,但你想升级的是计算机,而不是细线。如果你想升级它,只需进入那里,取下螺丝,然后装入下一个版本。这也是一种非常非常简单的手术。
第 755 段
比如你在皮肤上切一个切口,把这个滑进去,拧上螺丝。大概可以在10分钟内完成。
第 756 段
所以这在某种程度上可以让你重复使用那些细线。
第 757 段
[DJ] 没错。
第 758 段
所以,我是说,这自然会引出一个问题:增加细线数量并实现规模扩展的路径是什么?这是优先事项吗?其中的技术挑战是什么?
第 759 段
是的,这是一个优先事项。所以,对于下一版本的植入物,我们希望改进的关键指标是通道数量,也就是记录越来越多的神经元。我们有一条实现路径,可以从目前的1,000个增加到3,000个,希望如此,甚至可能在今年年底达到6,000个。然后到明年年底,我们希望达到更多,也就是16,000个。
第 760 段
哇。
第 761 段
这方面有几个限制。其中一个显然是能否通过光刻来印制这些导线。正如我提到的,其宽度和间距都是2微米。显然,有些芯片所采用的分辨率远比这类分辨率先进,而且我们已经把完成这项工作所需的一些工具引入了内部。因此,走线会变得更窄,因为你必须让更多导线接入芯片。
第 762 段
随着通道越来越多,芯片的能耗也不能线性增加。因此,为了降低功耗,电路和架构以及电路设计拓扑方面都有很多创新。你还需要考虑,如果有了所有这些尖峰信号,要如何把它们发送到最终应用?所以你需要考虑那里的带宽限制,以及信号处理方面可能需要的创新。
第 763 段
从物理层面看,最大的挑战之一将会是接口。出问题的总是接口。也就是把这种薄膜阵列与电子器件粘接起来。它开始变成密度非常非常高的互连。那么要如何对它进行表征?近年来,3D 集成领域出现了很多我们可以利用的创新。我们面临的最大挑战之一是形成这种气密屏障,对吧?
第 764 段
因为我们所处的是一个极其严酷的环境——大脑。那么,要如何保护它,一方面防止,没错,比如大脑试图杀死你的电子器件,另一方面也防止你的电子器件把你不希望出现的东西泄漏到大脑中。形成这种气密屏障将是一个非常非常大的挑战,而我认为我们实际上非常适合应对这个挑战。
第 765 段
你们要怎么测试?用来模拟那种严酷程度的开发环境是什么样的?
第 766 段
是的,所以这里的加速寿命测试仪本质上就是一个缸中之脑。它确实是一个由……构成的容器,而且再说一次,就这种特定类型的测试而言,实际上你的大脑就是盐水。你还可以加入另一组化学物质,比如活性氧物种,它们会作用于这类界面,试图引发反应,将其拉开。
第 767 段
但你也可以仅仅通过提高温度来加快这些界面的老化速度。所以温度每升高10摄氏度,基本上就是把时间加速到2倍。而温度能提高多少是有限度的,因为到某个点,就会出现其他一些非线性动力学现象,导致形成其他有害气体,而这在实际环境中并不现实。
第 768 段
所以我们的做法是把ALT测试舱的温度提高20摄氏度,这会使老化速度提高到4倍。所以从本质上说,在ALT测试舱里1天,相当于日历时间里的4天。然后我们观察植入物是否仍然完好,包括电极丝以及————还有运行情况等等?
第 769 段
还有运行情况等等。显然,它与大脑的环境并不完全相同,因为大脑有机械性的东西,还有其他更多会攻击它的生物群体。但至少对于外壳及其强度而言,这是一个很好的测试环境,测试环境。而且我的意思是,我们有一些植入物,也就是当前版本的植入物,已经在里面放了,我是说,接近2年半,这相当于10年。它们看起来没问题。
第 770 段
所以很有意思的是,这种灼烧,也就是说,近似度很高的环境是温盐水,热盐水是一个很好的测试环境。是的,顺便说一下,我正在喝LMNT,它基本上就是盐水,这让我有点……它不像大脑那样拥有计算能力,但也许就其他特性而言,它相当相似,而我正在饮用它。
第 771 段
是的,你还必须把它调到正确的 pH 值。(笑)——然后意识就会涌现。是的,不。
第 772 段
顺便说一下,我们的外壳还有一点也很有意思:如果你看看我们的植入物,它并不是常见的医疗植入物外观,后者通常封装在一个经过激光焊接的钛罐中。我们使用一种叫作PCTFE的聚合物,即聚三氟氯乙烯,它其实常用于泡罩包装。所以当你有一片药,试图把药片按出来时,会有那种塑料薄膜。这就是这种材料。
第 773 段
除了我们,其实从来没有人用过这个。我们想这么做的原因是它具有电磁透明性。所以,当我们谈到电磁感应充电时,如果使用钛金属外壳,通常你想做这样的事情,就必须设置一个蓝宝石窗口,而要扩大这种工艺的规模非常、非常困难。
第 774 段
所以你们在这里的各个方面都进行了大量迭代。材料、软件——整个、整个全套东西。
第 775 段
所以,好吧。你提到了扩大规模。作为扩大规模的一种方式,有没有可能使用多个 Neuralink 设备?植入多个 Neuralink 设备?
第 776 段
这就是目标,这就是目标。是的,我们的猴子植入过2个 Neuralink,每个大脑半球各1个。然后,我们还在研究这样一种可能性:在运动皮层植入1个,在视觉皮层植入1个,再在其他某个皮层植入1个。
第 777 段
所以是用1个 Neuralink 设备专注于某项特定功能。
第 778 段
对。
第 779 段
我是说,我在想计算端是否可以进行某种程度的定制。所以对于运动皮层——当然可以。这就是目标。我们在 Neuralink 谈的是构建一种连接大脑的通用神经接口。在市场推广方面,我们从战略上也是这样处理的。在监管方面也是如此,也就是说,嘿,你看,我们有机器人,而机器人可以进入皮层的任何部位。
第 780 段
目前,我们专注于运动皮层,使用的是当前版本的 N1,它专门用于运动解码任务。但归根结底,那里也提供了某种通用计算能力。不过通常来说,如果你真的想针对功耗和效率进行某种极致优化,就确实需要落实到某种专门功能,对吧?
第 781 段
但我们要说的是,嘿,你们现在已经习惯了这种机器人植入技术,而这经历了许多、许多年向 FDA 展示数据并与其沟通的过程。此外,我们也要在内部说服自己这是安全的。而现在的区别是,如果我们进入大脑的其他部位,比如我们有意将其作为第2款产品的视觉皮层,那么显然,那是一个完全不同的环境。
第 782 段
皮层的布局非常、非常不同。它将更多地以刺激为重点,而不是记录,只是以某种方式产生视觉感知。但最终,我们使用的是相同的薄膜阵列技术。我们使用的是相同的机器人植入技术。我们使用的是相同的封装技术。现在,讨论更多地集中在:有哪些差异,以及这些差异对安全性和有效性意味着什么?
第 783 段
你把它称为第2款产品的方式,在我看来既好笑又了不起。那款产品是为盲人恢复视力。那么,你能谈谈刺激视觉皮层吗?我是说,那里的可能性简直令人难以置信,能够把这份礼物还给没有视力的人,甚至恢复其中的任何一方面。你能谈谈其中的挑战吗,这里有若干项挑战。
第 784 段
哦,很多。
第 785 段
其中一项就像你说的,是从记录转向刺激。你们既感到兴奋、又看到挑战的任何一个方面都可以谈谈。
第 786 段
是的,我想我先从这一点说起:实际上,多年来,我们一直能够通过薄膜阵列以及其他电子器件进行刺激。实际上,我们已经展示了这方面的一些能力,用于使脊髓中的肢体重新活动。显然,对于目前的 EFS 研究,我们已经在硬件层面禁用了这项功能,所以我们希望把它作为一段独立的、独立的旅程来展开。
第 787 段
显然,有许多、许多不同的方式可以向大脑写入信息。我们采用的方式是通过电,也就是让电流通过,并以某种方式使其真正改变局部环境,从而可以说是人为地促使附近区域的神经元发生某种去极化。具体到视觉,我们的视觉系统如何工作,这既已得到充分理解。
第 788 段
我是说,任何与大脑有关的东西,都有一些方面已经得到充分理解。但归根结底,我们其实什么都不知道。不过视觉系统的工作方式是:光子照射到你的眼睛,而你的眼睛里有一些被称为感光细胞的特化细胞,它们会把光子能量转化为电信号。然后,电信号被投射到你的后脑,也就是视觉皮层。
第 789 段
它实际上会经过一个叫作 LGN 的丘脑系统,然后由该系统将它投射出去。接着,在视觉皮层中,有视觉区一,也就是 V1,然后还有许多其他更高级的处理层,比如 V2、V3。而这里其实存在一些有趣的相似之处。
第 790 段
当你研究这些卷积神经网络的行为时,比如网络的不同层在检测什么,它们首先检测的是这类边缘,然后检测一些更自然的曲线,接着开始检测物体,对吧?大脑中也会发生某种类似的事情。很多相关研究都受到了这一点的启发,而且看到其中的一些关联也一直令人颇为兴奋。
第 791 段
但从那里开始,认知在何处产生、颜色在哪里编码等问题,人们就是不太了解,缺乏根本性的认识。因此,就为盲人恢复视力而言,失明有许多不同的形式。美国实际上有100万人,100万人在法律上被认定为盲人。那意味着,比如在某种视力测试中,某个分数低于……
第 792 段
我想大概是,如果你在20英尺的距离才能看清某样东西,而正常人在200英尺的距离就能看清,比如说,如果你的情况比这更差,你就在法律上被认定为盲人。
第 793 段
所以这意味着你无法有效地正常活动。
第 794 段
对。
第 795 段
在这个世界上利用视力。
第 796 段
是的,比如用它在周围环境中行动。是的,失明有不同的形式。有些形式的失明是你的视网膜,也就是这些感光细胞出现了一定程度的退化,而我所描述的其余视觉处理系统则完好无损。对于这类人,你也许不需要把电极插入视觉皮层。
第 797 段
实际上,你可以制造视网膜假体设备,直接替代那些已经退化的视网膜细胞的功能。有许多公司都在研究这个方向。但这只是很小的一部分。尽管具有重要意义,但在法律上被认定为盲人的人群中,这部分人的占比更小。
第 798 段
如果沿着这套回路的任何位置存在损伤,无论是在视神经中,还是就在 LGN 回路中,抑或这条回路中存在任何中断,那对你都不会奏效。而因为你的生物机制没有产生视觉感知,你真正需要促使这种视觉感知发生的位置,是通过在你后脑的视觉皮层中放置电极。
第 799 段
而它的工作方式是,你会有一个外部摄像头,无论是像 GoPro 这样不太复杂的设备,还是 Meta 正在研发的某种可穿戴式 Ray-Ban 类眼镜,用来捕捉一个场景,对吧?然后,那个场景会被转换成一组电脉冲或刺激脉冲,你会通过这些薄膜阵列在视觉皮层中激活它们。
第 800 段
而通过像一支协调一致的管弦乐团那样演奏这些刺激模式,你可以产生所谓的光幻视,也就是这种偏黄白色的光点,你只需按压眼睛也能产生这种光点。实际上,你可以通过刺激视觉皮层来产生那些感知。
第 801 段
而真正的关键是要有许多个这样的光点,并让这些感知,也就是这些光幻视,尽可能小,这样你就能开始将它们区分开来,就像它们是屏幕上的单个像素一样,对吧?所以,如果你有非常非常多个这样的光点,那么从长远来看,你或许能够真正获得自然的视觉。
第 802 段
但在短期到或许中期,至少能够让物体检测算法在你的眼镜、术前处理单元上运行,然后至少能够看到物体的边缘,这样你就不会撞到东西。
第 803 段
太不可思议了。这真的太不可思议了。所以基本上,你会不断添加像素,而你的大脑会开始弄清楚那些像素意味着什么。是的,而且在各个方面,都有不同种类的助手参与信号处理。
第 804 段
是的。其实,这件事,所以有几点。一点是,显然,如果你先天失明,从大脑的工作方式来看,尤其是在幼年时期,神经可塑性其实无非就是你的大脑以及大脑的不同部分在争夺有限的地盘。
第 805 段
[Lex](笑)是的。
第 806 段
而且我的意思是,很快很快,你就会看到这样的案例,你知道有些人,我的意思是,你也听说过,盲人的听觉或其他某些感官会更加敏锐。其原因在于,未被使用的皮层会被皮层的这些不同部分占据。所以对于这类人,我的意思是,我想他们现在将不得不把他们其他感官的某些部分映射成他们所谓的视觉。
第 807 段
但显然,这会是一种与以前非常非常不同的意识体验。所以我认为,这是一个值得注意的有趣限定条件。另一件同样需要强调的重要事情是,就我们所能看到的波长而言,我们目前受到自身生物构造的限制。我们眼睛能够看到的可见光波长范围非常非常小。
第 808 段
但当你在这个脑机接口系统中使用外部摄像头时,就不受这个限制了。你可以使用红外线,可以使用紫外线,可以使用任何其他你想看到的光谱。至于这是否会对应某种奇怪的意识体验,我不知道。但我经常跟人们谈到,Neuralink 的目标是超越我们生物机能的限制。这差不多就是我的意思。
第 809 段
而且,如果你能够控制这种原始信号,当我们使用视觉时,我们接收到的是光子,并没有对它进行太多处理。如果你能够控制那个信号,也许就可以进行某种处理。也许你可以提前进行物体检测。
第 810 段
[DJ] 对。
第 811 段
你在进行某种预处理,而这方面有很多可能性可以探索。所以它不只是增强某种热成像之类的东西,还包括进行某种有趣的处理。
第 812 段
没错,对。我的意思是,我关于视觉系统如何运作的理论也是,我是说,世界上发生着太多事情,有大量光子进入你的眼睛,而其中一些预处理步骤究竟在哪里发生,并不清楚。但我是说,实际上从根本角度来看,我认为实在是太多了,我们所处的这个现实,如果它是现实的话,是如此——有如此多的数据。
第 813 段
而且我认为,人类其实就是无法真正吃下足够多的东西来处理所有这些信息。所以确实会发生某种过滤,无论它发生在视网膜中,还是发生在视觉皮层的不同层级中。并不清楚。
第 814 段
但我有时想到的类比是,如果你的大脑是一台 CCD 相机,而世界上的所有信息都是太阳,那么当你真的尝试用 CCD 相机直视太阳时,它只会让传感器饱和,对吧?因为那是极其庞大的能量。所以你所做的是最终加上这些滤镜,对吧?只是为了稍微缩窄传向你并被捕获的信息。
第 815 段
而且我认为,像我们的经历,或者像异丙酚这类药物,那种麻醉药,或者迷幻药,它们所做的就是在某种程度上换掉这些滤镜,装上新的滤镜,或移除旧的滤镜,并以某种方式控制我们的意识体验。
第 816 段
对,伙计,不是想偏离话题,但我刚在亚马孙丛林里服用了很高剂量的死藤水。所以没错,这是一个很好的思考方式。你在换上不同的体验,而 Neuralink 能够控制这一点,最初主要是为了改善功能,不是为了娱乐或享受,而是————对,恢复失去的功能。
第 817 段
嗯,恢复失去的功能。而且在那里,尤其是当某项功能完全丧失时,任何帮助都是巨大的。你会在自己的大脑中植入 Neuralink 设备吗?
第 818 段
当然。我的意思是,也许不是现在,但当然会。
第 819 段
当实现什么样的能力后,你会开始真正感到好奇,甚至变得有点坐立不安,比如看着别人接受植入时会嫉妒他们?
第 820 段
对,我是说,我认为,我是说,即使是我们早期的参与者,如果他们开始做一些我做不到的事情——我认为他们有可能达到 15、20,甚至可能 100 BPS,对吧?从根本上说,没有什么会阻止我们实现那种性能。我的意思是,我肯定会嫉妒他们能做到这一点。
第 821 段
我应该说,看着 Noland,我确实有点嫉妒,因为他玩得太开心了,而且这看起来是一种非常轻松的电子游戏玩法。
第 822 段
对。所以我的意思是,有件事有时也很难充分意识到,那就是他在说话的同时做着这些事。我的意思是,这是多任务处理,对吧?所以显然,这显然需要投入大量认知,但就像我们说话时会移动双手一样,这些事情就像是在进行多任务处理。我的意思是,他能做到这一点。而据我所知,用其他辅助技术是做不到的。
第 823 段
如果你显然使用的是眼动追踪设备,你就会高度专注于自己正试图做的那件事。而如果你使用语音控制,我是说,比如你说了些别的东西,对,你就没法使用它了。
第 824 段
对,其中多任务处理的部分真的很有意思。所以这不只是主要任务的 BPS,还包括多个任务的并行化。如果你测量整个人体的 BPS,也就是说,如果你一边说话,一边用意念做一件事,同时还四处张望,我是说,就会有很多瘫痪现象可能正在发生。
第 825 段
对。但我是说,我认为在某个时刻,对他而言,比如如果他想真正达到那些高水平的 BPS,确实需要全神贯注,对吧?而那是一套独立的神经回路,是一个巨大的谜团。比如注意力如何运作,还有,你知道吗?
第 826 段
对,注意力,比如认知负荷,我读过很多关于人们执行两项任务的文献。比如你有一项主要任务和一项次要任务。而次要任务是一个干扰源。这会如何影响主要任务的表现?而且根据任务的不同,有很多有趣的,我是说,这是一个有趣的计算设备,对吧?而且我认为————至少可以这么说。
第 827 段
可以从一切事物中获得许多新颖的见解。我的意思是,我个人很惊讶,Noland 能够在说话、同时还感到紧张的情况下,对光标进行如此不可思议的控制,因为他像我们所有人一样在说话,如果你对着镜头说话,就会紧张。所以所有这些因素都在发挥作用,而他依然能够取得高水平的表现。令人惊讶。我的意思是,这一切真的都很了不起。而且我认为,在真正深入研究过这件事之后,我有点想要 Neuralink 了。
第 828 段
(笑)排队去吧。
第 829 段
而且,安全方面也要排队。嗯,我们应该说明,登记名单是面向四肢瘫痪以及诸如此类情况的人,所以————没错。
第 830 段
会给人们另设一条队伍。他们只是好奇,就像我一样。那么现在,Noland,也就是 P1 患者,正在参与持续进行的 PRIME 研究,对 P2、P3、P4、P5 的高层愿景是什么?以及如何将范围扩大到其他能够体验这种植入物的人类?
第 831 段
对,我是说,我们研究的首要目标首先是达到安全性终点。就是了解这个设备以及植入过程的安全性。同时也要了解它的有效性,以及它可能对潜在用户的生活产生的影响。而且仅仅因为你患有四肢瘫痪,并不意味着你的情况就和另一位患有四肢瘫痪的人相同。
第 832 段
差异极其、极其巨大。我们也希望了解,我们的技术如何不仅服务于这些人中非常小的一部分,还能服务于更广泛的人群,并且能够获得反馈,真正为他们打造最好的产品。
第 833 段
显然,我们也有一些目标,而早期可行性研究的主要目的,是从每一位参与者身上学习,在我们开展所谓的关键性研究之前改进设备、改进手术。关键性研究是一项规模大得多的试验,开始考察各项终点的统计显著性;这是之后将设备推向市场之前所必需的。这就是它在美国的运作方式。
第 834 段
总体而言,世界各地也都是如此。你要遵循的就是这个流程。所以,我们的目标其实就是通过 Noland、P2、P3 以及未来的参与者等人,了解我们设备的哪些方面需要改进。如果人们最后说:“我真的不喜欢它只能持续 6 小时这一点。我希望能使用这台电脑,比如说 24 小时。”
第 835 段
我的意思是,这是一项用户需求和用户要求,而我们只有通过与他们交流才能发现这一点。
第 836 段
所以在关键性研究之前,会有一种基于个人体验的快速创新。你会从不同的个人那里了解他们如何使用它,既包括光标控制、信号以及所有这类方面的高分辨率细节,也包括生活体验。
第 837 段
对,对,所以既有硬件变更,也有固件更新。甚至在 Noland 经历了那次类似恢复的事件后,他现在也已经更新到了新的固件。这类似于你的手机会一直获得新固件更新,用于安全补丁、任何新功能、用户界面,对吧?而我们的植入物也能做到这一点。
第 838 段
它不是那种静态的一次性设备,只能做它声称自己能做的事情。我的意思是,这类似于 Tesla。你可以通过无线方式更新固件,然后就有了全新的用户界面。还有所有这些花哨功能,以及各方面的最新改进,对吧?我们所说的通用平台,指的就是这个。
第 839 段
对,Noland 正在使用的那个应用程序真的很酷,里面有校准之类的各种功能。然后还有更新。你只要点击一下,就能获得更新。你们还在关注哪些其他未来能力?你提到了视觉。这一点很令人着迷。那么,加速打字或说话之类的能力呢?还有别的吗?
第 840 段
对,这些仍然属于运动项目的范畴。大体来说,我们有两个项目。我们有运动项目,也有视觉项目。运动项目目前聚焦于数字自由。你很容易就能猜到,如果你能控制数字空间中的 2D 光标,就可以移动物理空间中的任何东西。
第 841 段
所以,机械臂、轮椅、你的环境,甚至实际上,比如说,无论是通过手机还是直接连接到那些接口,也就是连接到那些机器。因此,我们正在研究如何扩展这类能力,即使对 Noland 也是如此。
第 842 段
这需要与 FDA 沟通,并就安全数据进行某种展示,比如,如果有机械臂或轮椅,我们要能保证参与者不会意外伤到自己,对吧?在数字领域移动东西和在物理空间中移动东西非常不同,在物理空间中,你确实有可能对参与者造成伤害。所以我们目前正在处理这件事。
第 843 段
言语确实涉及大脑的不同区域。言语假体非常、非常迷人,而且学术界其实已经开展了许多非常了不起的工作。Sergei Stavisky(加州大学戴维斯分校)、Jaimie Henderson,以及已故的 Krishna Shenoy(斯坦福大学),都为改进言语神经假体做了数量惊人的工作。
第 844 段
而那些研究实际上更多着眼于运动皮层中控制这些局部发音器官的部分。即使只是无声地做出说那个词的口型,或想象说话,你也能够捕捉到那些信号。至于更复杂的高级处理区域,比如布罗卡区或韦尼克区,从所有这些运作方式的底层机制来看,它们仍然是非常、非常大的谜团。
第 845 段
不过,是的,我是说,我认为 Neuralink 的活动目标是以某种方式理解这些事情,并提供一个平台和工具,以便能够理解并研究它们。
第 846 段
这就到了我提那些嗑大麻者式问题的时候了。你认为我们能够开始深入了解思想之类的东西吗?所以言语就像你说的那样,包含肌肉成分。它涉及发出声音的行为。但那些内在的东西,比如认知呢?比如低层次思想和高层次思想。你认为我们会开始注意到某种能够被捕捉的信号吗?它们可以被理解,也许还可以被用来与外部世界互动。
第 847 段
从某些方面来说,我想这开始有点触及意识的心脏问题。而且,我是说,一方面,所有这些归根结底都是一组电信号,从那里开始,也许是它本身赋予了你认知或意义,或者不知怎么的,人类心智是一台极其惊人的讲故事机器。所以我们正在告诉自己,并欺骗自己相信这里存在某种有趣的意义。
第 848 段
不过,我是说,我当然认为 BCI,而说到底,BCI 确实是一套帮助你从局部以及更广泛的意义上研究底层机制的工具。至于是否存在某些有趣的电信号模式,意味着你正在想这个,而不是……以及你是否能够从许许多多组数据中学习,把其中一些信号关联起来,从而能够读心,或者不能。
第 849 段
我不确定。我当然不会排除这种可能性,但我认为单靠 BCI 大概做不到。可能还需要额外的一套工具和框架。此外,归根结底,意识的心脏问题也植根于这样一个哲学问题:这一切有什么意义?我们存在的本质是什么?心智从这个复杂网络的哪里涌现出来?
第 850 段
是啊,主观体验是如何仅仅从一连串脉冲、电脉冲中涌现出来的?
第 851 段
是的,是的,我是说,我们确实把 BCI 和我们正在构建的东西看作理解心智、大脑的工具。唯一重要的问题。其实,关于开始形成其中某些可能独一无二的体验需要什么,确实存在某种生物学上的存在性证明。如果你真的观察我们每个人的大脑,会发现有两个半球。
第 852 段
有左侧大脑,也有右侧大脑。而且,我是说,除非你存在其他一些状况,否则你通常不会感觉自己有左腿或右腿。你只会感觉像是有一条腿,对吧?那么那里发生了什么,对吧?如果你真的观察两个半球,会发现有一种在某种程度上表征两者的结构,叫作胼胝体,据认为它有大约2亿至3亿个连接或轴突。
第 853 段
所以,这是否意味着这就是我们创造某种心智融合所需的接口和电极数量,或者由此产生某种你能体验到的新型意识体验。但没错,我确实认为,我们所有人身上都有一种颇为有趣的存在性证明。
第 854 段
而这个阈值目前尚不明确。
第 855 段
哦,是的,这些事情,这个领域的一切都是推测,对吧?
第 856 段
然后你会,你会不断地感到惊喜。你能想象这样一个世界吗:数百万人,比如数千万人、数亿人,脑中装着一个 Neuralink 设备,或者多个 Neuralink 设备四处走动?
第 857 段
能。首先,比如从全球来看,有很多人正遭受运动障碍和视觉缺陷。我是说,那是数千万人,甚至可能数亿人。因此,仅凭这一点,我认为我们就能利用这类技术带来许多益处,并有潜力做很多好事。而一旦开始进入某种神经,比如精神医学应用,抑郁、焦虑、饥饿或肥胖,对吧?比如情绪、食欲控制,我是说,这就开始变得与每个人非常切身相关。
第 858 段
更不用说,地球上的大多数人都有智能手机。一旦 BCI 开始与智能手机竞争,成为同数字世界互动的首选方式,那也会变成一件有趣的事情。
第 859 段
哦,是的,我是说,是的。这甚至还没谈到那一步,对吧?我是说,几乎,我是说,整个世界都能从这类东西中受益。然后,是的,比如,如果我们谈的是我们与机器、甚至与自己交互的下一代方式,我认为 BCI 在许多方面都能发挥作用。我还谈过的一些事情是,我确实认为,看到 80亿人带着 Neuralink 四处走动,存在一种现实的可能性。
第 860 段
好的,非常感谢你推动这件事继续向前。我期待那项令人兴奋的功能。
第 861 段
谢谢你邀请我。
第 862 段
感谢收听这场与 DJ Seo 的对话。现在,亲爱的朋友们,接下来是 Neuralink 的神经外科负责人 Matthew MacDougall。你最初是什么时候开始对人脑着迷的?
第 863 段
从来都是。自我记事起,我就一直对人脑感兴趣。我是说,我小时候爱思考,也有点不合群。你坐在那里,用自己那小小的青春期大脑思考世界上最重要的东西是什么。
第 864 段
而我得出的、最终趋同的答案是:你所能设想的一切,也就是那些人类理应关心的重要事物,实际上都包含在头骨之中。既包括对它们的感知,也包括它们的相对价值。而我们所有问题的解决方案,以及我们所有的问题,也全都包含在头骨之中。
第 865 段
如果我们更了解它是如何运作的,了解大脑如何编码信息、产生欲望,以及产生痛苦和折磨,我们就能为此做更多事情。想想人类历史上所有真正伟大的胜利。再想想所有真正可怕的悲剧。想想大屠杀,想想任何一座充满人类故事的监狱,所有这些问题归根结底都是神经化学问题。
第 866 段
所以,如果你能稍微控制一下它,你就为人们提供了做得更好的选择。按照我对历史的理解,人们在拥有更好的工具后,最终往往会做得更好——这里要加一个巨大的星号。但我认为,给予人们更多选择、更多工具,是一项有趣、值得且崇高的追求。
第 867 段
是的,这是一个审视人类历史的迷人角度。你只要想象所有这些神经生物学机制:斯大林、希特勒、所有这些人、成吉思汗,他们都只是有一个大脑,只是一堆神经元,就像几吨的数十亿个神经元,在一段时间里获取大量信息。他们有一套负责语言、记忆以及所有那些东西的模块。而由此出发,就那些人而言,他们能够杀害数百万人。
第 868 段
[Matthew] 是的。
第 869 段
这一切都源自……并不存在什么被美化的观念,说独裁者拥有一个如此非凡的头脑之类的。就只是大脑。
第 870 段
是的,是的。我的意思是,其中很大一部分与这类人组织身边其他人的能力有多强有关。
第 871 段
其他大脑。
第 872 段
是的,所以我总觉得研究灵长类动物学很有意思,也就是研究与我们关系最近的非人类近亲,从中寻找人类将如何行事以及特定的人能够取得什么成就的线索。于是你观察黑猩猩和倭黑猩猩,会发现它们既相似又不同,尤其是在社会结构方面。
第 873 段
我曾就读于亚特兰大的埃默里大学,师从弗朗斯,那位伟大的弗朗斯·德瓦尔。他算是顶尖的灵长类动物学家,最近去世了。他研究黑猩猩时所采用的视角,就像你观看一集《老友记》并理解其中人物彼此互动的动机一样。他会观察一个黑猩猩群体,并基本上运用这种视角。我这是极大地简化了他的工作。
第 874 段
如果你这样做,而不是只说“473号受试者把粪便扔向了471号受试者”,你会从它们作为个体所经历的挣扎来谈论它们,赋予它们应有的尊严,把它们视为有着可以理解的目标和驱动力的行动者,理解它们想从生活中得到什么。而这主要就是我们想从生活中得到的东西:食物、性、陪伴、权力。以同样的视角,你会更容易理解黑猩猩和倭黑猩猩的行为。
第 875 段
而且我认为,这样做能给你提供所需的工具,剥去我们用语言层层附加在人类行为之上的那种虚假复杂性,从这些角度来看待它:哦,这些人类在寻求陪伴、性、食物、权力。我认为,在理解人类行为时,这是一个相当强大的工具。
第 876 段
而我刚刚去亚马孙丛林待了几个星期,这非常直观地提醒了我,地球上许多生命所做的事无非就是想交配。它们都在对彼此大叫。我看到了很多猴子,它们就是在努力给彼此留下深刻印象,或者也许是在争夺权力,但许多权力之争都与它们能否交配有关。
第 877 段
对。交配权往往与首领地位相伴。所以,如果你能从中分得一杯羹,那你的日子就会过得不错。
第 878 段
而且我们愿意认为自己在某种意义上有着根本性的不同,但尤其是在涉及灵长类动物时,我们其实并没有,你知道。我们可以使用更华丽、更富诗意的语言,但驱动我们的一些深层动机或许是相似的。
第 879 段
是的,我认为确实如此。
第 880 段
而这一切都源自这个,大脑。
第 881 段
是的。
第 882 段
那么,你最初是什么时候开始把大脑作为一种,我想,生物机制来研究的?
第 883 段
基本上,我一上大学就开始四处寻找可以让我从事神经科学研究的实验室。我最初是从研究大脑与免疫系统相互作用的角度切入的,这并不是最显而易见的起点,但我当时有这样一个想法:你的思维内容会对身体中的非意识系统产生影响,一种直接的、或许很强大的影响。
第 884 段
也就是那些我们认为属于体内稳态的自动机制,比如抵御病毒、修复伤口。果不其然,两者之间存在大量交叉。我的意思是,这触及了一个我认为没有得到充分认识的关键点。人们对人脑认识或重视不足的一件事是,它基本上控制着你的身体所做的几乎一切,或者在其中发挥着巨大作用。
第 885 段
比如,你试着举出身体中某种不受大脑直接控制或巨大影响的事物,其实相当困难。我的意思是,你可能会说骨骼愈合之类的事情,但即便在这些系统中,下丘脑和垂体最终也会参与协调内分泌系统,而内分泌系统确实会直接影响,比如说,你血液中的钙水平,而这又关系到骨骼愈合。
第 886 段
因此,这些事物之间并不明显的联系表明,大脑确实是整体健康中一个强有力的首要驱动力。
第 887 段
我也从另一个方向意识到了一件事,那就是身体中的大多数系统如何与人脑相互整合,比如它们也会影响大脑,例如免疫系统。我认为,就是那些研究阿尔茨海默病和这类疾病的人。令人惊讶的是,你能从免疫系统、从其他那些看起来显然与神经系统毫无关系的系统中理解这么多相关内容。它们全都协同作用。
第 888 段
是的,你也可以通过一些简单的例子理解这会如何由进化所驱动。如果你生病了,如果你得了传染病、得了流感,让免疫系统告诉你的大脑:“嘿,现在几天内别跟人来往。今晚别去成为聚会的焦点了。事实上,也许就该找个暖和的地方裹上毯子,待上1天或2天。”这会相当有利。
第 889 段
果不其然,这往往正是你在动物和人类身上都能看到的行为。如果你生病了,血液中升高的白细胞介素水平和肿瘤坏死因子α会要求大脑减少社交活动。甚至连四处活动也会减少,感染病毒的动物会表现出较低的运动活动水平。
第 890 段
那么,从早期从事神经科学研究到外科手术,这一步是什么时候发生的?
第 891 段
是的。
第 892 段
这是一次跨越。
第 893 段
这算是思想上的一种演变。我想研究大脑。本科期间,我开始在这个神经免疫学实验室研究大脑。从那里开始,我在某个时候意识到,我不想只是创造知识。我想在现实世界中、在真实的人们的生活中产生实际的改变。所以,尽管之前并没有真正考虑过去医学院,但我当时正沿着进入博士项目的轨道前进。
第 894 段
我说:“嗯,我希望拥有那个选项。我希望真的有可能帮助眼前活生生的人。”稍作调查后,我发现确实存在医学博士与哲学博士联合培养项目,让你可以不必在两者之间作出选择,而是两者都读。
第 895 段
于是我去了南加州大学医学院,并参加了一个与加州理工学院合办的博士项目。事实上,我选择那个项目尤其是因为加州理工学院有一位名叫理查德·安德森的研究人员。他是灵长类神经科学的奠基者之一,并拥有一个猕猴实验室,那里会把犹他阵列和其他电极植入猴子的大脑,试图了解意图是如何在大脑中编码的。
第 896 段
所以我最终去了那里,当时的想法是,也许我会成为一名神经科医生,并在业余时间研究大脑。后来我发现,神经内科,再说一次,我这么说会树敌,但让我感到苦恼的是,神经内科主要做的就是诊断出某种病,然后说:“祝你好运。我们能做的不多。”
第 897 段
而神经外科则截然不同,它是一种强有力的手段,能够让那些正走向糟糕结局的人改变进程,比如那些有可能通过手术治疗或治愈的脑肿瘤。甚至还有脑动脉瘤,那些将要破裂的血管,你可以挽救生命。归根结底,这才是对我而言重要的事。
第 898 段
所以,如我所说,我当时在南加州大学,而那里恰好拥有顶尖的神经外科项目之一。于是我遇到了这些真正堪称传奇的神经外科医生:亚历克斯·哈莱西、迈克·阿普佐、史蒂夫·贾诺塔和马蒂·韦斯,这些堪称传奇的人物就作为活生生的人站在我面前。
第 899 段
所以,这在某种程度上改变了我的想法:神经外科医生不再是住在另一个星球、偶尔前来探访我们的遥远神明,而是同样有烦恼、也是普通人的人类。没有什么东西从根本上阻止我成为他们中的一员。所以在医学院的最后关头,我改变了方向,不再进入另一个专科,转而选择了神经外科,这让我多花了1年时间。
第 900 段
我不得不再做1年研究,因为要转入神经外科时,我在整个流程中已经走得太远了。截止日期已经过了。所以这是一个耗费了时间的决定,但绝对值得。
第 901 段
在成为神经外科医生的培训道路上,最困难的部分是什么?
第 902 段
对,有两点。我认为神经外科住院医师培训有点像是一场痛苦竞赛,就是你能承受多少痛苦,还能面带微笑。
第 903 段
对。
第 904 段
所以有一些锻炼限制,但其实并不怎么——我认为,住院医师内部会把它们视为软弱的表现。所以大多数神经外科住院医师都会竭尽全力工作。而我认为,这必然意味着长时间工作,有时还会超过工时上限。我们很重视遵守摆在我们面前的任何规定。
第 905 段
但我认为比那更重要的是,人们想全力以赴,成为一名更优秀的神经外科医生,因为事关重大。所以要让住院医师在轮班结束时回家,而不是留下来做更多手术,真的是一场斗争。
第 906 段
你是认真地说,最困难的事情之一,真的就是强迫他们睡觉、休息之类的吗?
第 907 段
从历史上看,确实如此。我觉得下一代,我觉得下一代更守规矩,也更注重自我照顾——你是想说更弱吧。好吧,我只是开玩笑,我只是开玩笑。
第 908 段
我可没这么说。
第 909 段
现在我要树敌了。好吧,我明白了。哇,这太有意思了。那么第2点是什么?
第 910 段
性格。也许这两点是有关联的,但是——竞争很激烈吗?
第 911 段
竞争很激烈,而且正如我们之前略有提及的,灵长类动物喜欢权力,而我认为神经外科长期以来一直笼罩着某种神秘、卓越以及诸如此类的光环。所以我认为,这对那些披着这种权威外衣的人来说是一种邀约。通过专业委员会认证的神经外科医生,基本上就是一个行走的、诉诸权威的逻辑谬误,对吧?你有资格走进任何房间,在任何事情上都表现得像个专家。而对抗这种倾向,并不是大多数神经外科医生所擅长的。谦逊并非他们的强项。
第 912 段
对,我有一些认识你的朋友,每当他们谈起你时,都会说你具备一种对神经外科医生而言令人意外的品质,那就是谦逊。我认为这说明,这种品质在神经外科里可能不像在其他职业中那么常见,因为神经外科确实有一种极其宏大、近乎英雄式的色彩,而我觉得这多少会让人有点飘飘然。
第 913 段
是的。嗯,我觉得这让我能够在埃隆的公司里有出色表现。因为我认为,埃隆的优势之一,就是能立刻看穿诉诸权威的谬误。所以,没有人会走进他所在的房间说:“真该死,你必须相信我。我可是造了之前那10枚火箭之类的人。”而他会说:“嗯,你做错了,我们可以做得更好。”或者,“我是过去50年来维持福特生存的人。
第 914 段
在如何造车这件事上,你得听我的。”而他会说不。所以你不会走进他所在的房间说:“嗯,我是神经外科医生。让我来告诉你该怎么做。”他会说:“嗯,我是一个有大脑的人。我自己可以从第一性原理出发思考,非常感谢。而且,我认为应该这样做。我们去试试看,看看谁是对的。”
第 915 段
而且我认为,在他身上,这种方法已经一次又一次被证明非常有效。
第 916 段
如果我们顺着这个话题稍微岔开一下,Neuralink 有一支很有意思的跨学科团队,你可以和他们互动,其中也包括埃隆。你认为成功团队的秘诀是什么?仅仅通过观察这些人,你学到了什么?来自不同学科的世界级专家共同合作。
第 917 段
是的,这里有一个最佳平衡点:人们彼此意见不一,会强有力地表达自己的想法,充满激情地捍卫自己的立场,但仍然能够接受他人的信息,并在自己错了的时候改变想法。所以我喜欢用打磨石头来类比。你把坚硬的东西放进坚硬的容器里,让它旋转。人们彼此碰撞,最终产出的是一个更加精良的成品。
第 918 段
所以,为了在 Neuralink 打造一支优秀的团队,我们一直努力寻找那些不怕充满激情地捍卫自己想法的人。他们偶尔会与共事的人产生强烈分歧,并让最好的想法脱颖而出。这种平衡并不容易把握,再次说回灵长类动物的大脑。
第 919 段
灵长类动物的大脑并不是天生就会这样说:“我满怀激情地把所有筹码都押在这个立场上,现在我却要直接放弃它。承认你是对的。”我们大脑的一部分会告诉我们,那意味着权力的丧失。那是丢脸,是在群体中地位的丧失。而现在,你成了一只泽塔黑猩猩,因为你的想法被彻底击败了。
第 920 段
而你只需要认识到,脑后那个微小的声音是不适应环境的,它无助于团队获胜。
第 921 段
是的,你必须有足够的自信,才能放弃一个自己一直坚持的想法。是的。
第 922 段
是的。
第 923 段
而且,如果你足够频繁地这样做,实际上你会在自己所做的事情上成为世界最佳。我是说,就是那种快速迭代。
第 924 段
是的,至少你会成为一支获胜团队的成员。
第 925 段
乘势而上。你学到了什么?你提到南加州大学有很多杰出的神经外科医生。你从这些人身上学到了哪些关于手术和人生的经验?
第 926 段
是的,我认为是拼命工作,在作为团队一员履行职责的同时努力工作,完成一项难度高得难以置信的任务,长时间得难以置信地工作,彻夜不眠,照顾一个你认为无论做什么可能都活不下来的人。努力工作,好让那些你打心底里讨厌的人在第二天早上显得很出色。
第 927 段
这些人几十年如一日,不懈追求卓越的神经外科技术。而且我认为,我们也因这种卓越而广受认可。尤其是马蒂·韦斯、史蒂夫·贾诺塔、迈克·阿普佐,他们不仅为手术技术作出了巨大贡献,还建立了培训项目,培养了数十位乃至数百位杰出的神经外科医生。我只是很幸运,算是一路跟随在他们身后。
第 928 段
那是什么感受?你提到给一个很可能活不下来的人做手术。这会让你备受煎熬吗?
第 929 段
会。尤其困难的是,恕我对长者充分尊重,当你照顾的是一位80岁老人,而某种东西反正很快也会夺走他们的生命时,这种打击不会那么强烈。所以你失去这样一位患者,而不管怎样,这也是预计未来几年会发生在他们身上的自然进程的一部分。
第 930 段
而照顾一个有2个或3个、4个年幼孩子的父亲,一个才30多岁、本不该遭遇这种事的人;他们因为人生中第一次癫痫发作来到你的急诊室,结果你瞧,他们有一个巨大的恶性脑肿瘤,无法手术或无法治愈。我认为,这种事只要经历寥寥几次,就会真正开始侵蚀你的铠甲。
第 931 段
或者一位年轻母亲来到医院,她脑中发生了大出血,注定无法存活。他们把她4岁的女儿带来,让她在关闭呼吸机之前与母亲做最后一次告别。伟大的亨利·马什是一位英国神经外科医生,他对此说得最为贴切。我想他说的是,每一位神经外科医生都随身带着一片私人的墓地,而我对此深有同感,尤其是面对年轻的父母。那让我痛不欲生。
第 932 段
他们本来还能作出更多贡献。尤其是失去这些人,会产生连锁效应,让这个世界在很长一段时间里对人们而言变得更加糟糕。而面对这一切却感到无能为力,实在很难受。这就是为什么我认为,一个人必须近乎邪恶,才会去反对 Neuralink 这样的公司,或者不断对我们冷嘲热讽,因为我们所做的,就是努力解决这些问题。
第 933 段
我们正努力为人们提供选择,减少痛苦。我们正努力消除受损大脑给生活带来的痛苦。而且,是的,我想,这就是我们反抗熵增的一点小小方式。
第 934 段
是的,当我们视为理所当然的大脑能力被剥夺时,人所承受的痛苦是巨大的。而能够恢复其中一些功能,确实是一份珍贵的馈赠。
第 935 段
是的,我们才刚刚开始。我们还会做更多得多的事情。
第 936 段
那么,你能带我完整了解一下植入 Neuralink 的 N1 芯片之类的整个操作过程吗?
第 937 段
可以,这是一个非常简单、非常简单而直接的操作。手术中由我负责、需要人来完成的部分简单得不能再简单了。这是可以想象到的最基础的神经外科手术之一。而且我认为,有证据表明,某种形式的这种手术数千年来一直有人做。
第 938 段
我想,古埃及有一些已经愈合或部分愈合的环钻术实例,秘鲁或南美洲古代也有。当时这些原始外科医生会在人们的头骨上钻孔,大概是为了释放邪灵,但也可能是为了引流血凝块。而且孔洞边缘有骨骼愈合的证据,这意味着这些人在接受手术后至少存活了几个月。所以,我们做的就是这个。
第 939 段
我们会在头顶的皮肤上切开一个切口,位置就在大脑中最强烈表征手部动作意图的区域上方。所以,如果你是一位专业的音乐会钢琴家,在你演奏的整个过程中,大脑的这个部位都会活跃起来。我们称之为手部旋钮区。
第 940 段
手部旋钮区。
第 941 段
对。
第 942 段
所以这都像是手指的运动。所有这些都在不断放电。
第 943 段
是的,皮层上那里有一道小小的弯曲。大脑的一处褶皱正好在那个位置形成了某种双重折叠。所以你可以在 MRI 上看到它,然后说:“那就是手部旋钮区。”接着你会进行功能测试,并使用一种特殊的 MRI,叫作功能性 MRI,也就是 fMRI。当人们,甚至是那些大脑已不再与手指运动相连接的四肢瘫痪者,
第 944 段
想象手指运动时,大脑的这个部位仍然会亮起。因此,对于任何准备参加我们试验的人,我们都能识别出大脑的这个部位,然后说:“好,我们确认,大脑的这个部位就是你的手部意图区。”然后我会在皮肤上切一个小口,把皮肤掀开,就像打开汽车的引擎盖一样,只是要小得多。
第 945 段
在颅骨上开一个直径正好为 1 英寸的圆孔,取下那一小块颅骨,打开脑膜,也就是大脑的覆盖层。它就像一个装水的小袋子,大脑漂浮在里面。然后把大脑的那个部位呈现给我们的机器人。接下来就是机器人大显身手的地方。
第 946 段
它可以进入其中,将这些远比人类头发细小的电极精准地插入皮层,也就是大脑表面,达到非常精确的深度和非常精确的位置,同时避开覆盖在大脑表面的所有血管。
第 947 段
机器人完成它的部分之后,人类再回来,把植入物放进颅骨上的那个孔里并盖好,将它用螺钉固定在颅骨上,再把皮肤缝合起来。所以整个过程要持续几个小时。与一般涉及大脑的神经外科手术相比,它的风险极低,后者可能会,比如说,打开大脑深部区域,或者处理大脑中的血管。
第 948 段
这种在大脑表面开口、只进行皮层微插入的手术,其风险显著低于许多常规进行的肿瘤或动脉瘤手术。
第 949 段
所以,通过机器人和计算机视觉进行的皮层微插入,其设计目的就是避开血管。
第 950 段
没错。
第 951 段
我知道你在这方面有点偏向一方,但我们来比较一下人类和机器吧。
第 952 段
好。
第 953 段
那么,在人类文明发展的现阶段,人类外科医生擅长做什么,而机器人外科医生又擅长做什么?
第 954 段
是的,是的,这是个好问题。人类是通用型机器。我们能够适应异常情况。我们能够临场改变计划。我清楚地记得多年前在圣迭戈做过的一场手术,当时的计划是在耳后开一个小孔,然后重新调整一条血管的位置,这条血管压在了面神经,也就是三叉神经,那条通往面部的神经上。
第 955 段
当那条血管压在神经上时,可能会引起难以忍受、可怕的电击样剧痛,人们形容那种感觉就像被赶牛电棒电击一样。所以,这项漂亮而精巧的手术就是把这条血管从神经上移开。我们手术团队进入那里,开始移动这条血管,然后发现那条血管上有一个巨大的动脉瘤,而它在术前扫描中并不容易看出来。
第 956 段
所以计划必须动态调整,而人类外科医生对此完全没有问题。我们接受过应对所有这些情况的训练。机器人在那种情况下就不会做得那么好,至少以它们目前的形态而言是这样。完全由机器人实施的手术,比如Neuralink手术中植入电极的部分,是按照既定计划进行的。
第 957 段
所以人类可以中断流程并改变计划,但机器人实际上无法在中途改变计划。它按照编程方式和要求的运行方式工作。它非常精确地完成任务,但在如何应对不断变化的情况方面没有很大的自由度。
第 958 段
所以,当外科医生进入一种情境时,可能会遇到非常多种出乎意料的情况,其中可能包括一些你必须动态调整以应对的细微问题。
第 959 段
没错。
第 960 段
而机器人并不擅长这个。
第 961 段
目前是这样。我认为,随着AI让机器人的响应参数得到大幅拓宽,我们正处于一个新时代的黎明,对吧?我的意思是,你不能看着一辆自动驾驶汽车,说它是在非常狭窄的参数范围内运行。如果一只鸡跑过马路,程序未必专门针对这种情况进行过编程,但Waymo或自动驾驶的Tesla都能毫无问题地对此作出恰当反应。所以外科手术机器人目前还没有达到这个程度,但给它一些时间。
第 962 段
然后可能会有很多,有点像半自主的可能性,也许机器人外科医生可以判断,这种情况完全熟悉,或者这种情况并不熟悉。在不熟悉的情况下,可以由人类接管。但基本上就是要非常保守地判断:“好,这肯定没有问题,也没有意外”,然后让人类处理那些意外、边缘情况以及所有这些。这是一种可能性。
第 963 段
所以你认为最终你会失业吗?嗯,指的是你作为神经外科医生,你的工作是神经外科医生。人类,这个地球上将不会剩下多少神经外科医生。
第 964 段
在我的职业生涯期间,我并不担心自己的工作。我想我会告诉我的我的孩子,不一定要进入这一行,具体取决于20年后情况如何。
第 965 段
这太令人着迷了,因为我的意思是,如果说我从事什么工作,我会说是编程。如果你问我,在过去,我不知道,20年里,我会建议人们做什么,我会告诉他们,是的,去做吧。如果你是程序员,你永远都会有工作,因为计算机越来越多,还有诸如此类的东西,而且薪水也不错。
第 966 段
但随后你意识到,这些大型语言模型出现了,而且它们生成代码的能力真的强得惊人。所以一夜之间,你可能会惊讶地想:“哇,那么人类真正的贡献是什么?”但随后你开始思考:“好吧,人类似乎确实有能力,像你说的那样,应对新颖的情况。”就编程而言,就是能够想出新颖的点子来解决问题。
第 967 段
机器似乎还不能完全做到这一点。而当风险非常高、事关生死时,就像外科手术,尤其是神经外科手术那样,要让机器人真正取代人类,所承担的风险非常高。但有意思的是,在Neuralink这个案例中,存在人机协作。
第 968 段
对,对。我做它做不了的部分,它做我做不了的部分。我们是朋友。(Lex笑)——我看到你们进行了大量练习。所以我的意思是,Neuralink的一切都会接受极其严格的测试。但我看到的一件事是,有一个用来进行手术的替代模型。
第 969 段
对。
第 970 段
所以这既是给机器人练习的,也是给人类以及整个流程中所有参与者练习的。练习这种手术是什么样的?
第 971 段
相当紧张。所以在人类外科手术中,没有与此相似的做法。人类外科手术有点像一种手工技艺,历经世代,由师傅直接传授给徒弟。我的意思是,从字面上说,你学习如何成为一名给人做手术的外科医生,就是通过给人做手术。
第 972 段
我的意思是,一开始,你会看教授们做许多台手术,然后,他们终于会把手术中简单的部分交到你手上,之后再把更复杂的部分交给你。而随着你对手术要点和目的的理解不断加深,在理想情况下,你会获得更多责任。事情并不总是顺利。Neuralink采用的方法略有不同。当然,我们尽可能在动物身上进行了练习。
第 973 段
我们做了数百台动物手术。而到了要为第一个人类做手术的时候,我们拥有一支非常出色的工程师团队,他们制作了极其逼真的模型。其中一位工程师Fran Romano尤其制作了一个会搏动的大脑,将其置于一个定制的3D打印头骨中,这个头骨与患者的解剖结构完全吻合,包括其面部和头皮特征。
第 974 段
所以当我能够用它练习时,我的意思是,它已经达到了在合理范围内应当能够达到的最接近真实手术的程度,而且所有细节都具备,包括在这个定制头部上连接一个人体模型的身体。因此,我们进行手术练习时,会把那个身体推进CT扫描仪,进行一次模拟CT扫描,再把它推回来,并以口头方式执行所有常规安全检查。
第 975 段
“停,这位患者,我们正在确认他的身份,是人体模型几号,某某某。”然后使用标准的手术神经导航设备、标准的外科手术钻,在Neuralink用于进行所有手术练习的同一间手术室里,恰好在正确的位置打开大脑。
第 976 段
然后让头骨保持打开,让大脑搏动,这给机器人完美、精确地规划并将那些电极插入正确的深度和位置增加了一定难度。所以,是的,在我们为这台手术所做的练习有多么全面这方面,我们算是开辟了新天地。
第 977 段
所以今年1月,第一个人类接受Neuralink植入物,这不仅对Neuralink来说是一个历史性时刻、一个重大里程碑,在某种程度上对人类也是如此。带我了解一下为Noland进行的手术。参与其中是什么感受?
第 978 段
是的。嗯,我们很幸运,在巴罗神经学研究所有非常出色的合作伙伴。我认为,他们是世界上首屈一指的神经外科医院。他们尽可能让试验顺利启动,并凭借他们在如何安排细节方面的专业知识给予了我们极大的帮助。从某些方面来说,这是一场压力大得多的手术。
第 979 段
我的意思是,尽管就参与者的安全而言,结果并没有什么特别的悬念,但观察者的数量、人员的数量,还有医院里坐满了人的会议室,大家观看着直播,为这场手术能够完美进行而加油,这些都只会增添压力,而这种压力即使在最紧张的常规神经外科手术中也并不常见。比如切除肿瘤或放置脑深部刺激电极。
第 980 段
而且此前从未在人类身上做过。存在未知的未知。因此,对整个团队来说,确实有一定程度的紧张感,因为不知道我们是否会遇到,比如,超出预期程度的脑部移动,或者脑组织下垂到远离颅骨的位置、导致难以植入,又或者其他某种未知的未知问题。
第 981 段
幸运的是,一切都很顺利,那次手术的结果是我们所能想象的最顺利的结果之一。
第 982 段
你紧张吗?我的意思是,你有点像是处于超级碗比赛中的四分卫那种情形。
第 983 段
极其紧张。极其紧张。手术进展顺利并且结束时,我非常高兴。期待第2例。
第 984 段
是的。即便进行了所有那些练习,做了所有那些准备,你也从未身处这样一种局面:就被人们注视而言,风险依然很高。而且,考虑到媒体的运作方式,我们或许也应该提到,很多人可能怀着某种阴暗的心态,希望事情进展不顺。
第 985 段
嗯,我认为财富很容易招致仇恨、嫉妒或诸如此类的情绪。而且我认为,有一整个行业都围绕着如何推动点击量运转,而坏消息非常有利于点击量。因此,任何能够把一件事变成坏消息的方式,都会非常有利于点击量。
第 986 段
这真的很糟糕,因为我认为这会给人们带来压力。它会阻碍人们尝试解决真正困难的问题,因为要解决困难的问题,你就必须踏入未知。你必须去做以前没人做过的事,而且必须承担风险。
第 987 段
是的。
第 988 段
经过计算的风险。你必须采取各种安全预防措施,但风险依然存在。我只是希望人们能更多地赞颂这一点,赞颂承担风险,而不是像有些人那样只在一旁等待,等着失败发生,然后指出这个失败。是啊,这很糟糕。但在这件事上,一切都进行得毫无差错,确实非常棒,不过我会说,这种压力是不必要的。
第 989 段
现在有一个人真的是拿自己的血肉之躯参与其中,有一位参与者的福祉取决于这件事能否顺利进行,你得是个相当恶劣的人,才会盼着它出问题。所以希望人们照照镜子,在某个时候意识到这一点。
第 990 段
所以你真的坐在前排看机器人工作了吗?你看到了整个过程?
第 991 段
对,我是说,因为整个过程中所有的医疗决策都需要由一名医学博士负责,所以在暴露出大脑并将其呈现给机器人之后,我退出了手术的无菌操作,然后在机器人的软件界面上设置靶点,这些靶点会告诉机器人要在哪里插入每一根丝线。无论这算不算什么,那些操作都是我用手握着鼠标完成的。
第 992 段
所以是你设置的那些靶点?
第 993 段
对。
第 994 段
哦,酷。所以就像是机器人通过计算机视觉提供一批候选位置,然后由你来最终确定。
第 995 段
对。这个团队的软件工程师非常出色。所以他们实际上提供了一个界面,你基本上可以使用套索工具,选中一块优质的大脑区域,它就会自动避开该区域内的血管,并自动设置一批靶点。这样一来,人类操作员就可以选择非常理想的大脑区域,并在这些区域中密集设置靶点,也就是我们认为能够最为高保真地表征手指动作和手臂运动意图的区域。
第 996 段
我看过类似的图像。对有强迫症的我来说,不知为什么看着真的很舒服,我记得有个 Reddit 子版块叫“莫名令人满足”。
第 997 段
[马修] 对,我很喜欢那个子版块。
第 998 段
看到不同的靶点位置避开血管,同时还最大限度地提高那些位置对信号的效用,真是莫名令人满足。就是让人感觉很好。就像,啊。
第 999 段
作为一个看到大脑出血会产生强烈生理反应的人,我可以告诉你,看着电极本身进入大脑而不引起出血,是一件极其令人满足的事。
第 1000 段
对,对。所以你说,当一切都完美进行时,你的感受是如释重负。
第 1001 段
对。
第 1002 段
目前你们能深入大脑多深,将来又能深入多深,比如说,在 Neuralink 这方面。似乎在大脑中越深入,挑战就越大。
第 1003 段
是的,所以从广义的神经外科来说,我们可以到达任何地方。对我而言,把脑深部刺激电极放到接近大脑最底部的位置是常规操作:从顶部进入,将一根大约2毫米粗的导线一直送到大脑底部。这并不是什么革命性的技术。很多人都这样做。而且我们可以非常精准地完成。我每个月都会用 Globus 的机器人做几次这种手术。这是相当常规的。
第 1004 段
在那种情况下,什么充当你的眼睛?你能看到什么?你可以使用什么样的技术来可视化自己所在的位置,为你照亮前路?
第 1005 段
是的,所以从软件层面来说,这是一个很酷的过程。你会获取一份术前 MRI,它包含整个大脑分辨率极高的数据。你让患者进入睡眠状态,把他们的头放进一个能非常牢固地固定颅骨的框架中,然后趁他们戴着这个框架熟睡时对头部进行 CT 扫描,再通过软件合并 MRI 和 CT 图像。你会根据 MRI 制定计划,因为在 MRI 上可以看到大脑深处的这些神经核团。
第 1006 段
你在 CT 上看不到它们,但如果你相信两幅图像的融合,那么你就能间接知道它们在 CT 上的位置。因而也能间接知道,相对于用螺钉固定在他们头部的钛框架,这些目标位于哪里。所以,这是 60 年代的技术:根据进入点和目标点手动计算轨迹,然后调节一些看起来很滑稽的钛制执行器,手动调节那些上面带有小刻度线的执行器。
第 1007 段
这种技术的现代版本是使用机器人。就像一条小型 KUKA 机械臂,你可能会在 Tesla 工厂看到它制造汽车。
第 1008 段
这条小型机械臂可以向你显示根据术前 MRI 设定的预定轨迹,并建立一个非常牢固的固定装置,你可以通过它在颅骨上钻一个小孔,将一根细小的刚性导线深入大脑中那个中空的区域,再把电极穿过那根中空导线,然后移除除电极以外的所有东西。因此最终电极会被极其精准地放置在远离颅骨表面的位置。
第 1009 段
这是已经在世界上应用了一段时间的标准技术。目前 Neuralink 完全专注于皮层目标,也就是表面目标,因为没有一种简单的方法能把比如数百根导线送入大脑深处而不造成大量损伤。所以你的问题是,你看到了什么?嗯,我看到的是屏幕上的 MRI。我看不到 DBS 电极在抵达那个深部目标的途中穿过的所有组织。
第 1010 段
因此,采用这种方法时,人们接受这样一个事实:大约每一百名患者中会有一名因为将那根导线盲目送入大脑深部而在脑内某处出血。对 Neuralink 来说,这样的安全水平是不可接受的。我们的出发点是,希望它的安全性能显著提升,或许要安全两到三个数量级。
第 1011 段
真正安全到有朝一日,你或我即使没有严重的医疗问题,也可能在午休时说:“好啊,当然,我去装一个。我一直想升级到最新版本来着。”因此,在这种前提下,安全要求很高。所以对于如何任意接近大脑深部目标,我们尚未确定最终解决方案。
第 1012 段
这很有意思,因为你好像必须设法避开血管。也许可以用一些有创意的方式来做同一件事,比如绘制出血管的高分辨率几何结构,然后你就可以盲目地进入。但你要如何以一种超级稳定的方式把它绘制出来?这里面有很多有意思的挑战,对吧?
第 1013 段
[Matthew] 是的。
第 1014 段
不过幸运的是,在表面还有很多事情可以做。
第 1015 段
没错。所以我们在表面有视觉功能。实际上,我们已经在把电极缝入脊髓方面取得了巨大进展,可将其作为脊髓损伤的一种潜在绕行方案:让安装在大脑上的植入物把运动意图传递给安装在脊柱上的植入物,后者可以促使此前瘫痪的手臂和腿部肌肉收缩。
第 1016 段
这太令人震撼了。简直不可思议。所以,这里的努力就是试图把大脑与脊髓、周围神经系统连接起来。那么这做起来有多难?
第 1017 段
我们已经在动物身上以非常粗糙的形式实现了这一点。
第 1018 段
太惊人了。
第 1019 段
是的,我们已经做过——所以,这有点类似于 Noland 的情况,他能够以数字方式移动光标,而你们在这里进行的是同一种交流,但用的是你所拥有的实际因素。
第 1020 段
是的。
第 1021 段
[Lex] 这太迷人了。
第 1022 段
是的,所以我们已经让处于麻醉状态的动物做出抓握动作,并以某种行走模式移动它们的腿。再说一次,这还处于早期阶段,但这类技术前景光明。瘫痪人士应该期待那个光明的未来。他们将会有多种选择。
第 1023 段
是的,而且还有很多某种中间选项或额外选项,比如你拿一台 Optimus 机器人,例如它的手臂,然后能够控制那条手臂。把手臂的手指和手当作假肢。
第 1024 段
所以骨骼也在变得更好。
第 1025 段
所以是骨骼。是的,所以这两者是相辅相成的。不过,在深入思考并对 Neuralink 做了更多研究之前,我并不完全明白你能在数字层面做到多少事情。所以,对于这种数字心灵感应,我之前并不完全明白,你真的可以映射意图,就像你描述的手旋钮区域那样,你可以映射意图。只要想象它,想着它。
第 1026 段
这种意图可以映射为数字世界中的实际行动。而如今越来越多的事情都可以在数字世界中完成,以至于它能让你重新与外部世界建立联系。如果你是四肢瘫痪者,它可以让你拥有自由、拥有独立。这真的非常强大。你真的可以凭此走得很远。
第 1027 段
是的,我们的第一位参与者非常了不起。他接二连三地打破世界纪录。
第 1028 段
而且他从中获得了乐趣,这很棒。再回到手术这件事,以及你的整个历程,你私下跟我提过,你星期一有手术。所以你就像一直都在做手术。
第 1029 段
是的。
第 1030 段
也许这是个荒唐的问题,要怎样才能变得擅长做手术?
第 1031 段
练习,反复。其他任何事情也一样。人们有一百万种方式表达同一件事,还会写书来卖、讲这件事,但你可以称之为一万小时,也可以称之为,你知道,拿出人生中的一段时间、人生中的某个比例,专注于此,痴迷于让自己做得更好。反复练习,保持谦逊,认识到在这个过程中的任何阶段你都并不完美。认识到你的技术还有需要改进的地方。
第 1032 段
愿意接受那些对如何做这件事有不同看法的人的反馈和指导。然后就是始终想要做得更好的意志。幸运的是,如果你不是反社会人格,我想你的患者每天来就诊时都会把这种动力带给你。他们会迫使你一直想做得更好。
第 1033 段
是的,挺身而出。我的意思是,那是一个活生生的人,一个你能够帮助的活生生的人。
第 1034 段
是的。
第 1035 段
所以每一台手术,即便是完全相同的手术,在这个人和另一个人之间也会有很大的差异吗?
第 1036 段
是的,差异相当大。我的意思是,对我们来说,一个很好的例子是颅骨相对于身体轴线正常平面的角度,也就是手旋钮区上方颅骨的角度,变化范围相当大。我的意思是,有些人的颅骨非常平,而有些人那片区域上方的颅骨倾角非常陡。这会影响他们的头部如何固定在我们使用的那种支架中,也会影响机器人必须如何接近颅骨。
第 1037 段
是的,人们身体构造的差异,就像你在街上看到的行人彼此之间的差异一样。你在那里看到多少体形和体型上的差异,我们在大脑解剖结构和颅骨解剖结构中也会看到多少差异。有些人因为颅骨太厚或太薄,或者头皮太厚或太薄,我们不得不把他们排除在试验之外。
第 1038 段
我想我们覆盖了中间大约97%的人,但你不可能把人类解剖结构的所有变异都考虑进去。
第 1039 段
那里到底有多软糊、多杂乱?因为上生物课时,图示总是非常干净清晰。神经科学里,神经元的图片也总是非常漂亮,而且各不相同。但每当我看到真实大脑之类的图片时,我都不知道发生了什么。
第 1040 段
是的。
第 1041 段
所以生物系统在现实中到底有多么?要弄清楚发生了什么有多难?
第 1042 段
不算太难。一旦你真正习惯了,这就是经验、技能和教育真正发挥作用的地方:如果你观察过一千个大脑,就会更容易在脑海中一层层剥开,比如说,那些遮挡着脑沟和脑回的血管,也就是大脑表面的褶皱形态。
第 1043 段
偶尔,在你刚开始做这件事的时候,当你打开颅骨,看到的情况与根据MRI预想的并不一致。随着经验增加,你会学会在脑海中剥开那一层血管,看到下面大脑的褶皱形态,并把它作为判断自己所在位置的标志。
第 1044 段
[Lex] 褶皱是一个标志?所以就像—— ——是的。我之前描述过手旋钮区。它是大脑中的一种褶皱形态。它有点像希腊字母欧米伽形状的大脑区域。
第 1045 段
所以你能识别出手旋钮区。比如我给你看一千个大脑,每个只给你一分钟,你都会说:“对,就是那个?”
第 1046 段
当然。
第 1047 段
所以大脑的那个区域确实有某种独特性,比如就它的几何结构、它的拓扑结构而言。
第 1048 段
是的。
第 1049 段
它大概在—— ——你的大脑顶部有一条向下延伸的带状区域,叫作初级运动区。我相信你看过把“小人图”铺在大脑表面的那张图,就是那个嘴唇很大、双手巨大的奇怪小人。那个小人差不多是双腿位于大脑顶部,面部、手臂区域在更下方,然后某种嘴巴、嘴唇、舌头区域还要更靠下。所以手就在那里面。
第 1050 段
然后,至少对大多数人的左脑来说,控制言语的区域就在它的下方。所以,你身体中任何由你自主活动的肌肉,其中绝大多数都对应着那条带状区域,或者说那些意图都来自大脑中的那条带状区域。而手旋钮区的褶皱就在正中间。
第 1051 段
视觉区在这里后面。
第 1052 段
在后面,对。
第 1053 段
也靠近表面吗?
第 1054 段
视觉区要稍微深一点。这就回到了你关于为了实现视觉功能需要深入到什么程度的问题。我们不能只在大脑表面操作。我们必须能够深入进去,不需要像做DBS时那么深,但可能要比我们为手部功能植入时习惯的深度再深一厘米。所以这项工作仍在进行中。这是一系列有待克服的新挑战。
第 1055 段
顺便说一下,你提到了犹他阵列。我刚看到它的一张图片,那东西看起来很吓人。
第 1056 段
[Matthew] 是的,钉床。
第 1057 段
这是因为它是刚性的。而如果你看看那些丝线,它们是柔性的。关于这种柔性,有什么让你觉得有意思的地方?也就是用这种柔性丝线把电极送到神经元旁边的做法。
第 1058 段
是的,我的意思是,那里的目标源于经验。我的意思是,在我们出现之前,制造犹他阵列并使用了几十年的人为我们奠定了基础。
第 1059 段
Neuralink之所以诞生,以及这种技术路线之所以出现,部分源于人们在犹他阵列经常失效后认识到的一种需求,因为那些刚性电极,也就是那些实际上用气锤锤进大脑的尖针,会引发不良的免疫反应,导致电极尖针实质上被瘢痕组织包裹起来。
第 1060 段
所以我刚到加州理工学院的Andersen实验室时,那里正在进行的一个项目,是研究能否利用化疗阻止瘢痕形成。把一张钉床硬塞进大脑,然后再用化疗处理它,试图防止瘢痕组织形成,这种情况已经相当糟糕了。就像是,各位,也许我们走偏了。也许有必要进行一次根本性的重新设计。
第 1061 段
所以,Neuralink 采用高度柔韧的微小电极,这种方法避免了大量出血,也避免了将刚性电极猛力插入大脑时最终会发生的大量免疫反应。因此,我们看到,我们的电极寿命和功能,以及紧邻电极的脑组织的健康状况,都非常出色。我的意思是,在我们的动物模型中,现在已经能持续数年了。
第 1062 段
关于大脑的生物学特性,大多数人有什么不理解的地方?我们提到了脉管系统。那真的很有意思。
第 1063 段
我认为最有意思、也许也最被低估的事实是,它确实控制着几乎一切。我的意思是,我不知道,随便举个例子,想象一下,你想要一个控制生育能力的操纵杆,你希望能够开启和关闭生育能力。我的意思是,大脑本身确实存在可以调节生育能力的靶点。比如说血压。你想调节血压。大脑中确实存在能做到这一点的靶点。
第 1064 段
那些乍看之下并非大脑问题的事情,可能可以在大脑中解决。因此,我认为,对于所有困扰人们的问题而言,把大脑作为主要治疗对象是一个探索不足的领域。
第 1065 段
这是一个非常引人入胜的看待方式。比如,有很多病症,我们可能认为它们与大脑毫无关系,但它们可能只是某种实际上始于大脑的问题的症状。问题的实际源头。最主要的源头是大脑中的某种东西。
第 1066 段
是的,并不总是如此。我的意思是,肾病确实存在。但你可以操控大脑中的一些操纵杆,从而影响所有这些系统。
第 1067 段
还有旋钮。
第 1068 段
是的。
第 1069 段
大脑里有开关和旋钮,这一切都源自那里。你会在大脑里植入一枚 Neuralink 芯片吗?
第 1070 段
会。我想现在的使用场景是使用鼠标,对吧?我已经能做到这一点了。所以它没有什么价值主张。单从安全性来看,当然可以。我明天就会做。
第 1071 段
你说到鼠标的使用场景,是不是在研究了这一切之后,还有一部分原因只是看到 Noland 玩得那么开心?如果你能让用鼠标时的每秒比特率变得非常高,比如能够进行交互,因为仔细想想,在智能手机上,你滑动屏幕的方式,彻底改变了你与一个东西互动的方式。这很微妙。
第 1072 段
你没有意识到,但能够触摸手机、用手指滚动屏幕,这就像是,它改变了一切。人们曾确信打字需要键盘。其中有很多人机交互方面的因素,改变了我们与计算机交互的方式。所以,鼠标或许能达到某种足以改变一切的速度。
第 1073 段
是的。
第 1074 段
比如说,你也许能够极快地在屏幕上四处点击。而为了能与数字设备进行快得多的交互,我可以想象自己植入 Neuralink。
第 1075 段
是的,我认为从大脑中记录言语意图或许也会改变一些事情。对普通人来说,它的价值主张在于,键盘是一种相当笨拙的人机界面,需要大量训练。普通人所能达到的最高表现差异很大。我认为,把键盘从这个过程中去掉,直接采用一种自然的词语到计算机的界面,或许会改变很多人的情况。
第 1076 段
如果人们真是因为这个原因去做,那就太好笑了。即使你有极其准确的语音转文字功能——目前还没有,但假设它变得超级准确——如果人们选择 Neuralink,只是为了避免说话时令人尴尬的那一面,比如在公共场合对着手机说话,看起来像个混蛋,那就太好笑了。而这确实是一个,怎么说,这确实是一个限制。
第 1077 段
是的。我的意思是,比如用骨传导装置,它可以成为一副看不见的耳机,再加上把词语想进软件里并让它回应你的能力,那开始听起来有点像嵌入式超级智能。如果你能默默请求任何主题的维基百科文章,并让它读给你听,而外部世界看不出发生了任何变化,那么首先,标准化考试就过时了。(笑)——是的。
第 1078 段
如果用户体验方面做得好,它可能会带来改变。我不知道它是否会改变社会,但它确实可以像智能手机那样,让我们与数字设备互动的方式发生某种转变。只是必须研究所涉及的一切是否安全,我完全愿意试试。它不必发展到某种不可思议的程度,比如连接你的视觉,或者连接其他某个地方,就像连接你整个大脑一样。
第 1079 段
那可以只是连接到手部旋钮区。你可能会拥有很多有趣的互动、人机交互可能性。这真的很有意思。
第 1080 段
是的,而且学术界的技术正在以光速进步。我记得 UC Davis 的 Sergey Stavisky 实验室发表过一篇非常了不起的论文,基本上初步解决了言语解码问题。他们好像能以非常高的准确率识别 125,000 个词,这——所以你只是在脑中想着那个词?
第 1081 段
是的。
第 1082 段
想着那个词,然后你们就能把它读出来?
第 1083 段
是的。
第 1084 段
天啊。就像你必须有把它说出来的意图。
第 1085 段
对。
第 1086 段
所以就像发出内心的声音。天啊,对我来说,你们能够实现意图和信号的映射,实在太不可思议了。你只需要想象自己正在做这件事。而如果你得到反馈,知道它确实奏效了,你就能变得非常擅长。比如,首先你的大脑会作出调整,而你会像培养任何其他技能一样培养这项技能。就像盲打一样,你也是用同样的方式培养出来的。
第 1087 段
对我来说,哪怕只是能玩玩这个,都非常迷人。说实话,我会仅仅为了能玩这个而植入 Neuralink。只是为了试试我的思维学习这项技能的容量和能力。这就像学习打字和学习移动鼠标的技能。这是另一种移动鼠标的技能,不是用我的身体,而是用我的思维。
第 1088 段
我迫不及待想看看人们会用它做什么。我觉得我们现在就像穴居人。我们就像拿棍子敲石头,还以为自己在创作音乐。等这些东西在某个时候变得更加普及,就会出现钢琴的等价物,让人能以一种我们甚至未曾预料的方式用大脑创作艺术。我很期待。
第 1089 段
把它交给一个青少年之类的。每当我觉得自己很擅长某件事时,我总会去找,比如,我不知道。即使是玩电子游戏时每秒的比特数,你也会意识到,把它交给一个青少年,你把你的链接交给一个青少年,只要他们人数够多,那种东西,他们很擅长把事情练好。他们会达到每秒数百比特。即使只使用当前的技术也是如此。
第 1090 段
很可能。很可能。
第 1091 段
因为它也会让人上瘾,就是那种通过改进和训练让数字上升的部分,因为它几乎就像一项技能。另外,另一端还有会适应你的软件。尤其是如果这种适应过程的算法变得越来越好,你们就像是在共同学习。
第 1092 段
是的,我们目前才刚刚触及它的表面。还有太多事情要做。
第 1093 段
那么从完全相反的另一面来说,你体内植入了一枚 RFID 芯片。
第 1094 段
是的。
第 1095 段
我是这么听说的,不错。
第 1096 段
一个不起眼的小东西。
第 1097 段
它是一种无源设备,你用它来打开存放绝密资料的保险箱之类的吗,还是你用它做什么?这背后有什么故事?
第 1098 段
我不是第一个。这里有一整个由做过这种事的怪异生物黑客组成的社群,我想早期的使用场景之一是储存私人的加密货币钱包密钥之类的。我稍微尝试过一阵子,从中找了些乐子。
第 1099 段
你身体里的某个地方植入了一些比特币。你不能说在哪里,是吧。
第 1100 段
是的,实际上,是的。(Lex 笑)这就像现代版的从沙发垫缝里找到零钱:我把一些自以为一文不值的孤儿加密货币放在里面,然后忘了好几年。后来回头一看,发现有一个社群的人很喜欢它,并把它的价值撑了起来。所以它涨了 50 倍。
第 1101 段
哇。
第 1102 段
所以那些沙发垫缝里有很多零钱。(Lex 笑)——太好笑了。
第 1103 段
但主要用途大多是作为技术演示设备。里面存着我的名片。把它贴到手机上,就可以扫描读取。它能打开我家的前门,诸如此类的简单用途。
第 1104 段
这是很酷的一步。把某样东西植入身体,是一次很酷的跨越。我的意思是,也许这与 Neuralink 是一次相似的跨越,因为对很多人来说,把某样东西放进身体,把电子设备放进生物系统,这种概念是一次巨大的跨越。
第 1105 段
是的,我们对皮肤这道屏障有一种神秘主义式的看法。我们完全能接受膝关节置换、髋关节置换、牙科植入物。但对于头骨所代表的那道无法存活的屏障,仍然存在一种神秘主义式的看法。我认为,它需要像任何其他实际屏障一样被看待。问题不在于,打开头骨这件事有多么不可思议?问题在于,我们能带来什么益处?
第 1106 段
那么,根据你做过的所有手术,以及你对大脑的一切了解,神经可塑性在多大程度上发挥作用?大脑的适应能力有多强,比如说,哪怕只是在术后恢复或适应术后状况方面。
第 1107 段
对我和其他与我人口特征相同的人来说,令人难过的答案是,可塑性会随年龄增长而下降。愈合能力会随年龄增长而下降。我的白头发太多了,没法对此感到乐观。有一些利用电刺激提高可塑性的理论方法。但没有任何一种得到完全证实,成为足够可靠、可以广泛提供给人们的机制。
第 1108 段
不过,是的,我认为有理由保持乐观,我们或许能找到某种有用的东西,比如能改善学习的植入式电极。
第 1109 段
当然,Nicholas Schiff、Jonathan Baker 以及其他人最近开展了一些非常了不起的工作。他们有一组中度创伤性脑损伤患者,这些患者的大脑深部核团中植入了电极,那个核团叫作中央中核,或者就在中央媒体核附近。当他们向大脑的那个部位施加少量电流时,效果几乎就像电子咖啡因。
第 1110 段
他们能够改善人们的注意力和专注度。他们能够提升人们完成任务的表现。我想有一个案例是,有个人原本无法工作,但设备开启后,他们就能够找到一份工作。
第 1111 段
而对我来说,Neuralink 和其他这类技术的圣杯之一,从纯粹的功利主义角度看,就是我们能否让人们在经济上重新有能力照顾自己和家人?对于一个完全依赖他人、甚至可能需要大量照护资源的人,我们能否让他们处于完全独立的境地,照顾自己,并回馈他们的社区?
第 1112 段
我认为这是一个非常有吸引力的设想,它激励着我所做的许多事情,也是 Neuralink 的许多人为之努力的目标。
第 1113 段
这就是一种很酷的可能性:如果你把 Neuralink 植入那里,大脑会适应,就像大脑的其他部分也会适应一样。
第 1114 段
是的。
第 1115 段
并把它整合进去。大脑做到这一点的能力真的很有意思。你能在多大程度上做到这一点,可能还不得而知,但你现在正把一个外部的东西连接到它上面,尤其是当它开始进行刺激时,就像生物大脑和位于它外部的电子大脑协同工作。这里面的可能性真的很有意思。仍然未知,但很有意思。感觉大脑真的很擅长适应任何东西。
第 1116 段
是的。
第 1117 段
但当然,它本身已经是一个这样的系统,就像每样东西都有其用途,所以你不想对它干预太多。
第 1118 段
是的,这就像从一个生态系统中消灭一个物种。你不知道其中那些微妙的相互联系和依赖关系是什么。大脑无疑是一头精妙而复杂的野兽。而且,我们并不知道自己所做的一个改动可能会带来的每一种下游后果。
第 1119 段
你觉得自己会继续做下去吗?你提到了P1,那么P2、P3、P4、P5的手术呢?就是越来越多、越来越多、越来越多的人。
第 1120 段
我认为,如果所有手术都需要由我来做,那说明存在某种脆弱性,或者是公司这边的一种失败。我非常希望努力实现这样一种流程:在手术方面,它简单、稳健到几乎任何人都能操作。我们希望摆脱对高深专业知识或丰富经验的要求,不必具备这些也能成功完成手术,并让这个过程尽可能简单、尽可能容易推广。
第 1121 段
我的意思是,如果地球上的每一位神经外科医生都能毫无困难地做这件事,我会非常高兴。我认为,我们可能距离允许非神经外科医生做这件事的监管环境还很远,但并非不可能。
第 1122 段
好吧,我报名参加。你有没有把机器人R1拟人化过?比如说,你会给它取名字吗?你会把它视为朋友,视为一个与你并肩工作的伙伴吗?
第 1123 段
我的意思是,在某种程度上,它————或者解剖学,谁会得到那个空隙。
第 1124 段
在某种程度上,是的,这是一种复杂的关系。
第 1125 段
所有美好的关系都是如此。
第 1126 段
有意思的是,在手术进行到一半时,有一个环节我基本上会与机器人肩并肩站着。所以,如果你在房间里观察肢体语言,那就是我并肩作战的兄弟。我们正在共同解决同一个问题。是的,我没有感到它构成威胁。
第 1127 段
继续这样告诉自己吧。(笑)这些年来你做过的所有手术、你帮助过的人,以及你提到的风险,那些极高的风险,怎样改变了你对生与死的理解?
第 1128 段
是的。它会让你产生一种非常切身的感受,这听起来可能很老套,但它确实会让你非常切身地感受到,死亡是不可避免的。一方面,作为神经外科医生,你深度卷入这些简直难以理解的悲剧:比如年轻的父母去世,留下一个4岁的孩子。另一方面,它也会稍微减轻其中的刺痛感,因为你会看到死亡是多么普遍,普遍到令人麻木。
第 1129 段
我避开死亡的概率是零。我知道,现在的技术乐观主义者和长寿爱好者会不同意那个0. 0%的估计。但我看不到我们这一代人有任何可能避开它。熵是一股强大的力量,而我们是极其精巧、脆弱、易碎的DNA机器,承受不了我们所遭受的宇宙射线轰击。所以一方面,每一个曾经活过的人都已经死去或终将死去。
第 1130 段
另一方面,你所爱的人不复存在,这是你能想象施加在任何所爱之人身上的最难以承受的事情之一。我相信你也有已经不在人世的朋友,甚至想起他们都很难。所以,我希望自己已经抵达了那种死亡不再带来刺痛的涅槃境界。我并不为此担忧,但至少可以说,我已经能够坦然面对它的必然性。
第 1131 段
即使我还没有弄明白,当我想到孩子们失去我,或我失去他们,或失去我的妻子时,该如何消除其中的悲剧感。
第 1132 段
也许我已经在理智上接受了它的必然性,但对于失去所爱之人所带来的痛苦,我认为自己还没有理解其中关乎存在的那一面。比如,这一切终将结束。而且我不是以某种老套的方式来说这件事。我的意思是,感觉上它肯定不会结束。就像你在生活时,会觉得它永远不会结束。
第 1133 段
[马修] 对。
第 1134 段
而这束照耀着的光,这份意识,会在某个瞬间不复存在,也许就是今天。当我真正能够把这一切都装进脑中时,欧内斯特·贝克尔所说的那种恐惧便充满了我。那是一种真实的恐惧。我认为,人们并不总是坦诚面对它有多么可怕。我认为,你越能真正彻底地思考它,它就越可怕。这并不是一件如此简单的事情。哦,好吧,生命就是这样。
第 1135 段
如果你真的能把这一切装进脑中,那会很难。但我认为,这就是斯多葛主义者这样做的原因,因为这能帮助你振作起来,并让你觉得,好吧,此刻,你活着的每一个瞬间都无比美好。而它终将结束这件事非常可怕,就像你是一个无助的孩子,在寒冷中瑟瑟发抖,类似这样的感觉。
第 1136 段
然后,当你拥有温暖、拥有安全、拥有爱时,它会让你真正珍惜这一切。我觉得有时候,身处你的位置,当你提到铠甲时,只是不断看见死亡,可能会让你无法看见这一点,看不见生命的有限性,因为如果你一直盯着它看,它可能会把你压垮。
第 1137 段
所以,知道你仍在某种程度上为此挣扎是一件好事,既有那个神经外科医生,也有一个普通人。而这个普通人仍然能够为此挣扎,感受到对它的恐惧和它带来的痛苦。
第 1138 段
是的,这确实会让你问自己:还能持续多久,还能目睹多少次这样的事情?而不说:“我再也做不下去了。”不过,我的意思是,你说得很好。我认为,它给了你一个机会,让你纯粹地珍惜自己今天还活着。我有3个孩子和一位了不起的妻子,我真的很幸福。一切都很好。我能够参与帮助一个我认为很重要的项目。我认为它推动我们向前迈进。我是一个非常幸运的人。
第 1139 段
这是人类一次可能极其巨大的飞跃所迈出的早期步伐。这真的很有意思。而且很酷,因为你会读到历史上的所有这些事情,那些所谓的早期岁月。在前往亚马孙之前,我会阅读那些探险家的故事,他们会前去探索,甚至是第一次探索亚马孙丛林。那些就是早期的步伐。
第 1140 段
或者迈向太空的早期步伐,任何学科中的早期步伐,包括物理学和数学。而且这很酷,因为从宏观尺度来看,这些就像是深入探究人类大脑所迈出的早期步伐。所以,不只是观察大脑,还能够与人类大脑互动。它将帮助很多人,但也可能帮助我们理解那里面到底发生了什么。
第 1141 段
是的,我认为归根结底,我们希望给人们更多可以拉动的杠杆,对吧?也就是说,你想给人们提供选择。如果你能给某个人一个旋钮,让他们调节自己有多快乐,我认为这会让人们非常不安。但现在,谈谈重度抑郁症。谈谈这个国家里正以惊人速度自杀的人。然后从这个角度看,如果你能给人们一个旋钮,消除自杀意念和自杀意图,试着为那种不适感辩解吧。我会把那个旋钮给他们。我不知道你要如何为不这样做辩解。
第 1142 段
是的,你可以想象一下世界上正在发生的所有苦难。就像此刻每一个正在受苦的人,都会是一个发着红光的点。苦难越深,它就越亮。然后你会看到一幅人类苦难的地图,而任何能让你在大范围内调暗那苦难之光的技术都相当令人振奋,因为有很多人正在受苦,而他们中的大多数人都在默默承受。我们太常把目光移开了,而我们应该记住那些正在受苦的人,因为再说一次,他们中的大多数人都在默默承受。
第 1143 段
嗯,再从更宏大的尺度来看,社会的结构,人们对我们的社会结构如何运转或不运转、我们的政治如何运转或不运转有很多不满。而这些事物从总体上来说也是由神经化学构成的,对吧?
第 1144 段
就像我们的政治是由拥有人类大脑的个体构成的,而它运作或无法运作的方式,某种意义上可能是可以调节的,也就是说,我不知道,比如消除我们的成瘾行为,或者调节我们对社交媒体的成瘾行为,或是我们对愤怒的成瘾、对分享我们所能找到的最愤怒的政治推文的成瘾。我不认为这会带来一个正常运转的社会。
第 1145 段
而如果能为人们提供一些选择来缓和那种适应不良的行为,社会可能会从中获得巨大的益处。也许我们大家可以更和谐地合作,共同朝着有益的目标努力。
第 1146 段
这里存在一个最佳平衡点,就像你提到的,你不会想彻底消除人性中所有阴暗的方面,因为那些东西不知为何在某种程度上是让整个体系运转所必需的。但这里存在一个最佳平衡点。
第 1147 段
是的,我同意。我们得受一点苦,只是别苦到让你失去希望。
第 1148 段
是的。我们知道你做过的所有手术。你曾在里面看到过意识吗?那里有没有像是一道发光的亮光?
第 1149 段
我有种感觉,我从来没有找到过它。也从来没有像《哈利·波特》里的摄魂怪那样把它抽走。我有种感觉,意识远没有我们的本能想要声称的那么神奇。在我看来,要思考大脑中的意识是什么,一个有用的类比是:比如触碰你的皮肤,并知道哪里被触碰了,对这意味着什么,我们有非常好的直觉理解。
第 1150 段
我认为,意识不过是把那种层级的感觉映射应用到大脑本身的思维过程上。所以我的意思是,意识就是感知到你大脑的某个部分正在活动。因此你能感觉到它在工作。你能感觉到大脑中思考红色事物、有翼生物或咖啡味道的那些部分。你能感觉到大脑的那些部分正在活动,就像我现在能感觉到自己的手掌正被触碰一样,对吧?
第 1151 段
而那个感知大脑正在工作的感觉系统,就是意识。
第 1152 段
这太精彩了。道理是一样的,就像你触摸某个东西时产生的触觉。意识就是你感受到自己的大脑在运作、大脑在思考、大脑在感知的感觉。
第 1153 段
这并不是什么时空扭曲或某种量子场效应,对吧?它没有任何神奇之处。人们总想赋予意识某种真正与众不同的东西。有一段非常精彩而漫长的历史:人们总是试图用物理学中最新的发现来解释意识,因为它是你能想到的最神奇、最超乎寻常的东西。人们总想这样解释意识。我认为这没必要。它只是感受大脑运作的一种非常有用、令人满足的方式。
第 1154 段
而且正如我们所说,这可真是个了不起的大脑。
第 1155 段
[马修] 是的。
第 1156 段
我们在周围看到的一切、我们所爱的一切、一切美好的事物,都来自这样的大脑。
第 1157 段
这一切都是发生在你颅骨内的电活动。
第 1158 段
就我个人而言,我很感激有你这样的人,在探索它运作的所有方式,以及能够改善它的所有方式。
第 1159 段
谢谢你,莱克斯。
第 1160 段
非常感谢你今天与我交谈。
第 1161 段
这是一段愉快的经历。感谢收听这场与马修·麦克杜格尔的对话。现在,亲爱的朋友们,接下来是 Neuralink 的脑机接口软件负责人布利斯·查普曼。你告诉我,你见过数百名脊髓损伤或患有渐冻症的人,而你在 Neuralink 提供帮助的动力,源于想要帮助他们。你能描述一下这种动力吗?
第 1162 段
是的。首先,我只想感谢所有我有机会交谈过的人,感谢他们与我分享自己的故事。我认为,在任何情况下,我都不可能像他们自己那样有力地讲述他们的故事。
第 1163 段
但我想,只从非常宏观的层面总结一下,我一再听到的是,患有渐冻症或严重脊髓损伤、基本上已经无法再活动身体的人,归根结底真正寻求的是独立。而这对不同的人可能意味着不同的事情。
第 1164 段
对一些人来说,这可能意味着能够再次独立交流,不需要在脸上佩戴某种东西,也不需要护理人员把某种东西放进他们嘴里。
第 1165 段
对一些人来说,这可能意味着能够独立地重新工作,能够以足够高的效率在数字层面操作电脑,从而找到工作、养活自己、搬出去,并最终在家人也许已无法再照顾他们之后,依然能够养活自己。
第 1166 段
而对一些人来说,这只是能够及时回应自己的孩子,赶在孩子跑开或对别的东西产生兴趣之前。这些都是极其个人化、也可以说非常人性的问题。与这些人交谈时,一再令我感触深刻的是,这实际上是一个工程问题。这是一个只要拥有适当的资源和合适的团队,我们就能取得很大进展的问题。
第 1167 段
归根结底,我认为这是一条极具鼓舞力量的信息,也是一件让我每天都很期待起床去做的事。
第 1168 段
所以,一方面,这是一个关于脑机接口的工程问题,比如,它可以赋予他们与世界互动的能力。但另一方面,如何让世界其他部分对四肢瘫痪者而言更加无障碍,也是一个工程问题。
第 1169 段
是的,事实上,我想暂时用一种更宏观的视角来看待这件事。我非常支持任何在这个问题领域工作的人。所以,除了脑机接口之外,我也很高兴、很兴奋,并愿意尽我所能支持那些从事眼动追踪系统、语音转文字系统、头部追踪器、口控鼠标棒或 QuadStick 四肢瘫痪者控制器研发的人。我遇到过许多正是在做这些事情的工程师和社群人士。
第 1170 段
我认为,对于我们试图帮助的人来说,只要问题得到解决,解决方案有多复杂并不重要。我想强调的是,可能有许多解决方案能够帮助解决这些问题。脑机接口就是这类解决方案集合中的一种。具体来说,我认为脑机接口在这方面具有几项优势。
第 1171 段
我认为,立即意识到这一点的人,通常都是脊髓受伤或有某种形式瘫痪的人。通常,你不必向他们解释为什么这可能会有所帮助。这通常是相当显而易见的。
第 1172 段
但对于我们这些并未身患严重脊髓损伤、或不认识渐冻症患者的人来说,为什么会想要通过脑植入物来连接并操作电脑,往往并不明显。这件事出人意料地微妙复杂,而我仅仅通过在 Neuralink 的首次临床试验中与诺兰合作,并听他亲口讲述为什么这个设备对他影响重大,就学到了极多。
第 1173 段
这是一个微妙复杂的话题。比如,即便你能在操作电脑时用口控鼠标棒完成同样的事情,他也无法在一天中的每一分钟都使用那根口控鼠标棒。只有在有人能把它放到他面前时,他才能使用。
第 1174 段
因此,脑机接口确实能够提供一种独立和自主的程度;如果它不是实实在在地成为你身体的一部分,就很难通过其他任何方式达到这种程度。
第 1175 段
要让诺兰能够用意念控制屏幕上的光标,其中有许多引人入胜的方面。你给我发过一段我非常喜欢的话。你说:“我是负责面试并选出 P1 的团队成员之一。第一次人体手术期间,我就在手术室里监测从大脑传出的实时信号。我基本上每天都与用户合作,开发新的用户体验范式的解码策略。
第 1176 段
当信号质量下降时,我也是那个想出如何恢复实用脑机接口功能、使其达到新的世界纪录水平的团队的一员。”我想,我们会谈到其中的每一个方面,但先拉远一点看,成为那个团队的一员、参与那次历史性的,我会说,历史性的首次尝试,是什么感受?
第 1177 段
是的,我想对我来说,这是我近 10 年来一直为之兴奋的事情。因此,哪怕只是为它成为现实贡献一小部分力量,也令人极其兴奋。在整个过程中,有几个或许很特别、我永远都不会真正忘记的时刻,其中一个是在实际手术期间。到那个时候,我已经相当了解诺兰了。我也认识他的家人。
第 1178 段
所以我想,当诺兰被推进手术室时,最初的反应就是一种“糟了”的反应。但到了那一刻,肌肉记忆开始接管,你可以说就进入了那种状态,任由身体来完成一切。而在那场特定的手术中,我有幸承担的工作只是负责监测植入物。
第 1179 段
所以我的工作就是坐在那里,查看植入物传出的信号,查看当电极丝被插入大脑时从设备实时传出的脑数据,基本上就是观察并确保没有出现任何问题,也没有任何需要我们前去调查的危险信号或故障状况,或是需要暂停手术来调试的情况。
第 1180 段
而且,因为我是以那种旁观者视角观看手术的,所以我所处的视角,我觉得比房间里的大多数人稍微抽离一些。我可以坐在那里心想:“哇,那个大脑动得真厉害。”当你从侧面观察我们用来插入电极丝的开颅切口时,大多数人没有意识到的一件事是,大脑会动。你呼吸时、心脏跳动时,大脑都会大幅运动,而且你能清楚地看到。
第 1181 段
所以我觉得这件事令我很惊讶,也非常、非常激动,因为你能看到一个你现实中认识、并且与之深入交谈过的人的大脑,真的在他们的颅骨内搏动和移动。
第 1182 段
而且他们之前就是用那个大脑和你说话的,现在它就在那里动着。
第 1183 段
对。
第 1184 段
其实,我之前没意识到电极丝传输信号这方面的情况,所以 Neuralink 植入物在手术过程中是启用的,而且每次插入1根电极丝,你就能开始看到信号吗?
第 1185 段
是的。
第 1186 段
所以,这就是你们测试这东西是否正常工作的方法之一?
第 1187 段
是的,其实就在手术室里,我们差不多完成所有细线植入后,我就开始采集所谓的宽带数据。宽带数据基本上是能从 Neuralink 电极采集到的最原始的信号形式。它本质上是对局部场电位的测量,或者说,嗯,基本上就是该电极测得的电压。
第 1188 段
而且,我们的应用程序里有一种特定模式,可以让我们看到检测到的尖峰出现在哪里。所以,它算是把宽带符号中尖峰出现的位置可视化出来,而这是一种非常、非常原始的数据形式,表明某个神经元实际上正在产生尖峰。
第 1189 段
所以,在整个临床试验过程中,有一个时刻我永远不会忘记:在手术室里,他还处于麻醉状态时,我亲眼看到应用程序中显示出漂亮的尖峰,并且正实时传输到我手里拿着的一台设备上。
第 1190 段
所以这是未经信号处理的原始数据,然后上面叠加的是信号处理结果,你看到的是检测到的脉冲吗?
第 1191 段
[布利斯] 对,是的。
第 1192 段
而且那也是一种用户体验,因为————是的。
第 1193 段
那看起来也很漂亮。
第 1194 段
在那次手术过程中,房间里其实有很多摄影师。所以他们也很好奇,想要看看。房间里有几位神经外科医生,他们都很兴奋地看着机器人接替自己的工作,所有人都挤在一部小小的 iPhone 周围,看着实时脑数据从他的大脑中流出。
第 1195 段
看到机器人完成部分手术是什么感觉?也就是计算机视觉会检测出所有需要避开血管的位置,然后显然是在人工监督下,真正以极高精度将那些细线连接到大脑。
第 1196 段
这是个好问题。我的回答可能会很没劲,但那很无聊。我已经看过太多次了。是的,手术就应该是这样。你希望它很无聊,因为我已经看过太多次了。我真的看过机器人做手术几百次了,所以这只不过是再来一次。
第 1197 段
是的,所有那些练习手术和替代对象,而这只是平常的一天。
第 1198 段
对。
第 1199 段
那么 Noland 醒来时呢?你还记得这样一个时刻吗:他能够移动光标,不是移动光标,而是能从大脑中获取信号,从而证明连接已经建立?
第 1200 段
是的,是的。所以我们非常兴奋,想尽可能快地推进,而 Noland 也非常非常兴奋地想要开始。他其实想在手术当天就开始,但我们非常耐心地等到了第二天早上。所以那是漫长的一夜。第二天早上,他在重症监护室里康复时就想开始,真正着手了解我们能从他的大脑中测量到什么样的信号。
第 1201 段
也许对不熟悉 Neuralink 系统的人来说,我们会把 Neuralink 系统,也就是 Neuralink 植入物,植入运动皮层。运动皮层负责表征运动意图之类的东西,比如,如果你想象自己的手合拢再张开,这类信号表征就会出现在运动皮层中。
第 1202 段
如果你想象自己的手臂前后移动,或者晃动小拇指,这类信号可能会出现在运动皮层中。因此,我们开始梳理在某个特定个体的大脑中究竟能获取什么样的信号,其中一种方法是通过一项叫作身体映射的任务。身体映射本质上就是向用户展示一个视觉画面,然后对他说:“嘿,想象一下做这个动作。”
第 1203 段
那个视觉画面是一只不断张开、合拢的 3D 手,或者一根上下活动的食指。你让用户想象这个动作,显然你看不到他们真的这样做,因为他们瘫痪了,所以你看不到他们实际移动手臂,但在他们执行这项任务时,你可以记录神经活动,基本上可以进行离线建模并检查:我能否预测或检测到与这些不同动作相对应的变化?
第 1204 段
于是我们做了那项任务,并意识到,嘿,实际上有一些变化与他的某些手部动作相关。这是第一个迹象,表明,好吧,我们或许可以利用这种变化在现实世界中做一些有用的事情,比如控制电脑光标。我们第一次把它展示给他时,他就开始摆弄它了,而我们实际上只是把他大脑活动的同一个实时视图放到了他面前。
第 1205 段
我们说:“嘿,你来告诉我们发生了什么。我们不是你。你能够想象不同的事情,而我们知道这会改变其中一些神经元的活动,所以请你替我们弄清楚这实际上表征的是什么。”于是他摆弄了一会儿。他说:“我还没完全弄明白。”他又摆弄了一阵,然后说:“哦,当我移动这根手指时,我看到这个特定的神经元开始更频繁地放电。”
第 1206 段
我说,好吧,“证明给我看,再做一次。”于是他说:“好,三、二、一,砰。”就在他移动的那一刻,你可以看到这个神经元瞬间开始放电——单个神经元。如果你感兴趣,我可以告诉你确切的通道编号。它现在永远刻在我脑子里了。
第 1207 段
但那个单一通道的放电是一个很美妙的迹象,表明那确实是会随行为发生变化的神经活动,随后可以用于下游任务,比如解码电脑光标。
第 1208 段
你说单一通道时,那是与单个电极相关联的吗?
第 1209 段
是的,通道和电极可以互换使用。
第 1210 段
而这样的通道有 1,024 个?
第 1211 段
1,024 个,是的。
第 1212 段
这真的能奏效,太不可思议了。
第 1213 段
我在了解这一切、把这些知识装进脑子里时,简直觉得难以置信:意图,你可以在脑中想象自己移动手指,这能转化成一个信号;而且你随后还可以跳过这一步,想象光标移动,或者产生让光标移动的意图,而这会生成一个随后可以用来移动光标的信号。这里有太多令人兴奋的事情可以去了解,包括大脑以及大脑的运作方式。仅仅是存在可以利用的信号这一事实,就非常强大。
第 1214 段
但感觉这只是开始探索如何真正非常非常有效地利用这种信号。我还应该说,这里有太多引人入胜的细节了,但你提到了身体映射这一步。至少在我看到的、由 Noland 展示的那个版本里,有一个特别棒的界面,一个图形界面。
第 1215 段
但那感觉就像我置身于未来,因为它就像,你知道,我猜它会把你移动手的动作可视化。而且那个界面非常性感、精致。你好。我不知道里面有没有语音部分,但感觉就像你在一款制作精良的电子游戏中醒来,而这是那款游戏开头的教程。“这就是你应该做的。”很酷。
第 1216 段
不,我是说,未来就应该让人感觉像未来。
第 1217 段
但要做到这一点并不容易。我是说,它需要简单,但不能过于简单。
第 1218 段
是的,而且我认为,就 BCI 的整体开发而言,这里的用户体验设计部分被低估了。你向用户呈现指令的方式与能够获得的信号类型之间,存在完整的交互效应。而你的行为与神经信号之间这种匹配的质量,取决于你能多好地向用户表达你希望他们做什么。
第 1219 段
所以,是的,我们花了很多时间思考用户体验,思考如何构建应用程序、解码器实际如何运行,以及它向用户提供的控制界面。所有这些小细节都非常重要。
第 1220 段
所以,也许可以更详细地谈谈信号是什么样的,以及解码是什么样的。有一个 N1 植入物,正如我们提到的,它有 1,024 个电极,用于采集原始数据、原始信号。那个信号是什么样的?在它被传输之前,沿途有哪些不同的步骤?传输的又是什么?诸如此类的内容。
第 1221 段
对,是的。这会很有意思。我们开始吧。所以,在深入讨论我们所做的事情之前,值得先了解我们试图测量什么,因为这决定了我们所构建系统的许多要求。我们试图测量的,其实是单个神经元产生的动作电位。动作电位,你可以把它理解为一种微小的电脉冲,只要你离得足够近,就能检测到它。
第 1222 段
我说的足够近,是指比如在那个细胞的 100 微米以内。100 微米是非常非常小的距离。因此,任何一个给定电极能够捕捉到的神经元数量,仅限于该电极周围一小片半径范围内。关于这里的底层生物学,另一件值得了解的事情是,当神经元产生动作电位时,该动作电位的宽度约为一毫秒。
第 1223 段
所以,从尖峰开始到尖峰结束,神经元放电这种特征的整个宽度是一毫秒。如果你想检测单个尖峰是否正在发生,就需要以远高于每毫秒一次的频率,对该信号或该神经元附近的局部全电位进行采样。
第 1224 段
你需要在每毫秒内进行很多很多次采样,才能检测出这实际上是神经元产生动作电位时的特征波形。因此,我们在全部 1,024 个电极上以每秒约 20,000 次的频率采样。每秒 20,000 次意味着,我们已经给出了一个一毫秒的窗口。我们有大约 20 个样本,可以告诉我们那个动作电位的确切形状是什么样的。
第 1225 段
一旦我们以超高频率对这些细胞附近的底层电场进行了采样,就可以把那个信号处理成:我们在哪里检测到尖峰,或者在哪里没有检测到,也就是一种一或零的二进制信号。在这一毫秒内,我们是否检测到了尖峰?我们之所以这样做,是因为神经活动中实际承载信息的那部分子空间,就只是尖峰何时发生。
第 1226 段
从本质上说,我们在解码方面关心的一切,都可以通过脉冲序列的频率特征来捕捉或表征,也就是在任意给定时间窗口内,脉冲放电的频率是多少。
第 1227 段
因此,这让我们能够进行某种程度上惊人的压缩,把这种非常丰富、高密度的信号压缩成稀疏得多、也更易压缩的东西,从而可以通过无线电发送出去,比如蓝牙通信。
第 1228 段
这里快速岔开一下。你提到了电极、神经元。附近有一个神经元的局部邻域。要分辨尖峰来自哪里有多困难?
第 1229 段
对,所以有一整个学术神经科学研究领域,专门研究这个问题:基本上,在给定一个电极,或者给定一组测量一组神经元的电极时,如何进行某种分类,也就是尖峰分类,以确定哪些尖峰来自哪个神经元?学术研究之所以探索这个问题,是因为要理解大脑底层神经科学中正在发生什么,就需要关注它。
第 1230 段
如果你关心的是理解大脑如何表征信息,以及这种表征如何随时间演变,那么弄清这个问题就非常、非常重要。对于工程方面,至少在目前的规模下,如果每个电极对应的神经元数量相对较少,基本上可以完全忽略这个问题。
第 1231 段
你可以把它看作神经元到电极的一种随机投影,在某些情况下,一个电极可能对应不止一个神经元。但如果这个数量足够少,就可以把这些信号视为两者的某种并集。对于很多应用来说,这是一种完全合理的权衡,而且能大幅简化问题。
第 1232 段
而随着通道数量不断扩大,区分单个神经元的重要性会降低,因为你拥有了更多的整体信号,并且可以开始依靠数据中的相关性或协方差结构,来帮助理解这个通道放电时实际上代表什么。因为你知道,当这个通道与另外这50个通道协同放电时,意思是向左移动。
第 1233 段
但当同一个通道与另外这10个通道协同放电时,意思就是向右移动。
第 1234 段
好,所以你必须在设备上进行这种尖峰检测,而且必须极其高效,也就是速度快,并且不能耗费太多功率,因为你不希望产生太多热量。所以它必须是一个极其简单的信号处理步骤。
第 1235 段
[布利斯] 对。
第 1236 段
关于克服这项挑战需要做些什么,你有什么经验可以分享吗?
第 1237 段
对,所以我们尝试过很多不同的方案,基本上就是把这种原始信号转化成某种你可能想从设备发送出去的特征。而且我想说,我不认为我们已经走到了这个过程的最后一步。这是一段漫长的旅程。我们今天显然已经有了可行的东西,但未来可能还会发现很多远优于我们目前做法的方法。
第 1238 段
所以,我们目前的一些做法,在学术界都有很多传承,因此我不想声称这些是什么由 Neuralink 原创的想法之类的。但其中一个想法,基本上是构建某种近似卷积滤波器的东西,可以这么说,它会沿着信号滑动,寻找与某个特定模板相匹配的部分。
第 1239 段
而这个模板包括尖峰向下调制得有多深、恢复了多少,以及整个过程所用的持续时间和时间窗口。如果你能在信号中看到该模板在一定范围内得到匹配,那么就可以说:“好,这是一个尖峰。”
第 1240 段
这种方法极其方便的一个原因是,你实际上可以在硬件中极其高效地实现它,这意味着你可以用低功耗同时在1,024个通道上运行它。我们最近开始探索的另一种方法,可以与尖峰检测方法结合使用,叫作尖峰频带功率。
第 1241 段
这种方法的好处是,你或许能够捕捉到一些神经元的信号,这些神经元可能距离太远,无法被检测为尖峰,因为神经元离电极越远,实际的尖峰波形在该电极上看起来就越微弱。
第 1242 段
因此,你或许能够捕捉到一些略微超出正常记录半径的群体层面活动,也就是神经科学家有时所说的活动杂波,即正在发生的其他活动;你可以观察许多通道中这种背景噪声如何变化,并且或许能以这种方式从信号中榨取更多信息。但这是有代价的。
第 1243 段
这个信号现在是一种浮点表示,这意味着将它发送出去需要耗费更多功率。这意味着你必须寻找不同的压缩方法,与能够应用于二进制信号的方法不同。所以,这些不同模态都伴随着许多不同的挑战。
第 1244 段
所以在通信方面,你同样受到可发送数据量的限制。
第 1245 段
[布利斯] 对。
第 1246 段
另外,因为你目前使用的是蓝牙协议,所以必须把数据批量组合在一起。但与此同时,你还必须把延迟保持在低得离谱的水平。真的低得离谱。关于延迟,有什么可以谈谈的吗?
第 1247 段
对,这是我倾注热情的一个项目,所以我想打造世界上最好的鼠标。我不想造那种,你知道的,电动汽车里的雪佛兰 Spark 之类的东西。我想造鼠标中的 Tesla Roadster。而且我确实认为,在5到10年内,大多数电子竞技比赛很可能都会由瘫痪人士主导。出于很多原因,这是一种非常现实的可能性。
第 1248 段
一个原因是,他们将能使用最好的技术来有效地玩电子游戏。第二个原因是,他们有时间这样做。所以这两个因素结合起来,对电子竞技选手尤其有利。
第 1249 段
除非也允许没有瘫痪的人植入。
第 1250 段
(笑)对。
第 1251 段
这确实是与数字设备交互的另一种方式。如果它是一种截然不同、效率更高的体验,那是有意义的。即使它不是什么彻底的高带宽通信,如果只是能让鼠标移动速度提高10倍,比如每秒比特数,如果我能实现某个每秒比特数,是我用鼠标能做到的10倍,那么它能带来什么就是一种非常有意思的可能性,尤其是在你通过训练变得非常熟练之后。
第 1252 段
你的性能上限肯定会更高,因为你不必让自己的意图经过手臂、经过肌肉来缓冲。仅仅因为植入了脑部植入物,你在实际尝试执行的任何动作上就能获得大约75毫秒的提前量。
第 1253 段
这其中还有一些细微之处,比如有证据表明,运动皮层可以通过某种方式预先规划动作序列,因此你可能无法一直获得全部优势。但对于某种反应时间类的游戏,你只想要,比如有人在这边,就把他狙掉,诸如此类。你的确拥有一种固有优势,因为不需要经过肌肉。所以问题只是,你究竟能把它加快多少?
第 1254 段
而且从延迟角度来看,我们已经比你通过肌肉时会做的更……而我们还处于这方面的早期阶段。我认为我们可以进一步压低延迟——我们目前从脑部脉冲到光标移动的端到端延迟大约是22毫秒。如果你想想世界上最好的鼠标、最好的游戏鼠标,它们的延迟大约是5毫秒左右,具体取决于你如何测量。
第 1255 段
这取决于你的屏幕刷新有多快,其中有很多特性都会产生影响。不过,对,而且大脑中的一个神经元实际影响你对手部的控制,所需的大致时间约为75毫秒。所以如果看看这些数字,就会发现我们已经具备竞争力,而且比实际移动手部所能达到的速度略快。
第 1256 段
而这件事,如果你问Noland,当他第一次移动光标时,我们就问过他这个问题。这是我特别好奇的一件事,比如当你调节点击意图时,或者当你只是试图把光标向右移动时,那是什么感觉。
第 1257 段
他说,光标会在他像是真正产生移动意图之前就动起来,这有点超现实,也是我希望有一天能亲自体验的事情。那会是什么感觉,那东西如此即时、如此流畅,以至于感觉它在你真正打算让它移动之前就已经动了。
第 1258 段
是的,我想我们已经习惯了那种延迟,那种自然发生的延迟。所以目前的瓶颈是通信吗,比如蓝牙通信,实际的瓶颈是什么?我的意思是,总会有瓶颈。目前的瓶颈是什么?
第 1259 段
是的,有几个方面。所以有点好笑的是,低功耗蓝牙协议对通信速度有一些限制。协议本身制定了一个标准,即你能发送的最频繁的这类更新大约是每7.5毫秒一次。
第 1260 段
随着我们把延迟压低到单个脉冲就能影响控制的程度,达到那种分辨率,这类协议在某个尺度上将成为限制因素。这里另一个相当重要的细微之处是,这个等式涉及的不只是Neuralink本身。如果你开始把延迟压到低于屏幕刷新速度的程度,那么你就会遇到另一个问题。
第 1261 段
比如,你的整个系统都需要能够达到技术极限所能提供的那种响应速度。比如你需要屏幕——如果你想让某个东西以1毫秒这个量级作出响应,那么120赫兹就已经不够用了。
第 1262 段
这是个非常酷的挑战。我也喜欢把“世界上最好的鼠标”印在T恤上。跟我讲讲接收端,也就是解码步骤:现在我们已经弄清了这些脉冲是什么,把它们全部汇集到一起,现在正把它们发送到应用程序。解码步骤是什么样的?
第 1263 段
是的,那么也许先说说,什么是解码?我想,可能有很多听众完全不知道解码大脑活动是什么意思。
第 1264 段
其实,即使我们再退一步来看,那个应用程序是什么?所以,有一个植入物会与任何安装了该应用程序的数字设备进行无线通信。那么你能不能从高层次上告诉我,这个应用程序是什么,大脑之外的软件是什么?
第 1265 段
是的,那么也许从目标倒推,目标是帮助瘫痪者,在这个案例中是 Noland,让他能够独立操作自己的电脑。我们认为,实现这一目标的最佳方式,是为他们提供我们用来操作软件的同样工具,因为我们不想为了大脑而重建整个软件生态系统。至少现在还不想。也许有一天,你可以设想会有为 BCI 原生构建的用户体验。
第 1266 段
但就当下对人们有用的东西而言,我想大多数人会更希望能够直接控制鼠标和键盘输入,从而使用他们日常工作、与朋友交流等所需的所有应用程序。因此,这个应用程序的工作,实际上就是把从植入物传来的这股无线大脑数据流转化为对电脑的控制。
第 1267 段
我们实现这一点的方式,本质上是建立一种从大脑活动到实际硬件的某种 HID 输入之间的映射。HID 就是用于传达输入设备事件的协议。例如,把鼠标移动到这个位置,或者按下这个按键。因此,这种映射从根本上说就是该应用程序负责的事情。
第 1268 段
但这种映射如何运作,其中有很多细微之处,我们花了大量时间试图把它做好;而在这段弄清楚如何以最佳方式做到这一点的漫长旅程中,我们仍处于早期阶段。这个过程的一部分就是解码。所谓解码,就是接收通过蓝牙连接传输到应用程序的大脑数据中的统计模式,并把它转化为例如鼠标移动的过程。
第 1269 段
这个解码步骤,你可以把它看作几个不同的部分。和任何机器学习问题类似,其中有训练步骤,也有推理步骤。在我们的案例中,训练步骤是一个非常精细复杂的行为过程,用户必须想象自己正在执行不同的动作。例如,他们会看到一个带有光标的屏幕,并被要求把那个光标推向右边。
第 1270 段
然后想象把那个光标推向左边,向上推,向下推;这样,我们基本上就可以建立一种模式,或者说,使用任何一种现代机器学习方法,建立一个映射:给定这些大脑数据和那个想象中的行为,把一方映射到另一方。然后在测试时,你使用同一个模式匹配系统。
第 1271 段
在我们的案例中,它是一个深度神经网络;你运行它,接收从他们的植入物传来的实时大脑数据流,通过与校准时看到的内容进行模式匹配来解码,然后用它来控制电脑。
第 1272 段
现在,有几个我认为相当有意思的岔题,其中一个与你如何构建那个最佳的模板匹配系统有关,因为在与瘫痪者合作时,会遇到各种行为方面的挑战,也有调试方面的挑战。因为还是那句话,从根本上说,你观察不到他们试图做什么。你看不到他们尝试移动自己的手。
第 1273 段
所以,你必须想出一种办法,指导用户做某件事,并验证他们做得是否正确,这样之后你才能有把握地构建神经脉冲与预期动作之间的映射。而我所说的正确执行动作,真正指的是达到神经元活动这一层面的分辨率。
第 1274 段
所以,在理想世界里,如果你能获得一个行为意图信号,它在大约一毫秒的分辨率尺度上具有真实标准级别的准确性,那么我就能很有把握地建立从我的神经脉冲到那个行为意图之间的映射。但挑战仍然在于,你观察不到他们实际在做什么。
第 1275 段
所以,如何构建一种用户体验,其中有很多细微之处;它给你的不只是某种关于用户打算做什么的、平均而言正确的课程。如果你想打造世界上最好的鼠标,你真的会希望它尽可能灵敏。
第 1276 段
你希望它能够在沿途每个类似的步骤中,准确完成用户的意图,而不只是在你试图把它从左边移到右边时,平均而言是正确的。构建一种能提供这种分辨率的行为校准游戏,或者某种软件体验,正是我们投入大量时间研究的事情。
第 1277 段
所以在校准过程中,界面必须鼓励精准,意思是说,无论它做什么,都应该让人极其直观地觉得,这个人接下来很可能要做的事情,恰好就是你需要的那个意图,而且只有那个意图。而且除了他们事后也许会告诉你自己实际做了什么之外,你没有任何反馈。你不能,“哦,对。”
第 1278 段
对。
第 1279 段
所以从根本上说,这确实是一个非常令人兴奋的用户体验挑战,因为这全都取决于用户体验。它不只是要友好、令人愉悦或易用。
第 1280 段
是的。
第 1281 段
这就像——用户体验就是它如何运作。
第 1282 段
就是它如何运作。
第 1283 段
是的。
第 1284 段
对于校准而言,而且至少在 Neuralink 的现阶段,校准是这个东西运行的基础,并且不只是校准,实质上还包括持续校准。
第 1285 段
[Bliss] 是的。
第 1286 段
哇,是的。
第 1287 段
你刚才说了一点,我觉得值得在这里稍微探讨一下。
第 1288 段
你说这主要是一个用户体验方面的挑战,我认为其中很大一部分确实如此,但这里也存在一个非常有趣的机器学习挑战:给定某个数据集,其中包括一些平均而言正确的行为,比如要求用户向上、向下、向右、向左移动,再给定一个神经脉冲数据集,是否有办法以某种半监督或完全无监督的方式,推断出他们意图的那种高分辨率版本?
第 1289 段
仔细想想,可能确实有,因为数据集中有足够多的数据点,你的模型也受到足够多的约束,因此,通过恰当的某种形式,应该有办法让模型自己弄清楚。例如,在这一毫秒,他们向上推的力度究竟有多大。而在这一毫秒,他们正试图以多大的力度向上推。
第 1290 段
拥有非常干净的标签确实很重要,是的。所以,如果标签有噪声,从机器学习的角度来看,问题就会变得困难得多。
第 1291 段
[Bliss] 没错。
第 1292 段
然后,要获得干净的标签,这是一个用户体验方面的挑战。
第 1293 段
没错,不过说到干净的标签,我认为也许值得探究一下它究竟是什么意思。我认为,任何一种给定的标注策略都会对用户试图做什么作出一些假设。这些假设可以表述为损失函数,也可以表述为你可能采用的启发式方法,用来尝试估计或猜测用户试图做什么。
第 1294 段
而真正重要的是这些假设有多准确。例如,你可能会说:“嘿,用户,向上推动,并跟随这个光标的速度。”而你的启发式假设可能是,他们正试图精确地做那个光标试图做的事。另一种与之竞争的启发式假设可能是,他们实际上试图在动作开始时稍微快一点,在结束时稍微慢一点。
第 1295 段
而这些相互竞争的启发式假设可能准确反映了用户试图做什么,也可能没有。这个任务的另一个版本可能是:“嘿,用户,想象一下将这个光标移动一个固定的偏移量。所以不要跟随光标,只需尝试将它正好向右移动 200 像素。”那么,这是光标,这是目标。好了,光标消失了。现在尝试把那个已经不可见的光标向右移动 200 像素。
第 1296 段
在那种情况下,假设会是用户确实能够正确调节那个位置偏移量,但这种位置偏移量假设可能是一种较弱的假设,因此与那些试图在每一毫秒猜测用户想做什么的启发式方法相比,你或许可以让它更加准确。
第 1297 段
所以,你可以设想不同的任务,它们对用户意图的性质作出不同的假设,而这些假设是正确的,就是我所理解的干净标签。
第 1298 段
对于那一步,我们应该想象什么?有一个光标,而你想把那个光标向右或向左移动,或者上下移动,又或者把它们移动某个偏移量。所以这是一种方法,这是进行校准的最佳方法吗?比如,另一种疯狂的方法——它可能也在这里发挥作用——是像 Webgrid 这样的游戏,你可以获得非常大量的数据,由一个人在玩游戏;如果他们处于心流状态,也许你可以附带获得干净的信号。
第 1299 段
[Bliss] 对。
第 1300 段
那不是一种有效的初始校准方法吗?
第 1301 段
是的,很好的问题。这里有很多内容需要展开讲。所以,我首先会区分一种开环、第一闭环。所谓开环,我的意思是用户算是从零到一。他们完全没有模型,而他们正试图达到这样一种状态:至少拥有某种程度的控制。
第 1302 段
在那种设置下,你确实需要安排某种任务,给用户一个提示,让他们知道你希望他们做什么,这样你才能重新建立从大脑数据到输出的这种映射。然后,一旦他们有了一个模型,你可以设想他们使用这个模型,实际去适应它,自己摸索出使用它的正确方式,然后再用那些数据重新训练,从而让性能得到某种提升。
第 1303 段
这两种技术都伴随着许多挑战,如果你感兴趣,我们可以深入探究其中任何一种。但开环任务所面临的一个挑战是,用户自己无法获得关于其操作的本体感觉反馈。当他们使用一个开环……当他们尝试进行开环校准时,他们未必能感知到自己,或感觉到手下的鼠标。
第 1304 段
他们被要求执行某种操作。比如,想象一下,你的整条右臂都被麻醉了,你把它伸进一个箱子里,而且看不见它。因此,你既没有视觉反馈,也没有关于手臂位置或活动情况的本体感觉反馈。现在有人要求你,好,根据屏幕上这个从左向右移动的东西,匹配它的速度。
第 1305 段
基本上,你可以竭尽所能,在大脑中唤起任何一种让光标从左向右移动的想象动作。但无论如何,你都不会准确,而且执行这项任务的方式也可能不一致。所以,这差不多就是开环的根本挑战。
第 1306 段
闭环的挑战在于,一旦用户获得了一个模型,并且能够开始自行移动鼠标,他们就会非常自然地适应那个模型。而模型学习他们在做什么、用户学习如何使用模型,这两者之间的共同适应过程,可能无法为你找到最好的那种全局极小值。
第 1307 段
也许你的第一个模型在某些方面噪声很大,或者可能只是存在某种怪癖,比如数据分布中有某一部分覆盖得不是特别好,而用户现在摸索出了办法,因为他们是像 Noland 那样聪明绝顶的用户。他们找到了正确的想象动作序列,或者为了让它正常工作,他们必须以正确的角度保持手部姿势。
第 1308 段
他们会让它运行得非常好,但到了第二天,他们重新使用设备时,也许已经无法准确记得前一天用过的所有诀窍。因此,这里可能会形成一种复杂的反馈循环,让调试过程变得非常、非常困难。
第 1309 段
好,这里面有很多非常精彩的内容。对,其实,继续谈闭环,我见过一些情况,这其实发生在我观察心理学研究生的时候。他们不会自己编程时,就会使用某个软件。他们使用别人编写的软件,而那个软件有一大堆漏洞。然后他们摸索出了,比如说,而且他们已经用了很多年。他们摸索出了绕过这些问题的方法。
第 1310 段
哦,事情就是这样发生的。就好像没有人会考虑,也许我们应该修复这个问题。他们只是去适应。这是一个非常有意思的概念:我们确实很擅长适应,但你仍然需要,那可能并不是最优的。好,那么你要如何解决这个问题?是不是每隔一段时间就必须从头重新开始之类的?
第 1311 段
对,这是个好问题。首先,我要说这并不是一个已经解决的问题。对于学术界任何正在收听、从事 BCI 研究的人,我还要说,这不是一个只要扩大通道账户规模就能解决的问题。也许这会有所帮助,你可以得到某种更丰富的协方差结构,在设法制定良好的标注策略时加以利用。
第 1312 段
但如果你对那些本质上无法通过扩大通道账户规模来解决的问题感兴趣,这就是其中之一。对,那么要如何解决呢?这不是一个已经解决的问题。这是我首先想确保传达清楚的一点。第二点是,任何涉及闭环的解决方案都会变成一个非常困难的调试问题。
第 1313 段
而我在选择要处理哪些提示词时,有一条大致的通用启发法,就是你应当选择最容易调试的那个,因为如果你能做到这一点,即便性能上限较低,你也能推进得更快,因为你在调试问题时拥有一个更紧密的迭代循环。而在开环设置中,并不存在一个让用户参与其中的反馈循环调试。
第 1314 段
因此,有理由认为这应该是一个更容易调试的问题。另一件值得理解的事情是,即使在闭环设置中,也不存在什么特殊的软件魔法,能够推断出用户真正试图做什么。在闭环设置中,尽管他们正在移动屏幕上的光标,但他们试图做的事情可能与模型输出的内容不同。
第 1315 段
所以,如果你希望能够进一步改进模型,模型的输出并不是一种可以用来重新训练的信号。要弄清楚那个信号背后真正的用户意图是什么,你仍然面对着这个非常复杂的猜测或无监督问题。因此,开环问题有一个很好的特性,就是容易调试。
第 1316 段
它的第二个优点是,它拥有与闭环场景完全相同的信息和内容。我还想提到并指出另一件事:要为人们提供实用的控制能力,并不需要先解决这个问题。即使在今天,凭借我们现有的解决方案,以及学术界经过数十年积累建立起来的方案,如今可以赋予用户的控制水平也相当实用。
第 1317 段
要达到那种控制水平,并不需要解决这个问题。但同样,我想打造世界上最好的鼠标。我想把它做得足够好,好到你是否想要它根本都不再是个问题。而要打造世界上最好的鼠标,也就是超人版本,你确实需要彻底解决那个问题。
第 1318 段
我们以前在内部做过一些研究,其中也许有几个细节,我认为在思考如何解决这个问题时,理解这些细节非常有意思。第一点是,即使你拥有关于用户试图做什么的真实值数据,而你可以通过一只身体健全的猴子获得这些数据,也就是一只植入了 Neuralink 设备、正在移动鼠标控制计算机的猴子,即便拥有那样的真实值数据集,结果表明,要生成高性能 BCI,最佳预测对象也不只是对鼠标的直接控制。
第 1319 段
你可以设想建立一个数据集,记录大脑中正在发生什么,以及鼠标在桌面上究竟在做什么。结果表明,如果你建立从神经脉冲到准确预测鼠标行为的映射,那么这个模型的表现会比另一个模型更差,后者经过训练,预测的是关于用户可能试图做什么的某种更高层次假设。
第 1320 段
例如,假设猴子正试图沿直线移向目标,结果表明,作出这些假设在生成模型方面,实际上比预测底层的手部运动更有效。
第 1321 段
所以是意图,而不是什么身体动作之类的。
第 1322 段
对。
第 1323 段
这两者之间显然有非常强的相关性,但意图是一个更值得追寻的东西。
第 1324 段
[Bliss] 对。
第 1325 段
嗯,那也非常有意思。我的意思是,意图本身就很迷人,因为是的,在这里使用 BCI,在这种情况下,使用数字心灵感应时,你作用于意图,而不是动作,这就是为什么会有一种感觉,好像事情在你打算让它发生之前就已经发生了。这太酷了。也正因如此,就鼠标控制而言,你可以实现类似超人水平的表现问题。
第 1326 段
所以,关于开环,再澄清一下,每当要求这个人把鼠标向右移动时,你说没有反馈,所以他们无法获得那种真的让它移动了的满足感,对吧?
第 1327 段
你可以设想在屏幕上向用户提供反馈,但这很困难,因为在这个时候,你不知道他们正试图做什么。那么,你能向他们展示什么,才能基本上给他们一个信号,告诉他们我这样做是正确还是不正确的。我们来看这个非常具体的例子。比如,你的校准任务或许是尝试将光标移动某个特定的位置偏移量。
第 1328 段
所以,你给用户的指示是:“嘿,光标在这里。现在,当光标消失时,想象把它从原来的位置向右移动 200 像素,使其位于这个目标上方。”在这种情景中,你可以设想设计出某种一致性指标,向用户显示:“好吧,我知道当你执行这个向右的动作时,脉冲序列平均是什么样子。
第 1329 段
也许根据最近一次试验或你想象的轨迹,我可以生成某种概率估计,判断你采取的动作有多大可能是那个动作。”这可以向用户提供某种反馈,让他们知道自己在不同试验中的表现有多一致。
第 1330 段
你也可以设想,如果向用户展示这种一致性指标,也许他们从一开始就会在行为上更加投入,因为当你完全得不到任何反馈时,这项任务会有些无聊。所以,即使屏幕上显示的东西并不准确,它也可能改善用户体验,仅仅因为它能让用户保持动力,尝试提高那个数字或将其推高。
第 1331 段
所以这里面存在一个心理学因素。
第 1332 段
是的,完全正确。
第 1333 段
而且再说一次,所有这些都是用户体验方面的挑战。信号漂移有多大,是每小时、每天、每周、每月怎样变化的?由于信号漂移,你需要多频繁地重新校准?
第 1334 段
是的,所以这是我们一直在研究的一个问题,既包括在临床试验前用 NHP,也就是非人灵长类动物进行研究,也包括在临床试验期间与 Noland 一起研究。或许首先值得说明的是,我们在这里的目标是什么。我们的目标其实是让用户获得即插即用的体验——不过我想他们并不需要插入任何东西——而是一种随时想用就能用、想怎么用就怎么用的使用体验。
第 1335 段
而这正是我们的目标。因此,可以有一套解决方案,在不考虑这个非平稳性问题的情况下达到那种状态。所以,这里或许第一个重要的解决方案是,他们可以随时重新校准。Noland 如今已经能够这样做了。
第 1336 段
所以,他可以在凌晨 2:00、夜深人静的时候重新校准系统,无需护理人员、父母或朋友在身边帮他按按钮。解决方案中另一个重要部分是,当你校准出了一个良好的模型后,可以继续使用它,而不需要重新校准。所以,他现在需要多频繁地进行这种重新校准,其实取决于他对性能的要求。
第 1337 段
我们观察到,随着时间推移,任何单个模型的运行效果都会出现某种程度的下降,但用户可以通过调整自己的控制策略,在行为层面缓解这种情况。也可以通过我们向用户提供的一系列软件功能来缓解。例如,我们允许用户精确调整光标移动的速度。
第 1338 段
例如,我们称之为增益,也就是光标对任何给定输入意图作出反应的速度增益。他们还可以调整平滑度,也就是该光标意图的输出实际上有多平滑。他们也可以调整摩擦力,也就是让光标停下并保持静止有多容易。所有这些软件工具都为用户提供了很大的灵活性和故障排查机制,让他们能够自行解决这个问题。
第 1339 段
顺便说一句,所有这些操作都是通过看向屏幕右侧、选择混音器来完成的,而这个混音器里有——这就像 DJ 模式。你的 VCI 的 DJ 模式。
第 1340 段
所以我的意思是,这是一个做得非常好的界面。真的做得非常、非常好。然后,是的,Noland 在一次直播中谈到过那种偏置,也就是光标漂移,不过他说你们当时只是和他一起摆弄这个东西,然后不断改进。所以那可能只是那个特定时刻、某一天的一个快照。但他说,确实存在这种光标漂移和这种偏置,而他自己可以消除它,我想是通过看向屏幕右侧、屏幕左侧,来对偏置进行某种调整。
第 1341 段
对,对。
第 1342 段
我想,那是一种用来调整偏置的界面操作。
第 1343 段
是的,所以这其实是一个源自学术界的想法。此前有一些针对 BrainGate 临床试验参与者的研究,他们开创了这种偏置校正的理念。
第 1344 段
我认为我们的实现方式,怎么说呢,原型化程度很高,用户体验也非常漂亮:用户基本上可以将光标快速移到屏幕边缘,随即打开一个窗口,在那里实际调整或微调光标的确切偏置。所以,对于不熟悉这个概念的人来说,偏置或许就是当你什么都不去想象时,光标默认的运动方式。
第 1345 段
事实证明,这是最先受到神经非平稳性影响的光标控制体验感质之一。
第 1346 段
脱离光标体验的质量。
第 1347 段
我不知道还能怎么描述它。我不是那个移动——这非常有诗意,我喜欢。“光标体验的质量。”是的,我的意思是,这听起来很有诗意,但它确实非常真实。这里确实存在一种体验:当它运行良好时,那是一种令人愉悦的、真正令人舒适的体验。而当它运行不佳时,那是一种非常令人沮丧的体验。这其实就是 UX 的艺术。
第 1348 段
就像你既有可能让人感到沮丧,也有可能给他们带来快乐——归根结底,确实如此:UX 就是这个东西如何运作。所以,它不只是屏幕上显示了什么,还包括一次解码向用户提供了哪些控制界面?就像,我们希望他们觉得自己坐在 F1 卡片里,而不是某种类似小型厢式车的东西里,对吧?我们确确实实就是这样看待它的。
第 1349 段
Noland 本人是 F1 车迷,所以我们把自己称为维修站团队。他确确实实是那位 F1 车手,而不同种类的汽车和飞机会为用户提供不同的控制界面。在设计光标应当如何运动时,我们从中汲取了许多灵感……
第 1350 段
或许这里还有一个细微之处:即使是这样的细节——当你在 MacBook 触控板上移动鼠标时,你对触控板施加的输入如何转化为光标移动,其响应曲线也不同于使用鼠标时的运作方式。当你在触控板上移动时,动作以多大程度转化为计算机输入,有一种不同的响应函数、不同的曲线,不同于你用鼠标进行实体操作时的情况。
第 1351 段
这是因为很久以前,有人在设计任何计算机最初的输入系统时坐下来,仔细思考了使用这些不同系统时的确切感受。如今,我们正在设计这种计算机输入系统的下一代,它完全通过大脑来操作,而且没有本体感觉反馈。
第 1352 段
再说一次,你感觉不到手里的鼠标,也感觉不到指尖下的按键,而你需要一个仍能让用户轻松、直观地理解系统状态以及如何实现自己目标的控制界面。最终的目标是,让这种 UX 完全地、逐渐隐入背景。
第 1353 段
它会变得如此自然、如此直观,以至于对用户而言成为一种下意识行为;他们应该只会觉得自己基本上能够直接控制光标。它就是会按照他们的意愿行动。他们不会去思考为了让它按照自己的意愿行动,具体实现方式是什么。它就是在按照他们的意愿行动。
第 1354 段
有没有某种类似菲茨定律的东西,也就是说,你应该以某种特定方式移动鼠标,从而最大限度提高命中目标的概率?我甚至都不知道自己在问什么,但我希望我这个问题的意图能够落到一个深刻的答案上。不,我是说,在有人使用大脑进行控制的情境下,对于 UX 的规律是否存在某种理解?比如说,它和实际使用鼠标时有所不同?
第 1355 段
我认为我们还处在发现这些规律的早期阶段,所以我不会声称自己已经解决了那个问题。但我们确实学到了一些能让用户更容易完成事情的要点。当你把它们说出来时,它们相当直截了当,但当你身处一线调试这些东西的过程中时,实际上要花一段时间才能走到那一步。
第 1356 段
其中一点是,你构建的任何机器学习系统都会有一定数量的错误,而这些错误如何传导至下游用户体验非常重要。例如,如果你正在为照片开发一种搜索算法,当你搜索朋友 Joe 时,它却调出了你朋友 Josephine 的照片,那或许不是什么大问题,因为错误的代价并不高。
第 1357 段
在另一种场景中,比如你试图检测保险欺诈之类的事情,并因为某个机器学习模型的输出而直接将某人送上法庭,那么对错误保持谨慎就合理得多。你需要非常审慎地思考这些错误会如何转化为下游影响。脑机接口也是如此。
第 1358 段
举个例子,如果你构建的模型要从大脑中解码速度输出,而另一种输出则是你试图调控鼠标左键点击。这两者在对精确度的要求上有着不同的权衡,也就是要精确到什么程度,才会对最终用户有用。对于速度而言,平均而言正确就可以,因为模型的输出会随时间进行积分。
第 1359 段
所以,如果用户试图点击位置 A,而他们当前位于位置 B,他们会尝试随时间移动,从而在这两个点之间完成导航。只要模型的输出平均而言是正确的,他们就能在用户控制回路参与其中的情况下,随时间对其进行某种操控。他们可以到达自己想去的点。但点击并非如此。
第 1360 段
对于点击来说,以神经元放电的时间尺度看,你几乎是在瞬间完成它。因此,你要非常确定那次点击是正确的,因为一次错误点击可能会给用户带来很大的破坏。他们可能会意外关闭自己正试图在其中做某件事的标签页,丢失全部进度。他们也可能会意外按下某段文字的发送按钮,而那段文字只写了一半,之后读起来很奇怪。
第 1361 段
所以,在这个领域,错误对应着不同类型的成本函数。而用户体验设计的一部分,就是要弄清楚如何构建一种即使出错,对最终用户仍然有用的解决方案。
第 1362 段
为错误发生时的每个动作分配成本,这一点太有意思了。所以,每个动作一旦发生错误,都有特定的成本,而在解读意图、将其映射到动作时把这一点纳入考虑,真的非常重要。在你说出来之前,我并没有完全意识到,像过早发送文字也是有成本的。那就像是非常高昂的成本。
第 1363 段
是的。如果你不小心……比如说,假如你是一个光标,想象一下你的光标偶尔误点一次,那真的会超级烦人。而最糟糕的是,通常当用户试图点击时,他们也会保持不动,因为他们正停在想要点击的目标上方,并准备点击。这意味着,在我们构建的数据集中,平均而言,低速或保持静止的意图,往往与用户正试图点击相关。
第 1364 段
哇,那真的很吸引人。
第 1365 段
情况也并非如此。人们会想:“哦,点击是一个二进制信号。这肯定超级容易解码。”嗯,是的,确实如此,但要让它成为对用户有用的东西,门槛要高得多,而解决这个问题有一些办法。我的意思是,你可以采取某种复合方法,比如,好吧,就让这个,像是,让一次点击花5秒。让我们采用一个巨大的时间窗口,这样它就能对答案非常有把握。
第 1366 段
但还是那句话,世界上最好的鼠标。世界上最好的鼠标不会花1秒或500毫秒来完成点击。它只需要5毫秒或更短时间就能点击。所以,如果你的目标是达到那么高的标准,那你就真的得解决底层问题。
第 1367 段
所以,也许这里很适合问一下如何衡量性能,也就是每秒比特数这一整套东西。你能解释一下你说的是什么意思吗?也许可以先谈谈作为一款游戏的Webgrid,把它作为衡量性能的一个很好的说明。
第 1368 段
是的,也许我先退一步,从更宏观的角度解释一下我们为什么要衡量这个。所以还是那句话,我们的目标是让用户能够像我一样控制计算机,并且希望能比我控制得更好。这意味着他们能以和我一样的速度做到这一点。
第 1369 段
这意味着他们能使用我所拥有的全部相同功能,包括Command-Tab、Command-Space之类的所有细节。他们需要能够用大脑做到这些,并且达到和我用肌肉操作时相同的可靠程度。这是个很高的标准。因此,我们打算衡量并量化其中的每一个方面,以了解我们在实现这一目标的道路上进展如何。
第 1370 段
顺便说一下,衡量BPS的方法有很多。这并不是唯一的方法,但我们会向用户呈现一些创意目标,基本上,我们会计算一个分数,它取决于他们选择目标的速度和准确度,以及目标有多小。屏幕上的目标越多、目标越小,每次点击所呈现的信息就越多。
第 1371 段
所以,如果从信息论的角度来思考,你可以通过不同的信息论信道进行交流。其中一种信道可以想象成一个由网格构成的打字界面,就像屏幕上的软件键盘一样。每秒比特数是一项指标,其计算方式是取屏幕上目标数量的对数。
第 1372 段
如果你想为键盘建模,可以减去1,因为必须为键盘上的Delete键减去1;但具体来说,就是屏幕上目标数量的对数乘以正确选择次数减去错误选择次数,再除以某个时间窗口。例如60秒。这算是学术界衡量光标控制任务的标准方法。这一切都要归功于那位杰出的教授,博士。
第 1373 段
斯坦福大学的Shenoy,是他提出了那项任务。他也是激励我进入这个领域的人之一。所以,他提出了一项标准化指标,让我们如今能够以此一较高下,说Noland使用他的BCI完成这项任务时是世界上最出色的人,这一切都要归功于他。
第 1374 段
拥有可供人们在不同技术和方法之间进行比较的标准化指标,对进步非常重要。这个方法效果如何?所以,是的,要向他以及斯坦福大学的整个团队致以高度赞扬。是的,所以,对于Noland和我玩这项任务而言,这项任务还有不同的可配置模式。
第 1375 段
所以,Webgrid任务可以仅以在屏幕上进行某种左键点击的形式呈现,也可以设置只需将光标停留在上面的目标,或者设置需要左键、右键点击的目标。可以设置需要左键、右键点击、中键点击、滚动、点击和拖动的目标。在这个总体框架内,你可以做各种各样的事情。但最简单、最纯粹的形式,就是屏幕上出现蓝色目标。蓝色意味着左键点击。
第 1376 段
这是这款游戏最简单的形式。而此前学术研究中的记录,以及Neuralink内部使用某种NPS取得的记录,都已被使用Neuralink设备的Noland追平或打破。所以,在Neuralink之前,人类使用设备所创造的世界纪录大约介于4.2至4.6 BPS之间,具体取决于你读的是哪篇论文以及如何解读它。Noland目前的纪录是8.5 BPS。
第 1377 段
再说一次,Neuralink使用者的中位表现大约是10 BPS。所以可以粗略地认为,在使用光标选择屏幕上的蓝色目标时,他的控制水平相当于Neuralink使用者中位水平的85%。而且,是的,我认为,要让我们达到同样的10 BPS性能水平,前方还有一段非常有意思的旅程。
第 1378 段
让我们从4 BPS提高到6 BPS,再从6 BPS提高到8 BPS的那些技巧,并不是能让我们从8提高到10的技巧。在我看来,这里的核心挑战其实是标注问题。也就是如何以非常、非常精细的分辨率理解用户正试图做什么。而且,是的,我非常鼓励学术界的人们研究这个问题。
第 1379 段
在提高Webgrid BPS的这项追求中,Noland经历了怎样的过程?3月时,你说他在Webgrid中选择了89,285个目标。
第 1380 段
对。
第 1381 段
所以他很喜欢这款游戏。他非常认真地想提高自己在这款游戏中的表现。那么,设法弄清如何提高这种表现的过程是怎样的?其中有多少可以在解码端实现?有多少可以在校准端实现?又有多少可以在Noland这一端通过设法弄清如何更清晰地传达他的意图来实现?
第 1382 段
是的,这确实是个很好的问题。所以在我看来,Noland表现如此出色的主要原因之一就是Noland本人。Noland极其专注,而且精力非常充沛。他有时会在半夜玩Webgrid,一玩就是大约4个小时。比如从凌晨2:00到早上6:00,他会一直玩Webgrid,只因为他想把它推到自己能力所及的极限。而且这并不是我们要求他这么做的。我想说明这一点。
第 1383 段
我们并没有说:“嘿,你今晚应该玩Webgrid。”我们只是把这款游戏作为研究的一部分交给他,而他能够独立玩这款游戏,并在任何想练习的时候练习。他真的非常努力地推动它,把这项技术推向绝对极限。而且他确实把让我们成为瓶颈视为自己的工作。天哪,他把这件事做得太好了。
第 1384 段
所以,首先要承认的是,他极其渴望让这件事成功。我也有幸见过BrainGate和其他试验的临床试验参与者,他们在很大程度上抱有同样的态度,他们把尽其所能推动这项技术进步视为自己毕生的工作。而如果这意味着要从凌晨2:00到早上6:00连续4个小时选择屏幕上的目标,那就这样吧。
第 1385 段
这其中有某种极其令人钦佩、值得特别指出的东西。好,那么你如何从他起步时的状态,也就是完全无法控制光标,逐渐达到一个BPS?所以,我的意思是,他刚开始时,他那边和我们这边都需要进行大量学习,才能弄清楚对他来说最直观的控制方式是什么。
第 1386 段
而对他来说最直观的控制方式,某种程度上就是,你必须找出我们有哪些信号可供解码这一集合的交集。所以,我们并不会采集运动皮层中的每一个神经元,这意味着我们并没有身体每个部位的表征。因此,对于某些信号,我们的解码表现可能比其他信号更好。
第 1387 段
例如,在他的左手上,我们很难区分他的左手无名指和左手中指。但在他的右手上,我们能从可记录到的神经元中检测到他的小指、拇指和食指具有良好的控制和调制。因此,你可以想象,这些不同的调制活动子空间如何与对他而言最直观的方式相交。而且这一直在随时间演变。
第 1388 段
所以,一旦我们让他能够自行校准模型,他就可以去探索各种不同的方式,想象如何控制光标。例如,他可以想象通过左右摆动手腕来控制光标,或者通过移动整条手臂来控制。我想有一次,他还用了双脚。他尝试了一大堆方法,探索对他来说控制光标最自然的方式是什么。
第 1389 段
与此同时,我们也很容易解码规则。
第 1390 段
确认一下,你们是通过身体映射程序来判断他能移动哪根手指的吗?
第 1391 段
是的,对,那是一种方法。也许这里有一个细微之处:当他这样做时,他能想象的东西比我们在屏幕上的那个可视化中呈现的要多得多。所以我们以一种比较抽象的方式向他展示:“这里有一个光标。你自己找出最适合你的方式。”当然,通过身体映射程序,我们对什么方式最有效是有一些线索的,因为我们知道,这个特定动作是我们能够很好地表征的。但真正要去探索并找出最适合的方法,还是由他自己决定。
第 1392 段
但到什么时候,他不再想象自己身体的动作,而只是想象光标的移动?
第 1393 段
对。
第 1394 段
他从……要多快,他要多快才能到达那里?
第 1395 段
这发生在一个星期二。我对那一天记得非常清楚,因为那天有一阵子,看起来他的表现不是特别好。看起来模型运行得不太好,而且他似乎分心了。但他其实——事实并非如此。实际上发生的是,他当时在尝试一种新方法,直接控制光标。所以他不再想象移动自己的手了。
第 1396 段
他只是在想象,我不知道那是什么,某种移动屏幕上光标的抽象意图之类的。我没法告诉你这两件事之间有什么区别。我真的、真的说不出来。他以前试着向我解释过。我无法以第一人称描述那是一种什么样的体验。
第 1397 段
但他当时脱口而出的那些粗话,足以表明,对他来说,直接通过神经控制光标是一种在性质上截然不同的体验。
第 1398 段
我在想,是否有办法通过用户体验设计来鼓励一个人发现这一点,因为他发现了它,就像你对我说的,他是一位先驱。所以,他是在整个过程中独自发现这一点的,也就是尝试用不同类型的意图来移动光标的过程。但这显然是一种非常强大的境界,也就是不再试图控制手指和手,而是用意念控制实际的数字设备。
第 1399 段
没错,用户体验决定了它如何运作。而理想的用户体验,是用户无须思考需要做什么才能把事情完成。它就是直接完成了。
第 1400 段
这太令人着迷了。但我想知道,从生物学角度来说,大脑需要多长时间才能适应。
第 1401 段
是的。
第 1402 段
所以,这仅仅是像高层软件那样的学习,还是也存在神经可塑性的成分,也就是大脑在缓慢调整?
第 1403 段
是的,事实是,我不知道。我非常期待看到我们植入的第二位参与者会经历怎样的过程,因为我们会学到更多东西。或许我们能帮助他们更快地理解并探索那个方向。这是我之前不知道的。这并不是我提示 Noland 去尝试这么做。他只是在探索如何使用自己的设备,并自行摸索明白。
第 1404 段
但既然我们现在知道这是一种可能性,那么也许有办法,比如提示用户:“校准时不要过分用力。只要做一些感觉自然的事,或者直接控制光标。不要想象具体的动作。”然后,我们应该有希望了解,一个以前从未体验过这种感觉的人会是什么情况。也许这就是他们默认的操作模式。
第 1405 段
你不必经历明确动作这个中间阶段。
第 1406 段
或者,如果人们自然而然就会做到这一点,你可以偶尔鼓励他们允许自己移动光标。事实上,有时就像四分钟跑完一英里一样,仅仅知道这是可能的。
第 1407 段
就会推动你去做到。
第 1408 段
是的,让你能够做到,然后它就变得轻而易举了。这也会让你思考,这正是人类很酷的一点。一旦有更多人类参与者,他们就会发现各种可能做到的事情。
第 1409 段
是的,并分享他们的体验。
第 1410 段
是的,并且分享。
第 1411 段
彼此分享。
第 1412 段
而且,因为他们分享了这件事,他们就能做到。突然之间,这就对所有人解锁了,因为有时仅仅知道这件事,就是让它能够做到的东西。
第 1413 段
是的,我是说,也顺便谈一下这一点,我们大概尝试过一千种不同的方法,来处理解码的各个方面,而现在我们知道了应该继续深入探索的正确子空间。再次感谢 Noland,以及他在这件事上投入的许多个小时。
第 1414 段
所以,哪怕只是这种帮助,比如帮助约束我们可以探索的不同方法的束搜索,也确实有助于加快下一位参与者的进程,包括我们第一天可以让他们尝试的各种东西、我们希望多快让他们实现实用的控制、多快能让他们独立使用它,以及从系统中获得价值。
第 1415 段
所以,是的,向 Noland 以及在他之前为让这项技术成为现实而参与其中的所有参与者致以崇高敬意。
第 1416 段
那么,解码器多久更新一次?因为 Noland 提到过,比如,好吧,我们正在开发一个新更新;而且他在直播中说,他会玩贪吃蛇游戏,因为那个游戏非常难。这是他测试更新效果有多好的一种好办法。他还说,有时更新反而是一种倒退。这是一个持续迭代的过程。更新具体包括什么?主要是在解码器这一侧吗?
第 1417 段
是的,有几点要说。首先,可能值得区分两种情况:一种是我们积极尝试不同方法、以了解最佳方案是什么的研究会话;另一种是独立使用,也就是我们希望能够直接使用这台设备,就像任何人想使用自己的 MacBook 那样。
第 1418 段
所以,我认为他所指的通常是在研究会话的语境下,我们会尝试非常非常多不同的方法,甚至包括我们之前谈过的无监督方法,试图找出更好的方式来估计他的真实意图,并更准确地解码。在这些情况下,我是说,我们会尝试……在任何一次会话中,他有时一天会工作大约8个小时。
第 1419 段
所以我们在那一天可能会尝试数百个不同的模型。就是很多不同的东西。另外,值得注意的是,我们相当频繁地更新他使用的应用程序。我想有时一天最多会更新大约4到5次。我们会用不同的功能、错误修复,或者根据他给我们的反馈来更新他的应用程序。他非常善于表达,也是解决方案的一部分。他不是一个只会抱怨的人。
第 1420 段
他会说:“嘿,我发现这个东西在我的使用流程中并不是最优的。我有一些关于如何修复它的想法。告诉我你们怎么看。我们一起来想办法解决它。”而且经常是,他向我们提供反馈后的几个小时内,这些问题就得到了解决。这就是我们的迭代周期。
第 1421 段
所以有时,在会话开始时,他会给我们反馈;到会话结束时,他已经在对那个流程或那套设置的下一次迭代提供反馈了。
第 1422 段
这太令人着迷了,因为你提到的一件事是,从 BCI 会话中记下了 271 页笔记,而这还只是在三月份。所以,人类能够提供的一项惊人之处,尤其是那些聪明、兴奋,而且像 Nolan 一样充满积极正向氛围的人,就是他们能够提供反馈,持续不断的反馈。
第 1423 段
是的,而且,稍微夸一下这个团队,我和很多非常出色的人一起工作,而这要求团队绝对全神贯注于用户,以及什么对他们来说是最好的。这还需要这样一种投入程度:“好,这是用户的反馈。我有这么多会议。我们今天先不开那些会了,来做这件事。”我想说,这种专注和投入的程度在这个世界上没有得到足够重视。而且,显然你还必须具备相应的才能,才能有效地把这些事情执行好。是的,我们有很多这样的人才。
第 1424 段
是的,而且这是一个非常有意思的用户体验设计领域,因为这里有太多未知因素了。我能看出用户体验设计很难,因为太多人做得很差。这确实不是一件简单的事。
第 1425 段
是的,而且,你知道,用户体验并不是总能通过不断迭代各种不同的东西来解决的。有时候,你真的需要退一步,从全局思考:我所处的甚至是不是那种值得继续追寻解决方案的正确极小值?有很多问题,快速的迭代周期可以预测你最终会有多成功。
第 1426 段
举个很好的例子,比如在强化学习模拟中,你获得奖励的频率越高,进展就能越快。你获得反馈越频繁,它就越是一种更容易的学习提示。但用户体验不是这样。我的意思是,用户其实经常会判断错什么才是正确的解决方案,而要找到正确答案,需要深刻理解技术系统及其可能性,同时还要理解你试图解决的问题是什么。
第 1427 段
不只是用户如何表达它,还要理解真正的底层问题是什么,才能真正到达正确的位置。
第 1428 段
是的,这就像那些关于史蒂夫·乔布斯的老故事,他就那么闯进来,然后说,是啊,用户是一种有用的信号,但不是完美的信号。有时候,你必须移除软盘驱动器之类的东西。我已经记不清所有那些史蒂夫·乔布斯做出大胆设计决策的疯狂故事了。但其中有些关乎美感,有些关乎你在设计中投入的热爱,这非常像史蒂夫·乔布斯与乔尼·艾维的那种风格。
第 1429 段
但当一个人用大脑与它交互时,它也与功能有着深刻的关系。它不只是美感。而且你必须对眼前的人产生共情,同时又不能总是直接听从他们的话。你必须深度共情。这太迷人了。真的、真的太迷人了。与此同时,还要进行迭代。但不是小幅度地迭代。有时候,是彻底重建设计。
第 1430 段
他说过,诺兰说在早期,用户体验糟透了。但你们改进得很快。那个过程是怎样的?
第 1431 段
是的,我来举一个具体的例子。他非常想能够阅读漫画。这是他很想做的一件事,我是说,这听起来像是一件简单的事,但对他来说其实非常重要。而他没法用嘴控棒做到这一点。它无法无障碍使用。在他想用来阅读最新漫画的那个网站上,他没法用鼠标棒在 iPad 上滚动页面。
第 1432 段
这里或许适合快速暂停一下,说明所谓嘴控棒,就是他嘴里含着一根棒,用它在平板电脑上滚动页面的工具。
第 1433 段
对,是的,基本上,你可以把它想象成一支夹在牙齿之间的触控笔。是的,它基本上是一支很长的触控笔。
第 1434 段
而且这很累人,会疼,也很低效。
第 1435 段
是的,也许还有一点值得特别指出,还有其他替代性的辅助技术,但就 Noland 的具体情况而言,而且这并不少见,我认为人们对此也不太了解,他的痉挛相对严重,所以他会时不时出现肌肉痉挛。
第 1436 段
所以,任何要求他直接位于摄像头前方的辅助技术,比如眼动追踪器,或任何要求他把某样东西放进嘴里的技术,都完全行不通,因为他痉挛时要么会移出画面,要么如果嘴里含着东西,痉挛得太厉害时,那东西就会刺到他的脸。
第 1437 段
所以,在思考 BCI 能为某人的生活带来什么优势时,这类考量非常重要。如果它能以符合人体工学的方式融入你的生活,让你在照护者不在场时也能独立使用,想在哪里用就在哪里用,无论是在床上还是椅子上,取决于你的舒适程度以及你是否想长压疮,那么所有这些因素都会极大影响这个解决方案在那位用户生活中的好坏。
第 1438 段
所以,其中一个非常有趣的例子就是滚动。再说一次,漫画是他想要能够阅读的东西,而用 BCI 实现滚动有很多种方式。比如,你可以设想不同的手势。用户可以做出那种手势,页面就会移动。但滚动是一种非常迷人的控制界面,因为它是你面前屏幕上的一个巨大对象。
第 1439 段
所以,模型输出中的任何抖动、模型输出中的任何错误,都会在屏幕上引发类似地震的效果。比如,你真的不会希望自己正试图阅读的漫画页面仅仅因为滚动解码器不够准确,就上下偏移几个像素。
第 1440 段
因此,在这个例子中,我们必须弄清楚如何构建问题,使系统的错误无论何时出现——而且我们会尽最大努力将其降到最低——无论这些错误何时出现,都不会再次打断用户正在经历的体验之感质。不会打断他们阅读书籍时的流畅状态。所以,我们最终打造出了这项非常出色的功能。
第 1441 段
这是一个名叫 Ruse 的团队成员,他完成了这项非常出色的工作,叫作“快速滚动”。快速滚动基本上会查看屏幕,并识别屏幕上哪些位置有滚动条。它通过与 MacOs 深度集成,利用 MacOs 应用可用的那种辅助功能工具,了解屏幕上当前哪些位置存在滚动条。
第 1442 段
我们识别出那些滚动条的位置,并提供了一个 BCI 滚动条。BCI 滚动条看起来与普通滚动条相似,但其行为非常不同:一旦你移动到它上面,光标就像是变形并贴到上面。它仿佛附着或锁定在上面。然后,只要你向上或向下推动,就像用推动来控制普通光标一样,它实际上就会替你移动屏幕。
第 1443 段
所以,它基本上就像是把速度重新映射为滚动操作。它之所以感觉如此自然、直观,是因为当你移过去并与它连接时,感觉就像有磁性一样。于是你就像是粘在了上面。然后,这就是一个连续的动作。你不必切换自己想象中的动作。你仿佛一下吸附到它上面,然后就可以继续了。你可以立刻开始把页面往下拉或往上推。
第 1444 段
而且,即使把这一点做好了,滚动行为要如何运作才能显得自然、直观,仍然有非常多细微之处。一个例子就是惯性。比如,当你用手指在屏幕上滚动页面时,实际上会有某种流动感。它不会在你抬起手指时立刻停止。BCI 滚动也是如此。所以我们必须花一些时间弄清楚恰当的细微处理是什么。
第 1445 段
当你的指尖不再能感受到屏幕时,正确的动态效果是什么,或者可以说,当你推动页面时,页面应该有多大程度的延续移动,才能让它以恰当的幅度滑动,使用户在阅读书籍时获得自然的体验。
第 1446 段
而且有无数个,我是说,我可以告诉你,滚动究竟如何运作,其中有太多微小细节,我们大概花了1个月才把它们调对,让用户在操作时感觉极其自然、轻松。
第 1447 段
我的意思是,就连用手指在智能手机上滚动,感觉也极其自然、愉悦。而要把它做到位,可能需要极其漫长的时间。实际上,这也需要我们正在谈论的那种富有远见的用户体验设计:不要总是听用户的,但也要听他们的,而且还要有那种富有远见的宏大思维,比如把一切都扔掉,从第一性原理出发思考,但也不要。
第 1448 段
是的,是的,顺便说一句,这让我想到,桌面端的滚动条可能已经停滞不前,从未采用过那种,比如吸附,和你所说的吸附网格、吸附滚动条操作一样,这种功能甚至对普通用户而言,在桌面环境中也可能极其有用,能够改善体验,因为目前桌面端的滚动条体验糟透了。
第 1449 段
是的,同意。
第 1450 段
它很难找到,也很难控制。没有惯性。而且意图应该是明确的。当我开始朝滚动条移动时,应该触发吸附到滚动条的操作。当然,也许我愿意付出这个代价,但有数亿人一直在付出这个代价。不过不管怎样,在这个例子中,这是必要的,因为 Noland 还要为抖动付出额外代价。所以你必须在滚动和阅读之间切换。两者之间必须有一个阶段转换。比如,当你在滚动时,你就是在滚动。
第 1451 段
对,对,所以这是当前方法的一个缺点。这里也许还有另一个案例研究,所以再说一次,用户体验就是它如何运作,而我们会从整体角度考虑这一点,从我们在大脑中检测什么的特征检测层面,到我们如何设计解码器、选择解码什么,再到用户实际使用它时,它如何运作。
第 1452 段
所以,另一个很好的例子涉及他们真正开始使用解码器后它如何运作:屏幕上显示的输出并不只是解码器说了什么。它也取决于屏幕上正在发生什么。
第 1453 段
所以,比如说,我们可以理解,当你试图关闭一个标签页时,那个非常小、愚蠢的小 X 极其微小;如果你面对的是解码器某种带噪声的输出,就很难精准点中它。我们可以理解,那是一个你可能想要点中的小 X,并实际为你把它变成一个更大的目标。
第 1454 段
这类似于你在手机上打字时的情况,比如,如果你习惯使用 iOS 键盘,它实际上会根据底层语言模型调整各个按键的目标区域大小。所以,如果我正在输入“嘿,我要去见 L”,它实际上会理解。它会把 E 键变大,因为在那些情况下,Lex 是我要去见的人。
第 1455 段
所以,即使不改进底层解码器或技术栈中的特征检测部分,这种预测能力也能让体验顺畅得多。我们通过一项叫作“磁性目标”的功能来做到这一点。我们实际上会为屏幕建立索引,并理解:好,这些地方是非常小的目标,可能很难点中。
第 1456 段
这是该位置周围可能表明用户正试图选择它的那类光标动态。我们让它更容易些。我们把它的尺寸放大,让用户能更容易地大致吸附到那个目标上。所以所有这些小细节都非常重要,有助于用户在日常生活中保持独立。
第 1457 段
那么,解码器方面的工作有多少能泛化到 P2、P3、P4、P5、PM?你如何以一种具有泛化性的方式改进解码器?
第 1458 段
是的,问得好。所以,我们试图解码的底层信号在 P2 中看起来会与在 P1 中非常不同。例如,345 号通道在用户一身上所代表的含义会不同于在用户二身上所代表的含义,仅仅是因为对应于 345 号通道的那个电极,在用户一和用户二身上会分别靠近不同的神经元。
第 1459 段
但这种方法,也就是这些方式,以及如何让用户产生恰当类型的行为模式并将其与该神经信号关联起来的用户体验,我们希望这些能沿用到多代用户身上。除此之外,非常非常有可能,事实上很可能,我们已经对诺兰的用户体验诉求和偏好出现了某种过拟合。
第 1460 段
所以,我希望看到的是,当我们有了第 2、第 3、第 4 位参与者时,我们能找出覆盖所有情况的、某种恰当的宽阔极小值,让它对每个人来说都更加直观。而且希望不同做法之间会出现交叉借鉴,比如:“哦,我们面对这位用户时没有想到这一点,因为他们能说话。但对于这位从根本上完全无法说话的用户,这种用户体验并不是最优的。”
第 1461 段
而这实际上会,那些我们在那里做出的改进,应该有望继而也适用于那些无法说话、但因为身处公共场合而觉得那样做不自在的人,比如在他们的医生诊所里。
第 1462 段
所以,开环标注然后闭环标注的实际机制会是相同的,而且希望能在不同用户进行校准步骤时实现泛化。校准步骤也相当酷。我的意思是,Webgrid 本身有趣的一点在于,它有点像闭环,它很好玩。我很喜欢这种情况,比如,以前有一种所谓的人类计算理念,就是利用人类无论如何都想做的动作,从中获得大量信号。
第 1463 段
对。
第 1464 段
而 Webgrid 就是那种,像是一款不错的电子游戏,同时也能很好地用于校准。
第 1465 段
太有意思了。这种反应我听过太多次了。在第一位用户植入之前,我们内部原本觉得第一位用户不会觉得这有趣。所以我们其实认真想了很多,比如,我们是不是应该制作一些用户觉得更有意思的其他游戏,这样我们就能获得这类数据,并帮助推动长时间研究之类的工作?结果发现,人们很喜欢这个游戏。
第 1466 段
对。
第 1467 段
我一直都很喜欢它,但我不知道大家也有这种看法。
第 1468 段
对,以防大家还不清楚,Webgrid 是这样的:有一个网格,比如说是 35×35 个单元格,其中一个会亮成蓝色,你必须把鼠标移到上面并点击它。如果你没点中,它就是红色的——[布利斯] 我玩这个游戏玩了好多好多小时。好多好多小时。
第 1469 段
你刚才说你的纪录是多少?
第 1470 段
我想我的成绩在 Neuralink 是最高的。目前我的纪录是 17 BPS。
第 1471 段
17 BPS。
第 1472 段
这大概意味着,如果你想象一下那个 35×35 的网格,你每分钟大约能完成 100 次试验。也就是在那 1 分钟的时间窗口内正确选择 100 次。所以你每次选择平均大约需要 500 到 600 毫秒。
第 1473 段
我觉得自己玩那个游戏很吃力,其中一个原因是我太习惯用键盘了,所以一切都通过键盘完成。只要能避免碰鼠标,那就很好。那么你如何解释自己的出色表现?
第 1474 段
我玩 Webgrid 时有一整套仪式。其实还配有一套饮食计划。就像是完整的一整套东西,所以很棒。
第 1475 段
第一件事——你得禁食 5 天,还得上山。
第 1476 段
其实,这有点,我是说,禁食这件事很重要。所以这就像,你知道——能让精神集中,对吧?
第 1477 段
对,是真的。所以我实际的做法是,事先有一小段时间不吃东西。然后就在开始前吃掉大量花生酱,我就会——这是真的?
第 1478 段
这是真的,对。然后还必须是深夜。我觉得这也是我们共有的夜猫子习惯,但必须是在午夜到凌晨 2:00 左右的时间窗口。而且我会用一种非常特定的身体姿势坐着,这是因为,我以前,我从小在家上学,所以大部分功课都是在地板上做的,就在卧室里之类的地方。因此我在地板上有一个非常特定的姿势。
第 1479 段
我坐在地板上玩,然后你必须确保玩的时候肘部没有承受太多重量,这样才能快速移动。接着我会把光标的增益,也就是光标速度调得非常非常高。这样只要做很小的动作,实际上就能移动光标。
第 1480 段
你是在用手腕移动,还是从来不用——我用手指移动。所以我的手腕几乎完全不动。我只是在动手指。
第 1481 段
对。稍微岔开一下话题,我一直想深入研究一下那些创下《俄罗斯方块》世界纪录的人。那些人,他们玩的时候,有一种方法可以……你看过这个吗?
第 1482 段
我看到像是 3 根,像是所有手指都在动。
第 1483 段
对,你可以找到一种方法,像是利用一个漏洞,像是一个程序错误,让你能以难以置信的速度完成一些操作。所以这与那个有些相似,但又不完全一样。但你确实意识到,现在会有几位程序员正在听这些,快速地冷却并吃花生酱,然后说——对,请,请交易我的纪录。我是说,我这么做确实只是因为我希望标准定得高。团队,就像是我希望我们瞄准的数字不应该只是中位数表现。它至少应该能胜过我们所有人。那应该是最低标准。
第 1484 段
你觉得有可能达到多少,比如 20 次刮擦?
第 1485 段
对,我不知道极限是多少。我的意思是,你可以单纯根据屏幕刷新率以及光标立即跳到下一个目标之类的因素来计算极限。但我是说,我确信在达到那个极限之前,还会受到反应时间、视觉感知之类因素的限制。我猜会低于 40 但高于 20,大概就在这个范围内,这可能是我永远不会去思考的正确答案。
第 1486 段
任务的难度也很重要。你可以想象,有些人也许能在屏幕上处理大约 10,000 个目标,或许那样他们的表现会更好。你也可以进行一些任务优化,尝试提升自己的表现。
第 1487 段
你觉得要怎样才能让 Noland 超过 8.5?怎样才能让那个数字继续提高?你说过,数字每提高一次,可能都需要不同的——[Bliss] 对。
第 1488 段
对系统进行不同的改进。
第 1489 段
对,我认为这项工作的性质是,我是说,首先必须说明的重要答案是,我不知道。这是研究的前沿领域。还是那句话,此前没有人达到过那个数字。所以接下来只能由我做出启发式猜测。我们过去看到的情况是,技术栈的不同部分会紧挨着不同的时间点。
第 1490 段
所以大约 3 年前我刚加入 Neuralink 时,其中一个主要问题就是蓝牙连接的延迟。就像是那个径向设备不是特别好。那是植入物的早期版本,而且就像是,无论你的解码器有多好,如果你的设备每隔 30 毫秒或 50 毫秒才更新一次,运行起来就会卡顿。
第 1491 段
而且不管你有多优秀,那都会令人沮丧并带来挑战。所以在那个时候,非常明显,主要挑战就是以非常可靠的方式把数据从设备中传出来,从而让你能够着手解决下一个挑战。
第 1492 段
后来在某个时候,问题实际上变成了建模挑战,也就是如何构建良好的映射,比如这个监督学习问题:你有一大堆数据,还有一个试图预测的标签,那么究竟什么样的神经解码器架构和超参数才适合优化这个问题?这个问题持续了一阵子。一旦解决了它,瓶颈又变成了另一个。
第 1493 段
我认为,在那之后的下一个瓶颈其实就是某种软件稳定性和可靠性。如果你的系统中推理延迟变化很大,或者你的应用时不时就会卡住,就会降低你保持并进入心流状态的能力,而且基本上会扰乱你的操控体验。
第 1494 段
所以,我们修复了各种不同的软件错误并做出了改进,这些工作基本上提升了系统的性能,使其可靠得多、稳定得多,并让我们达到了能够可靠地收集数据、用来构建更好模型的状态。所以这有一段时间是个瓶颈。就是类似软件栈本身的问题。
第 1495 段
如果让我现在猜的话,为了进一步提升 BPS,大致有两个主要方向可以考虑。第一个主要方向是标注。所以,标注仍然是这样一个根本性挑战:给定一段用户正在表达某种行为意图的时间窗口,他们在每一毫秒的粒度上究竟想做什么?而这同样是一个任务设计问题。这是一个用户体验问题。
第 1496 段
这是一个机器学习问题。这是一个软件问题。某种程度上涉及所有这些不同的领域。为了进一步提升 BPS,你可以考虑的第二件事,是要么彻底改变你正在解码的东西,要么只是扩展你正在解码的东西的数量。所以,这是在服务于功能性这个方向,明白吗?因此,你可以设想提供更多种点击。例如,左键点击、右键点击、中键点击。
第 1497 段
例如点击并拖动等不同操作。而这可以提高你的通信辅助装置的有效比特率。如果你试图让用户通过任何给定的通信渠道表达自己,就可以用每秒比特数来衡量。但归根结底,操作所衡量的是他们在操控计算机时有多高效。
第 1498 段
所以,从你所关心的下游任务的角度来看,我们对功能性以及扩展功能性非常感兴趣,因为这不仅能提高某种意义上的 BPS 数值,还能提升用户在下游层面的某种独立性,以及他们操作计算机时的熟练程度和效率。
第 1499 段
增加线程数量是否也可能有所帮助?
第 1500 段
是的,简短的回答是肯定的。这条曲线或者说它如何体现在数字上,有一些细微之处。所以你会看到,如果你大致绘制一条曲线,将用于解码的通道数量与以下任一指标相比较:衡量解码效果有多好的离线指标,或者衡量实际中用户使用该设备效果有多好的在线指标,你会看到大致呈一条对数曲线。
第 1501 段
所以,随着通道数量进一步增加,你会在操控质量和离线验证指标上获得相应的某种对数式提升。这里重要的细微之处在于,每个通道都对应大脑中一个特定的、被表征的意图。例如,如果你有一个 254 号通道,它可能对应向右移动。256 号通道可能表示向左移动。
第 1502 段
如果你想扩大想要控制的功能数量,你确实会希望拥有覆盖更广泛想象动作的一组更广泛的通道。实际上,你可以把它想成有点像土豆先生。比如,如果你可以做出一大堆不同的想象动作,你会如何把这些想象动作映射为计算机的输入?你可以想象用手写来在屏幕上输出字符。
第 1503 段
你可以想象只是用手指打字,并让它在屏幕上输出文本。你可以想象用不同的手指调制来实现不同的点击。你可以想象扭动你的大鼻子或打开某个菜单,或者扭动你的大脚趾,让类似 Command-Tab 的操作发生,诸如此类。所以实际上,你能在世界中采取多少种不同的行动,取决于你拥有多少通道,以及它们所承载的信息内容。
第 1504 段
对,所以那更多关乎动作的数量。所以实际上,随着线程数量增加,那更多是在增加你能够执行的动作数量。
第 1505 段
这里还有一个值得一提的细微之处。所以再说一次,我们的目标确实是让一个具有过程的用户能够像我一样快地控制他们的计算机。所以那就是 BPS。拥有我具备的全部相同功能,也就是我们刚才讨论的内容,但还要像我一样可靠。而最后这一点与渠道账户的讨论密切相关。
第 1506 段
所以,随着通道数量的扩展,你的模型输入中任何特定特征对于用户输出控制的相对重要性都会降低。这意味着,如果这种神经非平稳效应是逐通道发生的,或者噪声是相互独立的,以至于通道越多,平均而言对输出的影响就越小,那么系统的可靠性就会提高。
第 1507 段
所以,至少我持有的一个核心论点是,扩展渠道账户应该能在不对解码器本身做任何工作的情况下,提高系统的可靠性。
第 1508 段
你能在这里多谈谈可靠性吗?首先,当你看到信号的非平稳性时,你指的是哪个方面?
第 1509 段
好,也许我们先简要谈谈实际的底层信号是什么样的。再说一次,我一开始非常简短地提到过,当你想象向右移动或想象向左移动时,神经元的放电可能会更多或更少。而那个信号的频率内容,至少在运动皮层中,与输出意图,也就是用户正在进行的行为任务,高度相关。
第 1510 段
实际上你可以设想,这一点并不显而易见,频率编码,也就是这种现象的名称,仿佛是大脑能够表征信息的唯一方式。你可以设想大脑编码意图的许多不同方式。而且实际上有证据表明,例如在蝙蝠中,存在时间编码。所以,像特定神经元究竟何时放电这样的时序编码,就是信息表征的机制。
第 1511 段
但至少在运动皮层中,有大量证据表明它是频率编码,或者至少,一个类似于一阶的事实是,它是频率编码。那么,如果大脑通过改变某个神经元放电的某种频率来表征信息,真正重要的就是神经元的某种基线状态与它受到调制时呈现的状态之间的某种差值。
第 1512 段
而我们观察到的,以及学术研究中也观察到的,是那个基线频率,也就是某种……如果你要以秤为目标,如果你想象一下烘焙时测量面粉之类东西的类比,那么容器重量的那个基线状态实际上每天都不一样。
第 1513 段
所以,如果你试图测量的是容器里有多少米,你在不同的日子会得到不同的测量结果,因为你使用的是不同的容器进行测量。所以,基线频率的漂移实际上就是造成这种下游偏差的因素,至少从对问题的一阶描述来看是这样。
第 1514 段
除此之外还可能有其他效应、非线性效应,但至少,作为对这个问题非常初步的一阶描述,我们日复一日观察到的情况是,在某个特定通道上观察到的任意特定神经元的基线放电率都在变化。
第 1515 段
所以你能不能直接根据基线不断进行调整,让它始终相对于基线?
第 1516 段
是的,这是个很好的问题。对于猴子,我们已经找到了各种办法来做到这一点。举一种方法,你让它们执行某项行为任务,比如用操纵杆玩游戏;你测量大脑中正在发生的活动,计算所有输入特征中正在发生的活动的某种均值,然后在进行脑机接口会话时从输入中减去这个均值。效果非常好。
第 1517 段
但不知为何,这对 Noland 的效果并不是特别好。实际上,我并不完全清楚其中的原因,但我可以想到几种解释。其中一种解释可能是,对 Noland 而言,某个开环任务和某个闭环任务之间的情境效应差异,要比对猴子而言显著得多。
第 1518 段
也许在这个开环任务中,他一边执行任务,一边观看 Lex Fridman 播客;或者一边吹口哨、听音乐、和朋友交谈,一边问他妈妈晚饭吃什么。
第 1519 段
因此,这两种状态之间的具体情境差异可能要大得多,从而导致你在开环阶段进行归一化的特征,与试图在闭环阶段使用的特征之间出现更大的泛化差距。
第 1520 段
这很有意思。就这一点而言,看着 Noland 能够做到、能够多任务处理、同时执行多项任务,能够在说话且感到紧张时仍有效地移动鼠标光标,确实有点不可思议,因为他正在……面前说话——还踢我的屁股和胸口,是的。
第 1521 段
把你打得屁滚尿流。而且做这些时还说垃圾话。所以这一切都是同时发生的。没错,如果你试图根据基线进行归一化,那可能会让一切都偏离。天哪,这太有意思了。
第 1522 段
也许还可以就这一点说一句。对于不熟悉辅助技术的人,我认为有一种普遍看法是,嗯,为什么不能直接使用眼动追踪器之类的东西,来帮助某人在屏幕上移动鼠标?这确实是个合理的问题,而在 Noland 之前,我实际上并没有信心认为,这会成为一项能给他这样的人带来深刻变革的技术。
第 1523 段
而现在我非常确信它会做到,但其中的原因很微妙。这其实与它如何从人体工学角度融入他们的生活有关。即使你只能提供与眼动追踪器或口控鼠标相同程度的控制,但你不需要让那件东西一直挡在脸前。你不需要保持某种特定姿势。你不需要照护者在身边替你设置它。
第 1524 段
你可以在自己想用的时候、以自己想要的方式、在任何想用的地方启用它。这种程度的独立性对人们而言具有颠覆性的意义。这意味着他们可以在晚上私下给朋友发短信,不需要让妈妈知情。这意味着他们可以在凌晨 2:00 打开设备浏览互联网,即使周围没有人替他们设置好 iPad。对于身处那种境况的人来说,这确实是一件具有深刻变革意义的事。
第 1525 段
而这甚至还没开始谈那些可能完全无法交流,或者想寻求帮助时也无法求助的人。这可能是他们与外界唯一的联系。而且,是的,我认为这一点不需要解释为什么它会产生如此重大的影响。
第 1526 段
你提到了神经解码器。这个解码器中有多少机器学习成分?有多少魔法、多少科学、多少艺术?要设计出一个能弄清这些脉冲序列意味着什么的解码器有多难?
第 1527 段
是的,好问题。这个问题有几种不同的回答方式。所以也许我先简要地拉远视角,然后再深入其中一个兔子洞。从宏观视角来看,构建解码器实际上就是构建数据集,再将其编译成权重的过程。而其中每一步都很重要。
第 1528 段
我认为,进一步改进的方向将主要在数据集这一侧,也就是如何为模型构建最优标签。但还有一个完全独立的挑战,那就是如何编译出最佳模型。所以我会简要深入第2个,也就是第2个兔子洞。为 BCI 设计最优模型时,主要挑战之一是离线指标未必与在线指标相对应。
第 1529 段
从根本上说,这是一个控制问题。用户正试图控制屏幕上的某个东西,而你如何输出意图这一具体的用户体验,会影响他们的控制能力。比如,如果你只看模型预测的验证损失,那么可能有多种方式可以实现相同的验证损失。并非所有方式都同样便于最终用户控制。
第 1530 段
而这可能简单到只是说:“哦,你可以加一些辅助损失项,来帮助你捕捉真正重要的东西。”但这是一个非常复杂、微妙的过程。所以,如何把标签转化为模型,比一个标准的监督学习问题更为微妙。
第 1531 段
这里有一件非常有意思的轶事。比如,我们尝试过许多不同类型的神经网络架构,将大脑数据转换为速度输出。而有一个几年前的例子一直留在我的脑海里:有一次,我们只使用全连接网络来解码大脑活动。
第 1532 段
我们尝试了 A/B 测试,在在线控制会话中测量一种对输入信号进行 1D 卷积的相对性能。所以,如果你设想一下,对每个通道而言,都有一个滑动窗口,同时为每个通道的每一个输入序列生成某种 Commvault 特征。
第 1533 段
实际上,使用这种卷积架构,你可以获得更好的验证指标,也就是说,你能更好地拟合数据,而且它在离线数据上的泛化也更好。你减少了参数。在处理时间序列数据时,这算是一种标准流程。现在结果表明,当在线使用该模型时,可控性反而更差,差得多,尽管离线指标更好。
第 1534 段
对此可以有很多种解读,但至少它让我明白了,嘿,至少目前的情况是,如果你只是向这个问题投入一大堆计算机,试图进行某种超参数优化,或者让某个 GPT 模型硬编码、提出或发明许多不同的解决方案,如果你只优化损失,那是不够的,这意味着这里仍然存在某种内在的建模缺口。
第 1535 段
这里仍有一些尚待发掘的技艺,涉及如何让你的模型随着更多算力而扩展。那可能从根本上说是一个标注问题,但其中也可能还有其他组成部分。
第 1536 段
目前是受数据限制吗?听起来是这样。你要怎样获得大量高质量标签?
第 1537 段
是的,我认为限制因素是数据质量,不一定是数据数量。
第 1538 段
但甚至,比如说,就连数量也是。我是说,因为它必须在这些交互上训练。我猜交互并没有那么多。
第 1539 段
是的,所以这取决于你说的是哪个版本。比如,如果你说的只是最简单的 2D 速度示例,那么我认为,是的,数据质量是主要问题。
第 1540 段
如果你说的是如何构建一种多功能输出,让你能够完成你我在计算机上可以进行的所有输入,那实际上是一个复杂得多、微妙得多的建模挑战,因为现在你不仅需要考虑用户什么时候在左键单击,而且在构建左键单击模型时,你还需要考虑如何确保当他们试图右键单击或试图移动鼠标时,它不会被触发。
第 1541 段
所以,与 Nolan 使用 BCI 的大约第 1 周有关的一个有趣错误示例是,当他移动鼠标时,点击信号会像坠崖一样骤降,而当他停下来时,点击信号又会上升。所以,这两个输入之间同样存在污染。另一个很好的例子是,有一次,他试图进行某种按住左键拖动。而他一开始移动,左键单击信号就像坠崖一样骤降。
第 1542 段
所以还是一样,因为两个信号之间存在某种污染,你需要想出某种办法,在数据集或模型中构建对这种情况的稳健性。你可以把它理解成过拟合,但实际上,只是模型以前没有见过这种可变性。所以你需要找到某种方法,在这方面帮助模型。
第 1543 段
这太酷了,因为感觉所有这些问题都非常有可能解决,但很难。
第 1544 段
是的,这从根本上说是一个工程挑战。这一点需要强调,同时也需要强调,它可能并不需要根本性的新技术,这意味着,比如那些在 Siri 内部使用 CTC 损失研究无监督语音分类的人,他们的技能可能非常适用于此。
第 1545 段
那么,对于 Neuralink 软件栈未来的发展,你对哪些事情感到兴奋?就是我们一直谈论的所有内容,解码、用户体验——我想,有些是我感到兴奋的,比如从技术层面让我兴奋的事情,还有些则是让我兴奋于了解如何让这项技术以最佳方式进入世界的事情。所以我倒着来说。
第 1546 段
在技术进入世界这一方面,我非常期待了解这款设备对那些完全无法说话的人效果如何。他们根本无法通过语音命令等方式,让自己起步并进入有用的控制状态,而且他们目前的能力受到极大限制。
第 1547 段
我认为,这将为我们理解一个问题提供极其有用的信号,我是说,这确实是所有初创企业面临的生存威胁,也就是产品与市场的契合度。这款设备在当前状态下是否有能力和潜力改变人们的生活?如果没有,缺口在哪里?如果存在缺口,我们怎样以最高效率解决它们?
第 1548 段
所以,这就是我对接下来大约 1 年的临床试验运营非常期待的地方。技术方面,我们正在做的基本上一切都让我相当兴奋。我认为这会非常棒。要说最突出的一个,我会说是扩大通道数量。所以目前,我们有一款 1,000 通道的设备。下一个版本将有 3,000 到 6,000 个通道。我预计这条曲线今后会继续延伸。
第 1549 段
目前还不清楚,在那种规模下,哪些问题会彻底消失,哪些问题会依然存在并需要重点关注。所以,我很期待它为我们提供的清晰梯度,让我们知道该把时间和资源集中在哪些用户体验上。而且就,嗯,甚至像非平稳性这么简单的事情而言也是如此。
第 1550 段
比如,在那种规模下,这个问题会不会彻底消失,还是即便到了那个时候,我们仍然需要提出新的、有创意的用户体验?
第 1551 段
而且,当我们到达那个时间点,开始大幅扩展一个大脑能够输出的功能集合时,要如何处理所有这些微妙之处:既包括用户无法感受到指尖下不同按键的用户体验,又仍然需要能够同步调节所有按键,以实现自己想做的事情。
第 1552 段
而且同样,你没有那套合适的反馈回路,所以,如何能让用户在没有实际触感的情况下,直观地控制一个高维控制界面?我认为这将是一个超级有趣的问题。我也相当期待弄清楚,这些缩放定律是否会继续成立?比如,当你扩大通道数量时,在真正达到那个饱和点之前,你还能走多远?
第 1553 段
而今天这一点并不明显。我认为我们只知道某种插值空间内的情况。我们只知道 0 到 1,024 之间的情况。我们不知道超出这个范围后会怎样。然后还有一整套有趣的神经科学和大脑问题,也就是当你把更多东西放进大脑、放到更多位置时,你就能快得多地了解那些大脑区域代表什么。
第 1554 段
所以,我对这种基础神经科学学习感到兴奋,这对于弄清楚未来如何以及如何最高效地植入电极也很重要。所以,是的,在所有这些维度上,我都真的、真的非常兴奋。而这甚至还远没有触及我们每天都在研究的软件栈,以及我们目前正在研究的东西。
第 1555 段
是的,对我来说,1,000 个电极就是饱和点几乎是不可能的。感觉这会成为未来那种显得很荒谬的想法之一,因为显然,你应该拥有数百万个电极,而这才是真正突破发生的地方。
第 1556 段
是的。
第 1557 段
你发推文说——哦。
第 1558 段
“有些思想用诗歌来描述最为精准。”你认为这是为什么?
第 1559 段
我认为这是因为语言的信息瓶颈相当严重。然而,即使不直白表达,你也能更有效地在另一个人的大脑中重构出信息。如果你能表达一种感受,使他们能够在自己的大脑中重构出你试图传达的事物真正的深层含义和美感。他们大脑中的那种生成函数,比语言所能表达的更强大。所以诗歌的机制其实只是为那个生成函数提供输入或播下种子。
第 1560 段
所以,直白表达有时对于你试图传达的东西来说,是一种次优的压缩方式。
第 1561 段
而实际上,正是在用户经历那一生成过程时,他们才理解你的意思。这就是美妙之处。这也像是,当你看一幅美丽的画时,美的并不是画中的像素,而是你看到它时发生的思维过程,是对它的体验。那才是真正重要的东西。
第 1562 段
是的,它正与某种深层的——[Bliss] 是的。
第 1563 段
你内在的某种东西,艺术家也体验过,并且能够通过像素将其传达出来。而这实际上会与完全的心灵感应有关。就像如果你只按字面阅读诗歌,它并没有说出多少有趣的东西。它需要人类来解读。
第 1564 段
所以,正是人类心智,以及那个人在全人类集体智慧的背景下所拥有的全部经历,共同让那首诗产生意义。而他们把这些载入其中。因此,以同样的方式,从一个人传递到另一个人、承载意义的信号或许看似微不足道,但由于接收端人类心智的复杂性,它实际上可能承载着巨大的力量。是的,这很有意思。
第 1565 段
诗歌仍然没有,是谁来着?我想 Yoshi Bako 最先说过一些关于所有那些认为我们已经实现 AGI 的人的话:解释一下人类为什么喜欢音乐。
第 1566 段
[Bliss] 哦,是的。
第 1567 段
而在 AGI 喜欢上音乐之前,你就还没有实现 AGI,或者类似这样的说法。
第 1568 段
你不认为那就像是某种下一词元熵惊奇之类的东西在起作用吗?
第 1569 段
我不知道。
第 1570 段
我也不知道。我听很多古典音乐,也读很多诗。是啊,我确实会想,这里面是不是存在某种下一个词元的惊喜因素。
第 1571 段
是啊,也许吧。
第 1572 段
因为我的意思是,诗歌和音乐里的很多技巧基本上都是,你先有某种重复的结构,然后再来一个转折。就像,好吧,诗节,或者说分句一、二、三是一回事,然后分句四就像,好吧,现在我们进入下一个主题了。他们会巧妙地把玩惊喜究竟何时发生,以及用户的预期。
第 1573 段
而且纵观历史,这一点甚至也成立:随着音乐家推动音乐演变,他们会采用某种人们熟悉的微妙结构,然后稍微调整一下。他们会调整它,并加入一个出人意料的元素。这在古典音乐传统中尤其如此。但我想知道的是,这一切是否都只是熵——所以,打破结构或打破对称性,似乎是人类喜欢的东西。也许就是这么简单。
第 1574 段
是啊,而且我的意思是,伟大的艺术家会模仿,而知道该打破哪些规则也是重要的一环。从根本上说,这必定与作品的听众有关。比如哪一条规则才是应该打破的,这取决于用户或观众是否觉得这很有趣。
第 1575 段
你认为人类存在的意义是什么?
第 1576 段
有一部我非常喜欢的电视剧,叫《白宫风云》。在《白宫风云》里,有一个角色。他是美国总统,正在和一位同事讨论《圣经》。那位同事说了些类似“《圣经》说了X、Y和Z”的话。总统说:“是啊,但它也说了A、B、C。”那个人说:“你相信《圣经》在字面上完全真实吗?”
第 1577 段
总统说:“是的,但我也认为,我们两个人都没有聪明到足以理解它。”我认为,这里关于生命意义的类比是,在很大程度上,我们不知道该问什么正确的问题。所以我觉得自己非常认同《银河系漫游指南》对这个问题的那种说法,也就是基本上,如果我们能提出正确的问题,我们就更有可能找到人类存在的意义。
第 1578 段
所以短期来看,作为一种在某种搜索策略空间中的启发式方法,我们应该努力增加提出这类问题的人群的多样性,或者更笼统地说,增加提出这类问题的意识和有意识存在者的多样性。所以再说一次,我想我在这里还是会打出“我不知道”这张牌,但我确实认为,我们可以做一些有意义的事情,提高回答那个问题的可能性。
第 1579 段
你对提出正确问题这项任务赋予了这么大的价值,这很有意思。最重要的不是答案,而是问题。
第 1580 段
顺便说一句,当你试图与一个无法说话的人交流时,这一点会以一种非常痛苦的方式得到强化,因为很多时候,最后才丧失的是,他们还能以某种方式动一动嘴唇,或者移动某个部位,从而让他们能够表达“是”或“不是”。在那种情况下,非常显而易见的是,真正重要的是:你问他们的问题是否正确,能让他们用“是”或“不是”来回答?
第 1581 段
哇,这很有力量。好吧,Bliss,谢谢你所做的一切,谢谢你成为你自己,也谢谢你今天和我交谈。
第 1582 段
谢谢。
第 1583 段
感谢收听这场与Bliss Chapman的对话。现在,亲爱的朋友们,接下来是Noland Arbaugh,第一位在大脑中植入Neuralink设备的人。你在2016年遭遇了一场跳水事故,导致你瘫痪,肩膀以下毫无知觉。那场事故如何改变了你的人生?
第 1584 段
那算是一件离奇发生的事情。想象一下,你正跑进海里,虽然这里其实是个湖,但你正跑进海里,水大约齐腰深时,你就那样一头扎进去,潜入水下,穿过海浪之类的。我就是这么做的。然后我就再也没有浮上来。不确定发生了什么。我当时和几个家伙一起跑进水里。
第 1585 段
所以我对事情经过的看法其实就是,我的头部侧面挨了不知从哪儿挥来的一拳、手肘、膝盖、脚之类的东西。之后大约一个月,我头部左侧一直很痛。所以肯定是挨了相当重的一下。然后他们两个都浮上来了,而我没有。所以我面朝下在水里待了一阵子。我当时有意识。最后,我意识到自己再也憋不住气了。
第 1586 段
而我总是说:“喝了一大口。”别人,我不知道他们是否喜欢我这么说。听起来好像我没把这一切当回事,但我就是这样的人。而且我不知道,我是个非常放松、基本没什么压力的人。面对这一切中的很多事情,我都是见招拆招。我算是坦然接受了。就像:“好吧,那我接下来能做什么?我怎样才能让自己的生活每天哪怕只改善一点点?”
第 1587 段
一开始,我只是努力寻找某种办法,让身体尽可能多地康复,努力恢复,努力摆脱呼吸机,尽可能多地学习,这样离开医院后,我或许能活下去。然后感谢上帝,我身边有家人。如果没有父母和兄弟姐妹,我绝不可能走到今天。他们为我做了太多,说实话,多到我永远都感谢不完。
第 1588 段
而很多人并没有这些。很多和我处境相同的人,他们的家人要么没有能力照顾他们,要么说实话,就是不愿意。所以他们会被安置到某个地方的某种照护院里。因此值得庆幸的是,我有家人。我有一群很棒的朋友,一群很棒的大学伙伴,他们全都聚集在我身边,而我们至今仍然无比亲近。
第 1589 段
人们总是说:“如果你幸运的话,最后会有一两个高中朋友陪伴你一生。”我大约有10或12个高中朋友一直都在,而我们所有人至今仍然每年聚两次。我们称之为春季系列和秋季系列。最近那一次,我们所有人都打扮成了X战警。所以我扮成了X教授,简直太棒了。真的特别好。
第 1590 段
所以是啊,我身边有这么棒的支持系统。因此,四肢瘫痪也没那么糟。我一直都有人伺候。别人会给我拿食物和饮料,而我可以坐在那里,想看多少电视、电影和动漫就看多少。我想读多少书就读多少。我的意思是,这很棒。
第 1591 段
看到你能发现这一切中好的一面,真的很美好。回到之前,你还记得自己第一次意识到颈部以下瘫痪的那一刻吗?
第 1592 段
记得,是的。就在不知道什么东西撞到我头部时,我正面朝下趴在水里。我试图起来,然后意识到自己动不了,接着就突然明白了。我心想:“好吧,我瘫痪了。动不了。我该怎么办?如果我起不来,翻不了身,什么也做不了,那我最终会淹死。”我知道自己不可能永远憋住气,所以我就憋着气,想了大概10、15秒。
第 1593 段
我听其他人说过,像是看看酒类,我猜,把我从水里拉出来的那两个女孩是我最好的两个朋友。她们是救生员。其中一个说,我的身体在水里看起来像是在发抖,仿佛我在努力翻身之类的。但我知道,我立刻就知道了。我就是有点,我意识到从此以后,这就是我的处境了。
第 1594 段
也许如果我到了医院,他们就能做点什么。我在医院的时候,就在手术前,我还在试着安抚我的一位朋友。我当时把她从大学带来参加露营,她就趴在我身上号啕大哭,我就说:“嘿,会没事的。别担心。”我还开了几个玩笑,想让气氛轻松一点。护士给我妈妈打了电话,我当时说:“别告诉我妈妈。
第 1595 段
她只会压力很大。等我做完手术再打给她。”因为至少到时候她会得到一些答案,比如说到底我能不能活下来。而且我不想让她在整个过程中一直承受压力。但我知道。然后我手术后第一次醒来时,药劲特别大。他们用3种方式给我用了芬太尼,那感觉太棒了。
第 1596 段
我不推荐这么做,但我在芬太尼的作用下看到了一些疯狂的东西,而且那仍然是我吸毒时感觉最好的一次。药物,对不起,是用药时。我还记得在医院第一次见到我妈妈的时候。我一直号啕大哭。当时我嘴里插着呼吸管,完全说不了话之类的,我就开始哭,因为那更像是因为见到了她。
第 1597 段
我是说,整个状况显然相当艰难,但第一次看到她的脸真的很难受。不过,是的,我从来没有过这样的时刻:“天啊,我瘫痪了。这太糟糕了。我不想再继续活着了。”我一直都只是想:“我讨厌自己不得不经历这些,但坐在这里沉溺于痛苦也无济于事。”
第 1598 段
所以你立刻就接受了。
第 1599 段
[诺兰德] 对,对。
第 1600 段
这一路上有过低谷吗?
第 1601 段
有,有,当然。我是说,有些日子我真的什么都不想做。现在没那么严重了。就像过去几年里,我不太会有那种感觉了。到了现在,我更多的是只想尽一切可能让自己的生活变得更好。但在一开始,确实有过一些起伏。有些事情真的很难适应。
第 1602 段
首先,就最初那几个月来说,我承受的疼痛真的非常、非常难熬。我的意思是,我记得自己在医院里扯着嗓子尖叫,因为我以为我的腿着火了。很明显,我什么都感觉不到,但那全是神经痛。所以那一晚真的很难熬。我让他们尽可能多给我止痛药。他们说:“你能用的剂量已经全用上了,所以就只能想办法忍着。
第 1603 段
去想些开心的事情之类的。”所以那是一个相当低落的时刻。然后时不时地,有些事很难受,比如意识到有些我这辈子想做的事,我再也做不了了。我一直想成为丈夫和父亲,而我只是觉得,现在作为一个四肢瘫痪的人,我做不到了。也许有可能,但我不确定自己是否会让所爱的人经受那些,比如不得不照顾我之类的。
第 1604 段
不能出去运动。我从小就是个非常热爱运动的人,所以这相当难受。还有些小事,当我意识到自己再也做不了时,也很难受。比如,能够捧着一本书、闻到书的气味,这件事真的很特别。那种触感、质感、气味,比如翻动书页时的感觉,我就是很喜欢。我再也做不到了。就是这类小事。满2年这个节点相当难熬。
第 1605 段
他们说,到了2年时,就动作和感觉而言,基本上你能恢复多少,也就恢复到头了。所以最初2年里,我脑子里唯一想的就是尽我所能地尝试活动手指、双手、双脚,一切可能的办法,努力恢复感觉和运动能力。然后到了满2年的节点,也就是2018年6月30日,我真的很难过,因为我的状况差不多就停在这里了。
第 1606 段
然后只是偶尔会毫无缘由地这样,但我从来没有长时间抑郁过。只是对我来说,那似乎从来都不值得。
第 1607 段
是什么给了你力量?
第 1608 段
我的信仰。我对上帝的信仰是一个重要因素。我明白这一切都有其目的。即使那个目的与Neuralink毫无关系,即使那个目的是,《圣经》里有一个关于约伯的故事,我觉得那是一个非常、非常广为人知的故事,讲的是约伯遭遇了所有这些可怕的事情,而在整个过程中,他始终赞美上帝。
第 1609 段
我曾经以为,而且我觉得很多人在一生的大部分时间里都会认为自己就是约伯,认为自己才是那个正在经历可怕之事的人,他们只需要在整个过程中赞美上帝,一切最终都会好起来。事故发生后的某个时候,我意识到我可能不是约伯,我可能是他的某个被杀害、绑架或从他身边夺走的孩子。
第 1610 段
所以它讲的是发生在你周围那些你所爱的人身上的可怕事情。那么也许在这个情形里,我妈妈才是约伯,她必须熬过一件极其艰难的事,而我只需要努力让这一切对她来说尽可能好过一些,因为真正经受这场巨大考验的人是她。这给了我很大的力量。当然,还有我的家人。
第 1611 段
我的家人和朋友,他们每天都给我所需的全部力量。所以身边有这样出色的支持体系,让事情容易了许多。
第 1612 段
从我在网上看到的关于你的一切,包括你的直播和你如今的样子,我真的很钦佩,可以说,你对生活始终不渝的积极态度。你一直都是这样吗?
第 1613 段
是的,是的。我的意思是,我一直觉得,凡是我想做的事,我都能做到。从来没有什么事大到无法实现。比如无论我下定决心做什么,我都觉得自己能做到。我想做的并不多。我想四处旅行,过一种有点像吉卜赛人的生活,去打各种零工。我曾梦想周游欧洲,然后当一个,比如,我不知道,在威尔士或爱尔兰之类的地方当牧羊人。
第 1614 段
然后再去意大利当渔夫,花大约1年时间做所有这些事。比如这些事都太老套了,但我只是觉得,去旅行、做各种不同的事会非常有趣。而且我也总是看到身边人最好的一面。我也一直努力善待他人。还有,我是跟着妈妈长大的,她就像是全世界最积极、最有活力的人。我们都就是喜欢与人打交道的人。
第 1615 段
我就是很能和人相处。我真的很喜欢结识新朋友,所以我就是什么都想做。我一直差不多就是这个样子。
第 1616 段
考虑到你经历的一切,看到愤世嫉俗没有占据你的内心,真是太好了。
第 1617 段
是的。
第 1618 段
那是你有意做出的选择吗,就是你不会让这件事一直击垮你?
第 1619 段
是的,有一点。另外,这也差不多就是我的性格。就像我说的,无论遇到什么,我都随遇而安。我以前总是跟别人说,比如我不太会为事情焦虑。每当看到别人感到焦虑时,我就会说:“这不难。别焦虑就行了。”你只需要这么做。他们会说:“事情不是这样运作的。”我会说:“对我有用。比如别焦虑,一切都会没事的。
第 1620 段
比如一切都会解决的。”显然,并非一切总会顺利,也不是所有事情每次都会得到最好的结果,但我只是觉得,从我小时候起,焦虑在我的生活中就没有任何位置。
第 1621 段
你被选中成为第一个在大脑中植入Neuralink设备的人类,那是一种怎样的经历?你当时害怕吗,还是兴奋?
第 1622 段
不,不,挺酷的。(Lex笑)比如我从来没害怕过。我得仔细考虑很多事情。我该这么做吗?比如成为第一个人?我可以等到第2个或第3个,然后用上更好的Neuralink版本。比如第1个可能无法运行。也许它实际上会有点糟糕。它会是植入人体的有史以来最差版本。所以我为什么要做第1个?比如我已经算是被选中了。
第 1623 段
我可以直接告诉他们,比如:“好吧,去找别人,然后我来做第2个或第3个。”比如我确信他们会允许我这么做。反正他们也在找几个人。但最终,我当时想,我不知道,第一个去做某件事有一种特别之处。挺酷的。
第 1624 段
我一直觉得,如果有机会,我会想成为第一个做某件事的人,这看起来是个相当不错的机会,而且我从来没有害怕过。我觉得我的,比如,信仰在其中发挥了很大的作用。我一直觉得上帝在为某件事预备我。我甚至有点希望不是这件事,因为我曾多次与上帝谈到,我不想以四肢瘫痪者的身份做这些事。我告诉他:“我会出去和人们交谈。
第 1625 段
我会出去周游世界,在体育场里对数千人讲话,讲述我的见证,这些我全都会做,但先治好我。别让我坐在轮椅上做这一切。那太糟了。”我猜这场争论是他赢了。我其实没有多少选择。我一直觉得有什么事正在发生。
第 1626 段
然后看到,我猜,我多么轻松地通过了面试流程,以及一切发生得多么迅速,所有这一切中的各种条件如何仿佛恰好都具备了,这只是在告诉我,比如,随着手术越来越近,这只是在告诉我,这一切都是注定要发生的。这一切都是命中注定的。所以我不该害怕接下来会发生的任何事。因此我没有害怕。
第 1627 段
我一直对自己说:“你现在是这么说,但手术一来,你可能马上就会慌得不行。你可是马上要做脑部手术了。”脑部手术对很多人来说都是件大事,但对我来说更是如此。因为那是我仅剩的一切。
第 1628 段
有太多次我都在说:“感谢上帝,你没有夺走我的大脑、我的个性和我的思考能力,还有我对学习的热爱、我的品格,一切的一切,真的太感谢你了。只要你给我留下这些,我想我就能过下去。”而我正要让人进去翻弄我的大脑,他们就像在说:“嘿,我们要往你脑子里放些东西。希望一切顺利。”
第 1629 段
所以这件事确实让我有所迟疑。但就像我说的,一切都进行得那么顺利,我一刻都没料想过会出任何问题。而且,我在 Barrow 那边和 Neuralink 那边认识的人越多,就越觉得他们简直是世界上最令人钦佩的一群人。我怎么强调都不足以表达我有多么愿意把生命托付给这些人,以及他们每个人给我留下了多么深刻的印象。
第 1630 段
看到他们脸上的兴奋,走进一个房间,或者坐着轮椅进入一个房间,看到所有这些人望着我,就像在说,我们太兴奋了。我们一直在为此拼命工作,而它终于要发生了。这种兴奋特别有感染力,只会让我更想去做这件事,帮助他们实现梦想。我也不知道,这太有成就感了。说实话,我真为他们所有人感到高兴。
第 1631 段
手术当天是什么样的?你什么时候醒的?有什么感觉?
第 1632 段
对。
第 1633 段
一分一秒地讲。
第 1634 段
对。
第 1635 段
你当时慌了吗?
第 1636 段
没有,没有。我本来以为自己会慌,但随着手术临近,前一天晚上、当天早上,我只是很兴奋。来吧,把这件事做成。我想我事先在电话里对埃隆说过类似的话。我们当时在用 FaceTime,我说:“开始大干一场吧。”他说:“那就来吧。”我也不知道,我并不害怕。然后我们起床了。我想我们大概得在早上 5:30 到医院。我想手术大概是早上 7:00。
第 1637 段
所以我们醒得很早。我不确定那天晚上我们有多少人睡了觉。5:30到了医院,走完了差不多所有术前流程。每个人都特别友善。埃隆本来应该早上到场,但他的飞机出了点问题,所以最后我们用 FaceTime 通了话。那很酷。我说出了我人生中最棒的妙语之一。那通电话之后,我跟他挂断了通话。
第 1638 段
当时我周围大概有20个人,然后我就说,我只希望他跟我说话时没有因为见到我而太过激动。
第 1639 段
不错。
第 1640 段
对,挺好的。
第 1641 段
说得好。
第 1642 段
对,对。
第 1643 段
你是提前想好那句话的,还是灵光一现?
第 1644 段
没有,就是灵光一现。我当时想:“这话感觉挺合适。”在手术室里,我问能不能在开始前祈祷一下。所以我算是为整个房间祈祷了一番。我祈求上帝,如果我出了什么事,希望您能陪伴我妈妈。也请让她在外面平静下来。醒来后,我对我妈妈搞了个小恶作剧。我不知道你有没有听说过。
第 1645 段
听说了,我读到过。
第 1646 段
对,她很不高兴。
第 1647 段
你能跟我讲讲那个恶作剧吗——对,这是件————你现在后悔那么做吗?
第 1648 段
不,不,一点也不后悔。那是我提前跟我的朋友贝恩聊过的一件事。我当时说:“我真的很想对我妈妈搞个恶作剧。”非常明确,就是我妈妈。她很容易上当。我想她甚至有一次做了膝盖手术。她做完膝盖手术出来后,整个人昏昏沉沉的。她说:“我的腿没知觉了。”我爸爸看着她,说:“你已经没有腿了。他们不得不把你的两条腿都截肢了。”
第 1649 段
我们一直都对她做非常过分的事。我太惊讶了,她居然还爱我们。但手术刚结束时,我真的很担心自己会太过昏昏沉沉,好像神志不清。我以前有一次接受过麻醉,它把我搞得很糟。之后有一段时间,我就像完全没法正常活动。而且我好像说了很多事情,我当时就想,我真的很担心自己会开始,我不知道,就像爆出一些猛料。
第 1650 段
而且我自己甚至都不会知道,也不会记得。所以我当时就想:“求求上帝,别让那种事发生。也请让我足够清醒,好对我妈妈这么做。”于是手术后她走了进来。那好像是他们在我手术后第一次得以见到我。她只是看着我,说:“嗨,那个,你怎么样?你还好吗?感觉怎么样?”
第 1651 段
我就那样看着她,我觉得麻醉起了作用,脸上带着一种非常昏昏沉沉、有点困惑的表情。就像:“你是谁?”然后她就开始环顾房间,看着外科医生、那些医生,像是在说:“你们对我儿子做了什么?你们现在就得把他治好。”眼泪开始往下流。我看到她有多么惊慌失措。我当时想:“我不能让这继续下去了。”所以我就说:“妈妈,我没事。那个,没关系的。”
第 1652 段
可即便如此,她还是对此不高兴。她到现在还说,总有一天要报复我。不过我的意思是,我不知道。我不知道那会是什么样子。
第 1653 段
这是一场终生的战斗。
第 1654 段
对,挺好的。
第 1655 段
从某种意义上说,这证明了你依然有—— ——我想要的就只是这个—— 幽默感。
第 1656 段
我想要的就只是这个。而且我知道,对她做那种特别刻薄的事会让她—— —— —— ——对。
第 1657 段
让她知道你依然还在,你爱她。
第 1658 段
[诺兰] 对,没错,没错。
第 1659 段
这种做法挺阴暗的,但我喜欢。你第一次感觉到自己能够用 Neuralink 设备影响周围世界,是什么时候?
第 1660 段
对,我第一次稍微尝到这种滋味,其实是在手术后没多久。Neuralink 团队的一些人拿来了一台小 iPad,类似一个小平板屏幕,然后调出了八个不同的通道,记录我的一些神经元脉冲。他们把它放在我面前。就像,这是你的大脑实时放电。太酷了。
第 1661 段
我的第一个念头是:“我的意思是,既然它们现在正在放电,那就看看我能不能以某种方式影响它们。”于是我开始试着动动手指,然后就这样逐个查看那些通道,而我当时做的一件事,就是上下移动食指。然后我看到一个黄色脉冲出现在大概最上面一排、从左往右第三个框之类的位置。每次我这么做,都会看到这个黄色脉冲。我当时就想:“哦,这很酷。”
第 1662 段
我周围的所有人都在问:“你看到了什么?”我说:“看,看这个。看最上面一排,从左往右第三个框里的这个黄色脉冲。那就是我,就在这里、这里、这里。”然后大家全都激动坏了。他们开始鼓掌。我说:“这真的太没必要了。”
第 1663 段
太棒了。
第 1664 段
这就是应该发生的情况,对吧?
第 1665 段
所以你想象自己一次活动一根手指,然后看到你好像能注意到某种变化,接着当你活动食指时,你就想:“哦。”
第 1666 段
对,我当时在动我所有的手指,想看看会不会发生什么。还有很多其他的变化,但那个巨大的黄色尖峰最引人注目。我敢肯定,如果我盯着它看得足够久,或许能把大约100种不同的变化都对应出来。但我注意到的就是那个巨大的黄色尖峰。
第 1667 段
也许你可以谈谈活动手指是什么感觉,想象一下,比如说,活动食指所需要的那种精神上的、认知上的努力。这做起来有多容易?
第 1668 段
对我来说相当容易。这件事,在最开始,我出事故之后,他们告诉我要尽可能多地尝试移动身体。即使你动不了,也要不断尝试,因为这会形成新的,比如说,神经通路,或者脊髓中的通路,来重新连接这些东西,希望有朝一日能恢复一些活动能力。
第 1669 段
这太不可思议了。
第 1670 段
是啊,我知道。这很离奇,但我——所以,不断尝试移动身体是康复过程的一部分?
第 1671 段
对,尽你所能。
第 1672 段
然后就是,神经系统会自行发挥作用。它开始重新连接。
第 1673 段
是的。对有些人来说,它会开始重新连接。有些人则始终不起作用。有些人会做到。比如我,我恢复了一些对肱二头肌的控制,也就这样了。如果我足够努力,我可以让几根手指动一动。不是那种随叫随动。更像是,如果我试着移动,比如说我的右手小拇指,然后不断尝试移动它,几秒钟后,它就会动一下。所以我知道那里是有反应的。
第 1674 段
比如我知道,而且这种情况会发生在我几根不同的——手指之类的部位。但对,这就是他们让你做的事。当时我住院时,有个人进来告诉我,有一个人恢复了大部分控制能力,他每天想的其实都是走路,就是一遍又一遍地想走路这个动作。所以我这样尝试了很多年。我试着单纯想象自己走路,这很难。
第 1675 段
很难想象迈出,嗯,一步所涉及的所有步骤,比如必须活动的所有部位,比如为了迈出一步,你的整条腿上必须发生的所有激活动作。
第 1676 段
但你不只是在想象。你就像是在实际做,对吧?
第 1677 段
我在尝试,是的。所以这就像,是一遍又一遍地想象为了迈出一步我必须做什么,因为我们所有人都不会去思考这件事。你想走路,然后就迈出一步。你不会去想身体里正在发生的所有不同事情。所以我必须尽我所能在脑海里重现这个过程。然后我一遍又一遍又一遍地练习。
第 1678 段
所以,这并不是第三人称视角,而是第一人称视角。你就像是——不是说你在想象自己走路,而是你真的在做这些动作,一切都和你实际走路时完全一样。
第 1679 段
对,这很难。一开始很难。
第 1680 段
是那种令人沮丧的难,还是认知层面上真的很难?是哪一种?
第 1681 段
两者都是。说来也怪,其实有一部《杀死比尔》电影里有这样一个场景:她像是瘫痪了,我不知道,好像是因为体内有某种药物。然后她不知怎么设法进了一辆卡车的后部之类的地方,她盯着自己的脚趾,说:“动。”就是让你的大脚趾动起来。过了荧幕上的几秒钟,她做到了。然后她对自己身上每个类似的身体部位都这样做,直到她又能动了。
第 1682 段
我一连好几年都在这么做,就是盯着自己的身体,说:“动动你的食指。动动你的大脚趾。”有时会大声说出来,有时只是在心里想。我用各种不同的方式尝试这样做,试图恢复一些活动能力。
第 1683 段
而且这很难,因为它实际上非常耗费体力,对我的身体来说就是很累,这一点我以前绝对想不到,因为我其实并没有在动,但感觉像是有什么东西在不断积聚。我不知道,我唯一能描述它的方式就是,好像有些信号没能从我的大脑向下传递,因为我的脊髓中间有一处断裂。所以信号从大脑向下传,然后再从我的手传回大脑。
第 1684 段
所以感觉那些信号会卡在我正试图移动的身体部位里。它们就不断积聚、不断积聚、不断积聚,直到爆发。等它们爆发后,我会有一种非常奇怪的感觉,仿佛一切都逐渐消散,重新恢复到正常水平,然后我再来一次。这也像是一种疲劳,就像肌肉疲劳,但你其实根本没有动肌肉。非常、非常怪异。
第 1685 段
而且,如果你试着盯着某个身体部位,或者想着某个身体部位,让它活动两、三、四个小时,有时甚至八个小时,这对精神的消耗非常大。需要高度集中注意力。刚开始时要容易得多,因为我没办法控制房间里的电视之类的东西。我无法控制周围环境中的任何东西。所以最初几年里,我做的很多事情就是盯着墙看。
第 1686 段
所以很显然,我思考了很多,也一遍又一遍、反反复复地尝试活动身体。
第 1687 段
所以你从来没有放弃过那种希望?
第 1688 段
没有。
第 1689 段
本质上就是刻苦训练?
第 1690 段
对,而且我现在还在这么做。我会下意识地这么做。老实说,我认为这对使用 Neuralink 时的很多事情帮助很大。这也是我前几天在 Neuralink 奥斯汀设施举行的全员会上谈到的一件事。
第 1691 段
顺便说一句,欢迎来到奥斯汀。
第 1692 段
是啊,嘿,谢谢,伙计。我上过学——帽子不错。
第 1693 段
嘿,谢谢,谢谢,伙计。超级工厂特别酷。我在得克萨斯农工大学上过学,所以我以前就在这一带待过。
第 1694 段
所以你应该对我说欢迎。
第 1695 段
是啊。
第 1696 段
欢迎来到得克萨斯,Lex。是啊,我懂你的意思。
第 1697 段
不过,是啊,我刚才在说,他们让我做的很多事情,尤其是在一开始,嗯,我现在仍然在做,就是身体映射。比如说,屏幕上会显示一只手或一条手臂的可视化图像,我必须做出那个动作,他们就是这样训练算法,让它理解我想做什么。所以我觉得,这让整个过程对我来说非常顺畅。
第 1698 段
那真的非常、非常酷。所以,是啊,了解到这一点很不可思议,因为我对使用这个接口进行身体映射的流程之类的学到了很多。得知你实际上一直在接受训练,要在这项任务上达到世界级水平,真的很酷。
第 1699 段
对,对。我不知道其他四肢瘫痪者,比如其他瘫痪的人,会不会放弃。我希望他们不会。我希望他们继续努力,因为我听其他瘫痪的人说过,比如永远不要停下来。他们告诉你是两年,但你永远不知道。人体能够做到不可思议的事情。所以我听其他人说过:“不要放弃。”
第 1700 段
比如,我记得有一个女孩曾通过一些家人跟我说,她已经瘫痪了18年,而她一直都在努力尝试活动她的食指,后来大约18年后,她终于又能动食指了。所以,我知道这是可能的,我永远不会放弃这样做。我躺着的时候会做。比如看电视时,我会发现自己也在做,有点儿几乎是自然而然地就在做。
第 1701 段
这只是我已经习惯去做的一件事,我不知道,我觉得自己永远都不会停下来。
第 1702 段
听到这件事真的很棒,因为我觉得从长远来看,这是那种真的能带来回报的事情。因为这就是训练。此刻你并没有直观地看到这种训练的成果,但就像,有一个奥运级别的神经系统正在为某件事做准备。
第 1703 段
说实话,我觉得这是Neuralink给予我的某种东西,我怎么感谢他们都不为过。就像,我怎么表达感激都不足以回报他们,因为我能够亲眼看到,自己所做的事情确实产生了一些效果。这是一个很重要的原因,让我现在知道自己会永远坚持下去,因为在Neuralink之前,我每天都在做,只是假定有些事情正在发生。
第 1704 段
我并不是真的知道。我没有恢复任何活动能力或感觉,什么都没有。所以据我所知,我可能一直都在徒劳地撞一堵砖墙。而有了Neuralink,我可以实时看到所有信号的出现,也可以看到我所做的事情确实能够被映射。当我们开始做点击校准之类的事情时,当我尝试用食指点击来执行左键单击时,它确实能识别出来。
第 1705 段
这改变了我对于通过重新训练让身体活动起来有哪些可能性的看法。所以,是的,现在我永远都不会放弃。
第 1706 段
而且,这也是一个信号,说明那里仍然有一个无比强大的大脑,它就像—— ——没错。
第 1707 段
而且随着技术发展,大脑,我是说,人体最重要的东西就是大脑。它可以完成很多控制。那么,当你第一次能够摆动食指,并看到外部环境作出响应时,是什么感觉?就像那一点点—— ——是的。
第 1708 段
不管在哪里,只是照你说的,表现得太戏剧化了。
第 1709 段
是的,那非常酷。我是说,确实很酷,但我一直跟人们说,这对我来说是说得通的。就像,我的大脑里仍然有信号在产生,而只要在它附近放置某种能测量、能记录这些信号的东西,那么你就应该能以某种方式把它可视化。就像亲眼看到它发生。所以这对我来说并不是很意外。我只是觉得:“哦,酷。”我们找到了一个。
第 1710 段
就像是我们找到了一个可行的东西。看到他们的技术奏效了,而且他们如此努力付出的一切似乎都将得到回报,这很酷。但那时我还没有移动过光标之类的东西。那时我已经用电脑进行过交互之类的操作。所以这一切就是很说得通,也很酷。那时我对 BCI 也不太了解,所以我不知道这实际上算是迈出了怎样的一步。
第 1711 段
比如我不知道这是不是一件大事,还是说这只是类似于,好吧,我们能走到这一步很酷,但我们其实希望以后能做到好得多的东西。就是,好吧。我只是以为他们知道它启动了。所以我就觉得,酷。这个很酷。
第 1712 段
那么,你有没有查阅过要植入你体内的硬件规格?比如线的数量,诸如此类的?
第 1713 段
有啊,这些我全都看了,但对我来说完全像天书。我当时就想,好吧,线,64 根线,16 个电极,1,024 个通道。好吧。这个计算对得上。
第 1714 段
听起来没错。
第 1715 段
对。
第 1716 段
你第一次能够移动鼠标光标是什么时候?
第 1717 段
我知道,那肯定是在最初大概一周之内,或者一两周的时候,我第一次能够移动光标。而且还是那样,这对我来说算是挺说得通的。感觉不是什么特别了不起的事。就是,好吧,嗯,我该怎么解释呢?当你身边的每个人都开始为你做成的某件事鼓掌时,你很容易就会说:“好吧,我做了一件很酷的事。这在某种程度上挺了不起的。”
第 1718 段
这究竟意味着什么、究竟是什么,我当时其实还没有真正意识到。所以还是一样,我知道,我试图移动身体的某个部位,然后通过某种机器学习算法把这种尝试映射出来,从而识别我的脑信号,再利用这些信号让我能够控制光标,这一切对我来说都算是说得通。我不了解其中所有的细节,但我当时就想,我的大脑里仍然有信号在发放。
第 1719 段
它们只是无法通过,因为我的脊髓里像是有一个缺口。所以它们没法一路向下再返回,但它们仍然存在。所以当我第一次移动光标时,我当时想:“这很酷,但我本来就预料到这应该会发生。”就像这对我来说是说得通的。当我第一次仅凭意念移动光标,而没有在身体上尝试移动时,所以我想我可以稍微讲一下这个。
第 1720 段
比如尝试运动和想象运动之间的区别。
第 1721 段
对,这是一个很有意思的区别。
第 1722 段
对。
第 1723 段
从一种到另一种。
第 1724 段
对,对,对。所以,像“尝试运动”就是我在身体上试图移动,比如说我的手。我试图尝试把手向右、向左、向前和向后移动。而这些全都是尝试。尝试把手指抬起和放下。尝试踢一下之类的。即使你看不出来,我也在身体上试图做所有这些动作。这就像是我试图耸耸肩之类的。
第 1725 段
那些全都是尝试运动。我最初几周就是这么做的,当时他们准备让我控制光标。在进行身体映射时,我会尝试做这个,尝试做那个。Nir让我想象自己在做那些动作时,我觉得好像有点道理,但这不是人们会练习的事情。比如,如果你小时候刚开始上学,他们说:“好,用这支铅笔写下你的名字。”
第 1726 段
于是你就照做。“好,现在想象你正用那支铅笔写自己的名字。”孩子们可能会想,我猜这好像有点道理。然后他们就会去做。但这不是我们受过教导的事情。我们学的全是如何在身体上做事情。我们会思考思想实验之类的东西,但那不像是在身体上实际做某件事。更多的是你在某些情况下会怎么做。
第 1727 段
所以,想象运动从来没有真正让我产生共鸣。我想,也许你可以把它描述成职业运动员挥动棒球棒或高尔夫球杆之类的情形。就是想象你应该怎么做。但接着你会立刻真正动手去做,拿起球棒,然后做出你一直在想象的动作。所以我没有那种联系。
第 1728 段
所以,让我想象某件事和让我尝试去做相比,我在心理上做不了太多。我只能算是接受正在发生的事情,然后尝试。但尝试运动这件事,我完全能理解。比如,如果我尝试移动,大脑中就会发出一个信号。只要他们能捕捉到这个信号,就应该能够把它映射到我正试图做的事情上。
第 1729 段
所以,当我第一次那样移动光标时,感觉就是:“对,这本来就应该发生。”我对此并不惊讶。
第 1730 段
但你能澄清一下吗?想象运动和尝试运动之间应该存在区别吗?
第 1731 段
是的,区别只是在想象运动时,你完全没有尝试移动。
第 1732 段
你是在想象自己正在做。然后从理论上讲,在这两种不同情况下,被激活的应该是大脑的不同区域吗?
第 1733 段
[布利斯] 对,不一定。我觉得所有这些信号仍然可以在运动皮层中得到表征,但我认为区别与想象某件事和尝试去做这件事之间的自然程度有关——明白了。
第 1734 段
[布利斯] 以及随着时间推移,由此产生的那种疲劳。
第 1735 段
顺便说一下,麦克风前的是布利斯。所以这些只是用不同的方式提示你,让你逐渐得出你最终得出的那个结论。
第 1736 段
[诺兰德] 对,对。
第 1737 段
尝试运动听起来确实是正确的做法——试试看。
第 1738 段
对,我是说,我觉得这说得通。
第 1739 段
因为对我来说,要是想象,我会开始在脑海里进行视觉化,就是在脑子里想象。要是尝试,我实际上会开始试着去……就是,我是说,我一辈子都在练格斗运动,比如摔跤。当我想象一个动作时,你看,我就像是在动我的肌肉。
第 1740 段
没错。
第 1741 段
就像是几乎有一点激活的感觉,而不是像看着一幅自己正在做这件事的画面那样。
第 1742 段
是的,我觉得这就像是任何人自然而然都会做的事。如果你试着让某人想象自己在做某件事,他们可能会闭上眼睛,然后开始真的做出动作。但就是——就是没能领会。
第 1743 段
是的。很难。一开始非常难。
第 1744 段
但尝试奏效了。
第 1745 段
尝试奏效了。它就像本该有的那样奏效了。非常顺利。
第 1746 段
[布利斯] 我记得有一个星期二,我们当时在摆弄这个,然后我想,我忘了你用了哪个脏字,但当你发现自己可以直接控制光标时,你嘴里蹦出了一个脏字。
第 1747 段
对,就是这样。这让我大为震撼。真不是有意说双关语,我第一次只凭想法、不尝试做出动作就移动了光标时,真的大为震撼。这是我在那之前几周逐渐发现的:随着我对光标的控制越来越好,模型也变得更好,然后对我来说就变得更容易了,就像,我不需要那么用力尝试就能移动它。
第 1748 段
这其中有一部分,是有一天我观察自己大脑的信号时,甚至还跟他们谈过的事情。我当时在看,当我尝试向右移动并看着屏幕时,就像我看到了那些尖峰。就像在我真正尝试移动之前,我就看到了尖峰,也就是正在发送的信号。
第 1749 段
我的想法是,这只是因为当你要移动手或任何身体部位时,信号会在你真正移动之前发出,必须一路传下去再传回来,然后你才会真正做出任何动作。所以这里存在延迟。而且我注意到,在我真正尝试移动之前,我的大脑里就已经有某种活动了,就像我的大脑在预判我想做什么。
第 1750 段
然后这一切就开始有点儿,我不知道,像是在我的脑海里慢慢酝酿。它就那么一直存在,像是始终藏在意识深处。它居然能做到这一点,真是太奇怪了。这似乎也说得通,但我想知道,对于使用 Neuralink 来说,这意味着什么。
第 1751 段
然后,当我不断尝试做出动作、不断摆弄光标时,我看到,随着光标控制变得更好,它也越来越能预判我的动作以及我想让它做什么。就像对光标移动、对我想做什么的判断,一点一点地变得更好。后来有一天,我在玩 Webgrid 时,只是很随意地在开始尝试移动之前看向了一个目标。
第 1752 段
我只是想要克服这个问题,就像训练自己的眼睛开始提前看。比如,好,我现在在这个目标上,但如果我看向这边这个目标,我知道自己或许能更快一点到那里。然后我看了过去,光标一下子就冲过去了。太疯狂了。我不得不停下来缓一缓。我当时就像:“这不应该发生。”我一整天都在笑,兴奋得不得了。我说:“各位,你们知道这真的能用吗?
第 1753 段
就像我只要想一下,它就发生了。”而他们其实一直以来都在这么说。就像,“真不敢相信你完全是在用意念做这一切。”我就说:“对,但这真的是用意念吗?”就像,我是在尝试移动,而它只是捕捉到了这个信号,所以感觉并不像是用意念。当我第一次像那样移动它时,那感觉,哦,天啊。
第 1754 段
这让我想到,这项技术、我正在做的事情,其实远比我曾经以为的要了不起得多、得多。它比我曾经以为的酷多了。而且它就此开启了一个充满可能性的全新世界,比如这项技术可能会带来什么,以及借助它我或许能够做到什么。
第 1755 段
因为你第一次感觉到,这就像是数字心灵感应。就像你在用意念控制一台数字设备。
第 1756 段
[诺兰] 对。
第 1757 段
我是说,那是真正的发现时刻。真的很酷。就像你发现了某种东西。我见过科学家谈论那种重大的恍然大悟时刻。就像诺贝尔奖得主,他们会有这种“我的天啊”的感觉。
第 1758 段
对。
第 1759 段
就像,哇。
第 1760 段
感觉就是那样。我并没有觉得,怎么说呢,我觉得自己发现了某种东西,但只是对我而言。未必是对整个世界或整个领域而言。对我来说,那就像一个恍然大悟的时刻。就像,“哦,这真的有用。”就像,很显然,它确实有用。所以我现在一直都是这么做的。我会把尝试做动作和想象做动作混合起来。
第 1761 段
我会把它们全都结合起来做,因为我发现两者之间存在某种相互作用,能让光标的操作效率最大化。所以并不完全是二选一。并不是说,我只使用尝试做动作,或者只使用想象中的动作。更多的是,我会同时使用它们,而且我也可以只用其中一种。我完全可以只想着自己正在做的事。但我不知道。我喜欢琢磨尝试这些东西。
第 1762 段
我也喜欢直接拿这些东西做实验。比如时不时地,我脑中会冒出一个想法:“嗯,不知道这样有没有用。”然后我就会直接开始这么做,之后我会告诉他们:“顺便说一句,我刚才没有按你们希望的那样做。我想到了一个点子,想试一试,所以就试了。看起来似乎有用,所以也许我们应该稍微探索一下。”
第 1763 段
所以我认为,这个发现不只是对你而言的,至少从我的角度来看,这是让其他所有使用 Neuralink 的人发现这是可能做到的。比如,我不认为这是一件显而易见、甚至确实可能做到的事。这就像,我之前对 Bliss 说的,就像4分钟跑完一英里。人们曾认为4分钟跑完一英里是不可能的,而第1个人做到之后,所有人就都开始做到了。
第 1764 段
所以,比如只要证明这是可能的,就能为任何人都可以不这么做铺平道路。关键是,这件事实际上是可能的。你不需要做尝试动作。你可以直接来。这太疯狂了。
第 1765 段
这太疯狂了,太疯狂了。
第 1766 段
对于不了解的人,你能解释一下 Link 应用是如何工作的吗?你有一场关于这个话题的精彩直播。我想,那是你在 X 上的第一次直播,介绍了这款应用。你能简单描述一下它是如何工作的吗?
第 1767 段
可以,所以它就是 Neuralink 创建的一款应用,用来帮助我与电脑交互。在 Link 应用里,有几个不同的设置和不同的模式,还有一些我可以用它做的事情。比如我们稍微提到过的身体映射。还有校准。校准是我真正获得光标控制能力的方式。也就是校准我大脑中正在发生的活动,把它转化为光标控制。所以它会生成模型。
第 1768 段
我想,他们采用的衡量方式类似于时间。比如进行5分钟,校准就会给我一个达到相应水准的模型。然后,如果我在里面进行10分钟、15分钟,模型就会逐步变得更好。所以一般来说,我进行得越久,模型就会变得越好。
第 1769 段
这真的很酷,因为你经常提到模型。模型就是你完成校准步骤后构建出来的东西。你还谈到,有时你会玩像《贪吃蛇》这样非常难的游戏,只是为了看看模型有多好。
第 1770 段
对,对,所以《贪吃蛇》算是我检验模型的试金石。如果我能相当不错地控制《贪吃蛇》,那我就知道自己有了一个很好的模型。所以没错,Link 应用具备所有这些功能。它现在也内置了 Webgrid。一般来说,我也是通过它连接电脑的。所以到目前为止,他们为它加入了很多语音控制功能,因此我可以说“连接”或者“植入物断开连接”之类的指令。
第 1771 段
只要我手边有那个充电器,我就能连接上。所以,我也是通过充电器连接 Link 应用,进而连接电脑的。我想连接时,必须把植入物充电器放在头顶上方来唤醒它,因为植入物在我不使用时一直处于休眠模式。我想有一个设置,可以让它每隔一段时间唤醒一次。
第 1772 段
所以,如果我只是想让它定期唤醒,我们可以把间隔设为半小时、5小时之类的。对,我会连接 Link 应用,然后完成各种事项。做当天的校准,也许再做身体映射。我让他们给我加了一个小小的作业标签页,因为我非常健忘,经常忘记做事情。所以我有很多他们希望我完成的数据收集事项。
第 1773 段
身体映射是数据收集的一部分吗,还是说那也是————对,是的。这是他们希望我每天做的事情,但我一直在偷懒,因为我一直在做太多媒体活动,又到处奔波。所以我一直————你已经超级有名了。
第 1774 段
是啊,作为第一位候选者,我在功课上偷懒得这么厉害,实在是很糟糕。不过没错,这只是他们希望我每天都做的事情,用来追踪 Neuralink 随时间推移表现得有多好,也为了有东西可以提交。我想是提交给 FDA,用来制作各种精美的图表之类的,并展示比如:“嘿,这就是 Neuralink,这就是它在第 1 天、第 90 天和第 180 天的表现”,诸如此类。
第 1775 段
校准步骤是什么样的?是类似向左移动、向右移动吗?
第 1776 段
这是一个气泡游戏。屏幕上会弹出一些黄色气泡。一开始是开环。所以说到开环,这是我至今仍未完全理解的东西,就是开环和闭环这回事。
第 1777 段
我和 Bliss 从技术层面谈了很久这两者之间的区别。
第 1778 段
好的。
第 1779 段
所以很想听听你这边的说法。
第 1780 段
开环基本上就是,我无法控制光标。光标会自行在屏幕上移动,而我会用意图跟随光标移向不同的气泡。然后算法会根据我这样做时接收到的那些信号进行训练。他们采用过几种不同的方法。他们称之为“中心向外目标”。也就是说,中间会有一个气泡,周围还有8个气泡。
第 1781 段
光标会从中间移向一侧。比如说,从中间到左边,再回到中间,然后向上,再回到中间,比如向右上方。他们会这样绕完整个圆圈。我会全程跟随那个光标,然后它会根据我的意图进行训练,也就是在整个过程中它预期我的意图是什么。
第 1782 段
你能具体说说吗,当你说“跟随”——可以。
第 1783 段
你指的不是用眼睛跟随,而是用意图跟随。
第 1784 段
对,所以一般在校准时,我会尝试做出动作,因为我觉得这样效果更好。我认为,随着校准的推进,模型越好,使用想象动作就越容易。
第 1785 段
等等,等等,等等,等等。所以,用尝试动作进行校准,会创建一个模型,让你之后能够非常有效地使用原力?
第 1786 段
是的,我试过用想象运动来做校准,但不知为什么,效果就是没那么好。所以那是从中心向外的目标。还有一种是屏幕上会随机弹出一个目标,做法也一样。我就是移动,跟着光标走到那个目标,不管它出现在屏幕上的什么位置。
第 1787 段
我试过用想象运动来做那些校准,但不知为什么,等我们进入闭环时,这些模型就是达不到那么高的质量水平。我还没有对此做太多尝试,所以也许我们现在采用的不同校准方式可能会让效果好一点。但我发现,在校准过程中会有一个节点,从那时起我可以使用想象运动。在那个节点之前,它并不怎么奏效。
第 1788 段
所以,如果我做45分钟的校准,最初15分钟我不能使用想象运动。不知为什么,它就是不奏效。到了某个节点之后,我就能隐约感觉到。我能看出它移动得不一样了。这是我能想到的最佳描述。就好像它又开始在我着手做某件事之前,就预判到我要做什么。
第 1789 段
所以在使用尝试动作 15 分钟后,到了某个时候,我大概能感觉到,当我把视线移向下一个目标时,光标开始有点跟上了。就像它开始理解了,它正在学习我接下来要做什么。
第 1790 段
所以首先,这真的很酷,我是说,你确实是这一切的真正先驱。你就像是在探索怎样最有效地完成其中的每一个方面,而且我想,这里面有太多经验教训可以汲取了。所以,感谢你以各种不同的、极其技术性的方式在这些事情上开拓先河。
第 1791 段
而且,听说以不同方式校准时,这种体验会有某种不同的感受,也很酷,因为我是说,我想你的大脑正在做不同的事情,所以才会有不同的感觉。然后,试着为那些感觉找到相应的词语和衡量方式,也会很有意思。
第 1792 段
但归根结底,你也可以衡量自己无论是在《贪吃蛇》还是 Webgrid 上的实际表现,你可以看到什么方法实际上效果好。而你是说,就开环校准而言,目前尝试做出动作的方式效果最好。
第 1793 段
对,对。
第 1794 段
所以在开环模式下,你不会得到自己已经做了某个动作的反馈。
第 1795 段
对——这会让人沮丧吗?
第 1796 段
不会,不会,我觉得这很合理。比如在开环模式下,我们做过有光标的,也做过没有光标的。所以有时候,就比如做中心向外任务时,开始校准时会有一个气泡亮起来,我就朝那个气泡推动。然后,当朝那个气泡推动了比如3秒后,气泡就会破掉,接着我再回到中间。所以我完全是凭自己的意图在做。反正它要学习的就是这个。所以我觉得,只要我按照他们希望我做的去做,就像沿着黄砖路走,一切就都会顺利完成。
第 1797 段
你引用了很多精彩的典故。气泡游戏好玩吗?
第 1798 段
好玩,他们每次让我做校准时总觉得特别过意不去。比如,“哦,我们马上要做一次40分钟的校准。”我就会说,“好啊,你们想做2次吗?”就像我总会主动问,不管他们需要我做什么,我都非常乐意去做。而且这并不难受。比如我可以躺在那里,或者坐在椅子上,和一些很棒的人一起做这些事情。我可以进行很棒的交谈。我可以给他们反馈。我可以谈各种各样的事情。我还可以在电视上随便放点什么当背景,算是把注意力分散在他们和电视之间。比如这完全不难受。我不介意。
第 1799 段
你会得到一个分数吗?比如说,你能在泡泡游戏里做得更好吗?
第 1800 段
没有,我很希望能有。我很希望——对。记下诺兰的建议。
第 1801 段
那就是——让它更有趣。游戏化。
第 1802 段
对,这正是我非常、非常喜欢 Webgrid 的一点,因为我太争强好胜了。比如 BPS 越高,分数越高,我就知道自己做得越好。我想我曾经问过其中一个人,他能不能给我某种关于校准的数字反馈,比如我想知道他们在看什么。
第 1803 段
比如:“哦,你进行校准时,我们会看到这样一个数字,而这意味着,至少在我们这边看来,我们认为校准进展得很好。”我会非常希望能有这个,因为我想知道自己做得好不好。但他们也告诉过我,比如:“对,这不一定是一一对应的。”从某些方面来说,它实际上并不意味着校准进展得很好。
第 1804 段
所以它并不是 100% 准确的,而他们也不想影响我的体验,或者让我根据那个数字改变做法。如果那个数字并不总能准确反映模型最终会怎么样,或者最终结果会怎么样,至少我是这么理解的。有一件我确实问过他们、而且我非常喜欢努力追求的事是,在校准接近尾声时,目标之间会有一个时间间隔。
第 1805 段
所以我喜欢在最后把那个数字保持得尽可能低。一开始,我每戳破一个泡泡之间可能会间隔 4、5、6 秒。但接近尾声时,我喜欢把它保持在大约 1.5 秒以下。或者如果可以的话,把每个泡泡之间的间隔缩短到大约 1 秒,因为在我看来,这可以很好地对应到 Webgrid 之类的东西上,在那里我知道,如果我每秒能击中 1 个目标,那我就做得非常、非常好了。
第 1806 段
这就有了,这是一种给校准打分的方法。比如看速度,看你能多快地从一个泡泡移到另一个泡泡。
第 1807 段
对。
第 1808 段
所以先是开环,然后进入闭环。
第 1809 段
闭环。
第 1810 段
闭环已经可以开始让你有所感知,因为你会得到反馈,比如模型有多好。
第 1811 段
对,所以闭环就是我第一次获得光标控制权,而按他们向我这个不懂这些东西的人所描述的那样,每次我和——在一起时,我都是房间里最笨的人——我喜欢这种谦逊。
第 1812 段
对,就是我在闭合这个环。所以现在实际上是我在完成这个不管是什么的环。我甚至不知道这个环是什么,他们从来没告诉过我。他们只是说有一个环。在某个时刻,它是开环的,我无法控制。然后我获得了控制权,它就闭合了。所以是我在完成这个环。
第 1813 段
那么校准通常需要多长时间?你说大概10、15分钟。
第 1814 段
嗯,对,他们正努力把这个数字降得相当低。我们最近一直在做的一项重要工作,就是尽可能把它降到最低。这样,如果这是人们每天都需要做的事,或者有些人需要每隔一天做一次,或每周做一次,他们就不希望人们长时间坐在那里进行校准。
第 1815 段
我想他们希望把它降到7分钟或更短,至少就我们目前的进展而言。要是永远都不必校准就好了。所以我相信,随着我们对大脑了解得越来越多,总有一天会做到的,我觉得那就是梦想。我想现在,为了得到非常、非常好的模型,我要校准40或45分钟。而我并不介意。
第 1816 段
就像我说的,他们总是很过意不去,但如果这能让我得到一个可以在 Webgrid 上打破这些纪录的模型,那我愿意继续校准足足 2 小时。
第 1817 段
我们来谈谈正事。那么,Webgrid。我看过一个演示,Bliss 在里面说,截至三月,你在 Webgrid 中选中了 89,000 个目标。你能解释一下这个游戏吗?Webgrid 是什么?随着你不断打破世界纪录,要成为 Webgrid 的世界级选手需要具备什么?
第 1818 段
是的。
第 1819 段
就像一位金牌得主一样,那个。
第 1820 段
是的,我想感谢,我想感谢所有帮助我走到这里的人,我的教练,还有我的父母,感谢他们每天早上 5 点开车送我去训练。还要感谢上帝。以及总的来说,我对自己这门技艺的投入。
第 1821 段
对运动员的采访总是那样,就像是套用那个模板。
第 1822 段
是的。
第 1823 段
所以 Webgrid 是一个会卖东西的网格。
第 1824 段
Webgrid 是,嗯。它真的就只是一个网格。他们想把网格做多大或多小都可以。网格上的一个方格会亮起来,你过去点击它。这是他们衡量脑机接口性能有多好的一种方式。所以非常直截了当。你只需点击目标。
第 1825 段
[Lex] 只会出现一个蓝色单元格,然后你应该把鼠标移到那里并点击它。
第 1826 段
对。所以我喜欢在那种更大的网格上玩,因为网格越大,BPS 就越高。它指的是比特每秒,也就是你每点击一个目标时获得的数值。比如说,我会在一个 35×35 的网格上玩,然后那些小方格、单元格中的一个,我们管它叫目标之类的东西,会亮起来,你把光标移到那里,点击它,然后永远重复下去。
第 1827 段
而且你一开始已经能够达到每秒 8 比特。然后你最近又打破了这个纪录。
第 1828 段
对,我现在是8.5。我本来在来奥斯汀的前一天就能超过这个成绩的。但就在快结束时,我遇到了大概,我不知道,大概5秒钟的延迟。我只能等延迟缓下来,然后继续点击。但我当时大概是8.01,接着延迟了5秒,然后我点击的接下来大概3个目标全都停留在8.01。
第 1829 段
所以,如果我在延迟的那段时间里还能点击,我可能就能达到,我不知道,我可能就能达到9。所以我已经到那儿了。我真的很接近了。然后这整个奥斯汀之行确实妨碍了我玩Webgrid的能力。
第 1830 段
这让人很沮丧。
第 1831 段
对,这很——所以你现在满脑子想的就只有这个?
第 1832 段
对,我知道。我想在 9 上做得更好。我想做得更好。我想达到 9,我觉得。嗯,我知道 9 是非常、非常可以实现的。我已经很接近了。我觉得 10 也许下个月就能达到。比如,如果我真的拼一把,可能接下来几周就能做到。
第 1833 段
我觉得你和埃隆基本上是同一个人,因为上次我和他录播客时,他来的时候极其沮丧,因为他以机器人的身份打不过超级莉莉丝。那大概是 1 年前吧,我忘了,好像是单人打。我看得出来,在整个过程中,他大脑里有一部分一直在想:“真希望我现在正在尝试。”
第 1834 段
我想他做到了。
第 1835 段
他那天晚上就做到了。
第 1836 段
对。
第 1837 段
他那天晚上熬夜把它做了出来。这对我来说简直太疯狂了。我的意思是,从根本上说,这真的很鼓舞人心。而你正在做的事也以这种方式鼓舞着人,因为,我的意思是,这不只是关乎游戏。你在那里做的一切都有影响。通过努力在 Webgrid 上取得好成绩,你正在帮助大家弄清楚如何从头到尾创建这个系统,比如解码、软件、硬件、校准,所有这些,如何让这一切运转起来,这样你就能把其他一切都做得非常好。
第 1838 段
是啊,就是非常好玩。
第 1839 段
嗯,那也是,那也是其中一部分,就是让它变得好玩。
第 1840 段
是啊,它让人上瘾。我开玩笑说过,他们进去把这东西植入我大脑时到底做了什么。他们肯定拨动了某个开关,让我更容易沉迷于这类游戏,让我对 Webgrid 之类的东西上瘾。你知道 Bliss 的最高分吗?
第 1841 段
知道,他说好像是 14 之类的。
第 1842 段
17。
第 1843 段
哎呀。
第 1844 段
17.1 之类的,17.01。
第 1845 段
17 点,17.01。
第 1846 段
是啊。
第 1847 段
他告诉我,他好像会趴在地板上玩,还吃花生酱,而且他好像还会禁食。很奇怪。那听起来像作弊。听起来像是在提升表现。
第 1848 段
[Bliss] 不,就像 Noland 第一次玩这个游戏时,他问:“我们玩这个游戏的水平有多高?”我觉得你当时就对我说:“你会试着在这上面打败我。”
第 1849 段
总有一天我会达到那个水平的。
第 1850 段
是啊。我完全相信你。
第 1851 段
我觉得我可以。
第 1852 段
我很期待。
第 1853 段
对。所以首先,我一直在用悬停光标,这极大地妨碍了我玩 Webgrid 的能力。基本上,每次点击我都得等 0.3 秒。
第 1854 段
哦,所以你没法直接点击。所以你是通过悬停来点击的,你说是 0.3 秒?
第 1855 段
0.3秒,这很糟糕。它真的限制了我的速度,呃,就是我能达到多高的水平。我还是做到了大约50,我想我每分钟大概完成了50多次试验,净试验次数,在那种情况下已经很不错了,因为我可以,就是其中一个设置还涉及你需要移动得多慢才能触发一次点击,开始一次点击。所以我大概能判断什么时候达到那个开始触发点击的阈值,然后稍微提前一点开始,这样我点击时就不用在目标上方完全停住。我会在移向目标的途中稍微这么做,试着把时机卡得恰到好处。
第 1856 段
哦,哇。所以你是在减速。
第 1857 段
对,就在到达目标前稍微减一点。(Lex笑)——这简直是顶尖水平的表现,好吧。但还是,0.3秒这个限制很糟糕。
第 1858 段
嗯,我可以降到0.2和0.1。0.1是什——我明白了。
第 1859 段
对,我也稍微试过那个。为了用0.1来玩,我必须调整大量不同的参数,而我这边目前还不能控制所有那些参数。这也会改变模型的训练方式。
第 1860 段
比如,如果我在Webgrid里训练一个模型,就像我基于一个模型进行引导训练,这基本上就是他们在我玩Webgrid时,根据Webgrid的数据训练模型,所以如果我玩10分钟Webgrid,他们就能专门使用那些数据进行训练,从而给我一个更好的模型。如果我分别用0.3和0.1这么做,得到的模型会不一样。它们交互的方式就是非常、非常不同。所以我必须格外小心。
第 1861 段
我发现,用0.3来做在某些方面其实更好,除非我能用0.1来做,并改变所有不同的参数,那样会更理想,因为显然,0.3比0.1更快。所以我能做到。我能做到。
第 1862 段
你能用大脑点击吗?
第 1863 段
目前是用停留光标进行悬停点击。在所有电极丝回缩的问题发生之前,我们当时正在校准点击——左键点击、右键点击。那是我之前的上限。在我后来用停留光标再次打破纪录之前,我想当时是在一个35乘35的网格上使用左键和右键点击。而且使用多种点击时,你能获得更高的BPS,也就是每秒更多比特,因为那样更困难。
第 1864 段
哦,因为那是什么,你应该要么左键点击,要么像右键点击?这种情况你们会用不同的颜色吗?
第 1865 段
对,蓝色目标用于左键点击;橙色目标用于右键点击,他们当时就是这么做的。
第 1866 段
明白了。
第 1867 段
所以我之前用蓝色和橙色目标创下的纪录是 7.5,对,我觉得如果我现在回去做点击校准,并且能够自己发起点击之类的,我想我最多几天就会、会突破 10 这个上限。
第 1868 段
就像,对,你会开始让 Bliss 为他的 17 感到紧张——你应该紧张。
第 1869 段
你觉得我们为什么还没给他那个——对,没错。那么发生回缩时是什么感觉?有些线回缩了?
第 1870 段
糟透了。真的、真的很难受。他们告诉我的那天,正好是我在 Neuralink 的弗里蒙特设施进行大型参观的日子,而且他们就在我们过去之前告诉了我。听到这个真的很难受。我的第一反应是:“好吧,进去,修好它。就是进去,把它取出来,然后修好。”第一次手术非常轻松。就是我睡着了。几个小时后,我醒了,然后就到了现在。
第 1871 段
我没有感觉到任何疼痛,也没有吃任何止痛药之类的。所以我只知道,如果他们愿意,他们第二天就可以进去装一个新的,如果必须这样做的话,因为我希望它变得更好,也不想失去这种能力。我用它玩了几个星期、一个月,玩得非常开心。它为我打开了那么多扇门。它为我开启了更多得多的可能性,我不想在一个月后失去它。
第 1872 段
我当时觉得,如果我得以从这座山的山顶之类的地方看到风景,然后一个月后一切轰然崩塌,那会是命运一次残酷的捉弄。而且我知道,就像,说是山顶,但在我看来,我那时才刚刚开始爬山。我知道还有那么多事情是可能实现的。所以要让这一切都被夺走,真的、真的很难受。
第 1873 段
但后来在开车去设施的路上,我不知道,大概5分钟车程吧,不管具体多久,我和父母聊了聊这件事。我为此祈祷过。我当时就想:“我不会让这件事毁掉我的一天。我不会让这件事毁掉他们为我安排的这次令人惊叹的、怎么说呢,参观。就是,我想去向每个人表明,我有多么感激他们正在做的所有工作。
第 1874 段
我想去,怎么说呢,见见所有促成这一切的人,而且我想去度过我人生中最美好的一天之一。”我确实做到了,那太棒了,而且那绝对是我有幸经历过的最美好的日子之一。然后有几天,我非常沮丧。但在之后最初的几天里,我就是,我不知道它以后是不是还会、还会再次工作。
第 1875 段
我做出了决定,即使我失去了使用 Neuralink 的能力,即使我失去了,即使我错过了未来的一切,只要我还能以任何方式继续为他们提供数据,我就会那样做。如果我需要每天只做一些数据收集,或者每天做身体映射,持续1年,那我就会去做,因为我知道,我所做的一切都会帮助每一个后来者。而那就是我想要的一切。
第 1876 段
我想,我做这件事的全部原因就是帮助人们,而且我知道,只要是我能做的任何有助于此事的事,我都会继续做,即使我再也不能使用光标,我也只是很高兴能参与其中。而我做过的一切都只是额外的好处。那是我得以经历的事情,而且我知道,对每一个后来者来说,它将会有多么令人惊叹。所以不妨就这样继续坚持下去。
第 1877 段
话虽如此,你还是能够一步步努力,把性能恢复回来。所以这就像是从《洛奇1》走到《洛奇2》。那么你最早是在什么时候意识到这是可能的?又是什么给了你某种力量、动力和决心去做到这一点?重新提高成绩并打破你之前的纪录?
第 1878 段
对,是在几周之内。
第 1879 段
再说一次,这感觉就像我在采访一名运动员。(笑)这太棒了。我喜欢感谢我的父母。
第 1880 段
回去的路漫长而艰难,充满了许多困难。曾有过黑暗的日子。我想,那是几个星期。然后就出现了一个转折点。我想他们改变了测量我大脑中神经元尖峰的方式。布利斯,帮我补充一下。
第 1881 段
[布利斯] 对,就是我们测量单个神经元活动的方式。
第 1882 段
对。
第 1883 段
所以我们正从某种单独尖峰检测切换到一种叫作尖峰频带功率的东西,如果你看过之前有我或 DJ 出现的片段,你可能已经了解一些相关内容。
第 1884 段
对,好,所以他们这么做以后,那有点像是头顶亮起了一盏灯,像是灵光一现。就像:“哦,这行得通。”而且这看起来像是我们可以沿着这条路做下去。我立刻就看到了性能的提升。他们切换过去时,我能感觉到。我当时就觉得:“这个更好。这很好。”
第 1885 段
就像此前过去几个星期里的一切,过去大概随便吧,3或4个星期,因为那甚至是在他们告诉我之前,就像在这之前的一切都糟透了。就继续做我们现在正在做的事吧。到了那时,并不是说:“哦,我知道自己仍然只有,比如用 Webgrid 的术语来说,只有4或5 BPS,而之前是7. 5。但我知道,如果我们继续这样做,那么我就能回到那个水平。”
第 1886 段
然后他们给了我驻留光标,而驻留光标一开始糟透了。它显然不是我想要的,但它给了我一条前进的路,让我能够继续使用它,并希望能继续帮上忙。所以我就这样做下去了,再也没有回头。就像我说的,反正我就是那种会随机应变的人。
第 1887 段
这个过程是怎样的?在弄清楚如何以一种真正适用于诺兰德的方式进行尖峰检测时,反馈循环是怎样的?
第 1888 段
对,这是个很好的问题。所以也许先描述一下实际的更新是如何进行的。它基本上是对你的植入物进行了一次更新。所以我们只是通过无线方式对他的植入物进行了软件更新,就像更新你的 Tesla 或 iPhone 一样,而那次固件变更让我们能够记录单个电极附近神经元群体的某种平均值。
第 1889 段
所以,对于具体哪个神经元在做什么,我们得到的分辨率算是比较低,但我们能更全面地了解一根电极附近总体上正在发生什么。而且那种反馈,我是说,基本上,诺兰描述说,我们打开那个开关时,反馈是即时的。我想就在我们这么做的第一天,你一上来就达到了3或4 BPS。那一刻我们恍然大悟:“好,这就是正确的方向。”
第 1890 段
从那之后,我们收到了很多关于如何让它适合独立使用的反馈。所以我们最终关心的是,你能独立使用它来做任何你想做的事。为了达到这一点,我们需要重新设计用户体验,也就是你提到的停留光标,让它成为一种你无需我们一直参与也能独立使用的东西。
第 1891 段
而且,是的,显然,这段旅程仍处于起点。希望我们能回到你可以进行多次点击的阶段,并利用它更流畅、更自然地控制一切,以及你正尝试与之交互的应用程序。
第 1892 段
最重要的是,把那个 Webgrid 数字提高。
第 1893 段
是的。
第 1894 段
对。那么,关于悬停点击,你有时会不小心点到格子吗?要避免误点有多难?
第 1895 段
基本上,我必须让它一直保持移动。就像我说的,有一个阈值,低于它就会触发点击。所以只要我降到那个阈值以下,它就会开始,而我有 0.3 秒的时间在它点到任何东西之前把它移开。如果我根本不想让它进入那种状态,就会让它以一定速度移动,比如一直在屏幕上画圈,来回移动,以防它点到东西。
第 1896 段
其实几周前我注意到,我没在使用植入物时,只是在来回移动手,或者画圈。就好像我在努力阻止光标点击一样,而我只是在准备入睡时这么做,然后我心想:“好吧,这是个问题。”(两人都笑)——为了避免点击。我想,这会不会在你玩游戏时造成问题,比如不小心点到某个东西?
第 1897 段
会,会,下棋时会发生这种事。我因为不小心点到东西,已经输了好几局。
第 1898 段
[布利斯] 我觉得我第一次打败你,就是因为一次意外。
第 1899 段
对,我点错了,对。
第 1900 段
这是个不错的借口,对吧?
第 1901 段
对。
第 1902 段
你随时都可以,每次你输的时候——你可以直接说——那是误操作。
第 1903 段
是的。
第 1904 段
你说这款应用相比第1版改进了很多。你最初开始使用它时,它非常不一样。所以你能谈谈你和团队经历的反复试验吗?比如200多页的笔记。那种过程是怎样的——是的。
第 1905 段
来回沟通、共同努力改进这个东西的过程?
第 1906 段
很大一部分就是我日复一日地使用它,然后说,比如:“嘿,你们能帮我做这个吗?比如给我这个。我想要能够做到那个。我需要这个。”我觉得,其中很多事情可能要等到有人真正开始使用这款应用、使用植入物时,他们才会想到。这些就是他们以前绝不会想到的事情。或者有些甚至非常针对我个人,可能是我想要的东西。
第 1907 段
对于接下来参与的人,我有点担心的一件事是,他们想要的东西可能会和我设置的方式,或者我给团队的建议非常不同。他们会看到团队为我添加的一些东西,然后说:这主意真蠢。他为什么会提出这种要求?
第 1908 段
所以我真的很期待让接下来的人参与进来,因为我敢保证,他们会想到一些我从未想过的东西。他们会想到各种改进。我会想:“哇,这个主意真不错。真希望我当时能想到。”然后他们也会对我提出一些反对意见,比如:“是啊,你在这里要求他们做的事情是个坏主意。我们换这种方式做吧。”
第 1909 段
我非常乐意看到这种情况发生。但这里面就是有很多与不同游戏或应用、互联网,以及总体上与电脑进行的各种不同交互。各种各样的漏洞最后会从四面八方冒出来。所以我就是尽量多使用它,向他们展示哪些有效、哪些无效,以及我希望哪些地方做得更好。
第 1910 段
然后他们会采纳这些反馈,而且通常会为我创造出令人惊叹的东西。他们解决这些问题的方式是我绝对想象不到的。他们做的每件事都非常出色。所以我真的非常感激,我能够给他们反馈,而他们能把反馈变成实际成果,因为我的很多反馈其实真的很笨。基本上就是:“我想要这个。请想办法解决。”
第 1911 段
而且我们会回来,而且经过了极其周密的思考,它比任何我自己能够想到或实现的东西都要好得多。所以他们就是很棒。他们真的、真的很酷。
第 1912 段
随着 BCI 社群不断壮大,你愿意和其他使用 Neuralink 的人相处吗?如果愿意,你想和他们建立什么样的关系?因为你说过,他们对于如何使用这个东西可能会有一套不同的想法。
第 1913 段
是的。
第 1914 段
你会因为他们在 Webgrid 上的表现而感到压力吗?
第 1915 段
不,不,我希望竞争。我希望从第一天起,他们就把我打得落花流水。我希望他们超过它,并彻底击败它。如果可以,就把它翻倍。因为一方面,这只会推动我变得更好,因为我非常好胜。我希望其他人推动我。我认为,对于任何试图取得伟大成就的人来说,这一点都很重要:他们身边需要有其他人推动他们变得更好。我甚至有一次在 X 上拿这件事开过玩笑。
第 1916 段
比如,等下一批人选出来以后,比如 qubadi cot music,我只是很兴奋能有其他人和我一起做这件事,并和他们分享经历。他们想和我交流多少,我都非常乐意;我也非常乐意给他们建议。我不知道自己能给他们什么样的建议,但如果他们有问题,我非常乐意回答。
第 1917 段
对于下一位临床试验参与者,你会给出什么建议?
第 1918 段
他们应该享受这件事,因为它真的非常有趣。而且我希望他们非常、非常努力,因为这不仅是为了我们,也是为了所有后来的人。如果他们需要任何东西,就来找我。需要任何东西,也可以去找 Neuralink。天啊,Neuralink 能移山。他们真的会竭尽所能为我做任何事情。这是一个非常了不起的支持体系。
第 1919 段
这让我对很多事情都安心了,比如那些我一直有疑问的事情,或者我想做的很多事情。而且他们一直都在,这真的非常、非常好。我会告诉他们,不要害怕带着他们的任何问题、任何顾虑,以及他们希望用这个做的任何事情去找 Neuralink。只要是 Neuralink 有能力提供的任何帮助,我知道他们都会提供。而且我不知道,我不知道。
第 1920 段
就拼命努力吧,因为我们努力为此付出全部真的很重要。
第 1921 段
所以要开心,也要努力。
第 1922 段
对,对,就是这样。也许以后我就对人们这么说:开心点,努力干。
第 1923 段
现在,你真是个职业运动员。说简短点。(诺兰笑)也许可以谈谈,现在有了 Neuralink 植入物之后,你都能做些什么。比如说,以这种方式与外界互动给你带来的自由。比如你会通宵玩电子游戏。而且你是独自做到的。那是一种自由。你能谈谈你获得的这种自由吗?
第 1924 段
对,这就是所有,我不知道,处在我这种境况的人想要的。他们只是想更加独立。我能为身边的人减轻的负担越多,就越好。如果我能够在不依赖家人、不通过任何朋友的情况下与世界互动,比如不需要他们帮我做事,那就越好。
第 1925 段
如果我能整晚坐在电脑前,而不需要有人把我扶起来,比如把我的 iPad 放到一个我能使用的位置,然后还得让他们整晚不睡等着我,直到我准备好不用了,这就减轻了我们所有人的负担。而这真的就是我所能要求的一切。这是我无论如何感谢 Neuralink 都不为过的事情。而且我知道我的家人也有同样的感受。
第 1926 段
仅仅是能够拥有在白天或夜晚的任何时间独自做事的自由,对我来说就意味着整个世界。而且我不知道。
第 1927 段
当你凌晨 2:00 独自玩 Webgrid 时,我会想象周围一片漆黑,只有一盏灯亮着。而你全神贯注。你脑子里在想什么?(诺兰笑)还是说你当时处于一种心流状态,脑子是空的,就像那些禅宗大师一样?
第 1928 段
是的,一般来说,就是我在播放某种音乐。我有一个非常庞大的播放列表,所以我就跟着音乐尽情摇摆。然后这也像是一场与时间的赛跑,因为我一直在查看我的植入物还剩多少电量百分比。比如,好吧,我还有30%,这相当于x的时间,也就是说我必须在接下来的1个半小时内打破这项纪录,否则今晚就没戏了。
第 1929 段
所以发生这种情况时会有点紧张。电量高于50%时,我会想:“好吧,我还有时间。”然后电量开始降到30%,接着是20%。就像,好吧,10%,这里马上会弹出一个小弹窗,它会严重打乱我玩Webgrid的节奏。它会告诉我电量不足,低电量弹窗会出现,然后我就会想,它真的会把我坑惨。
第 1930 段
所以如果我必须,如果我要打破这个纪录,我就必须在接下来大概30秒内做到,否则那个弹窗就会挡住,像是盖住我的 Webgrid。之后,我就去点它,再回到 Webgrid。然后我会想:“好吧,这意味着在这东西没电之前我还剩10分钟。”通常我脑子里就是这些,还有当时正在播放的歌。我太想打破那些纪录了。
第 1931 段
就像我玩 Webgrid 时只想要这个。它已经不太像是:“哦,这只是一项休闲活动。”就像我只是喜欢做这个,因为感觉太舒服了,而且能让我放松。不,一旦我进入 Webgrid,你最好打破这个纪录,否则你现在就要浪费掉大概5个小时的人生。我不知道,就是很好玩。很好玩,伙计。
第 1932 段
你有没有试过用比如2个目标和3个目标玩 Webgrid?那样能获得更高的 BPS 吗?
第 1933 段
你能那么做吗?
第 1934 段
[布利斯] 你的意思是,比如不同颜色的目标,还是说你是指——哦,多个目标,因为那会改变这个东西。
第 1935 段
对,所以 BPS 是目标数量的对数乘以(正确次数减去错误次数),再除以时间。所以你可以把不同的点击理解为基本上让活跃目标的数量翻倍。
第 1936 段
明白了。
第 1937 段
所以你知道,基本上选项越多,任务越难,你得到的 BPS 就越高。而且还有你以前玩过的那种禅模式,它就像是——对。对,它用一个网格铺满整个屏幕。然后我不知道。
第 1938 段
[莱克斯] 对,然后你就会说:“这也太疯狂了。”
第 1939 段
对。
第 1940 段
[布利斯] 他不喜欢它,因为它不显示 BPS。
第 1941 段
我让他们在背景里放了一个巨大的 BPS,所以现在就像是禅模式的反面。就像是超级困难模式,就像纯金属模式。就像后面的计数器上有一个巨大的数字。
第 1942 段
[布利斯] 我们应该给它改名。现在“金属模式”这个名字好得多。
第 1943 段
所以你也玩《文明 VI》?
第 1944 段
我很喜欢《文明 VI》,对。
第 1945 段
[莱克斯] 你说过,你通常会选韩国?
第 1946 段
是的,对。韩国很棒的一点是,他们专注于那种科技胜利,而这并不是计划好的。比如我玩韩国已经很多年了,然后 Neuralink 的那些事都发生了。所以这算是契合上了。但我注意到,走科技胜利路线时,只要你能猛冲科技、猛冲科学,那你就什么都能做。
第 1947 段
比如到了游戏里的某个阶段,你在技术上会遥遥领先于所有人,你会有火枪兵、步兵,有时还有飞机,而别人可能还在用弓箭作战。所以,如果你想赢得征服胜利,只要让科学发展到某个阶段,然后去消灭世界上的其他国家就行了。
第 1948 段
或者你也可以一路发展科学,以那种方式获胜,而且你会远远领先于所有人,因为你产出的科学值实在太多了,根本没有可比性。我曾经只专注于科学,却意外地以不同方式赢得了胜利。
第 1949 段
因为专注于科学而意外获胜。
第 1950 段
很明显,我当时只玩科学。就是一路发展科学,只搞科技。我想把科技树里的每一项科技之类的都解锁。然后我却意外地通过外交胜利赢了,我气坏了。(Lex 笑)我气坏了,因为游戏就那么结束了。某个回合,它就像是:“哦,你赢了,你太会外交了。”我就想:“我不想这么干。我本来应该向更多人宣战之类的。”
第 1951 段
太糟糕了,但走科技路线不需要那种庞大的文明,尤其是用韩国时。你可以让规模保持得相当小。所以我通常就是发展到某种军事单位,然后把他们全都部署在边境周围,把所有人挡在外面,之后我就只管发展。所以非常孤立主义。
第 1952 段
不错。
第 1953 段
是啊。
第 1954 段
就专注于科学和技术。
第 1955 段
[诺兰德] 对,就是这样。
第 1956 段
你说得这事好像特别有趣。
第 1957 段
真的特别有趣。
第 1958 段
我还看了《文明 VII》的预告片。
第 1959 段
天啊,我太兴奋了。
第 1960 段
是啊。而且那可能会在————拜托,《文明 VII》,联系我吧。Alpha 测试、Beta 测试,什么都行。
第 1961 段
等等,什么时候发布?
第 1962 段
2025年。
第 1963 段
对,对,明年,对。关于 Neuralink 应用和整个体验,你还希望看到哪些方面得到改进?
第 1964 段
我希望能像我说的那样,重新实现那种按需点击,也就是常规点击。那会很棒。我希望能连接更多设备。现在还只能连接电脑。我希望能在手机上使用它,或者在不同的游戏主机、不同的平台上使用它。老实说,我希望能够控制尽可能多的东西。比如能控制一个 Optimus 机器人就会很酷。
第 1965 段
如果我能控制一个 Optimus 机器人,那就太酷了。Link 应用本身,似乎我们正在逐渐明确它未来可能会是什么样子。至少我想从中获得的大部分东西,似乎都已经解决了。我要说的另一件事就是,希望能更全面地控制所有那些可以调整的参数,比如与光标相关的参数之类的。光标以特定方式移动,背后涉及很多因素。
第 1966 段
而且我有,我不知道,大概3个或4个那样的参数,而且可能—— ——比如增益和摩擦力之类的?
第 1967 段
增益和摩擦力,对。而且仅仅速度方面可能就有两倍数量的参数,然后还有实际的悬停光标。所以我想要全部。我希望尽可能多地控制我的环境,尤其是—— ——所以你想要类似高级模式的东西?通常会有菜单,有基础模式。而你就像那种人—— ——我会选—— ——高级用户的高级模式。
第 1968 段
对,对。
第 1969 段
明白了。
第 1970 段
那就是我想要的。我希望尽可能多地控制这个。所以,是的,这其实就是我所能要求的一切。把所有东西都给我就行。
第 1971 段
语音功能有用吗?比如除了其他所有功能之外,也能直接说话?
第 1972 段
对,你是指比如我在使用它的时候吗?
第 1973 段
你在使用它的时候,比如语音转文字?
第 1974 段
哦,对。
第 1975 段
还是说你会打字,或者比如,因为也有一个键盘?那个真的很好用。
第 1976 段
对,对。所以有一个虚拟键盘。我还想进一步研究的另一件事,就是找到某种不同的打字或输入文字的方式。目前基本上是听写,还有一个我可以用光标操作的虚拟键盘。但我们尝试过像手指拼写,也就是手语中的手指拼写。这个看起来非常有希望。
第 1977 段
所以我脑子里有这样一个想法:它会有一条和我使用光标时非常相似的学习曲线,当时我在某个时刻从尝试做出动作转变为想象动作。我有一种感觉,这只是我的直觉,在某个时候,我会做手指拼写,而我将不再需要实际尝试做出手指拼写,我只需要想着我想要的那个,比如说,字母,它就会弹出来。
第 1978 段
那会非常精彩。
第 1979 段
是的。
第 1980 段
那很有挑战性,很难。你要实现那种飞跃,得做大量工作。但那会非常棒。
第 1981 段
然后,从字母到单词又是一步。比如你会从目前这种只是用手指拼写手语字母表,发展到下一步。但如果它能识别这些,那它应该也能识别整个手语,像一门语言那样。
第 1982 段
所以,如果我能做一些类似的事情,或者只是用手语拼出单词,如果我能以合理的速度拼写,而它能识别出来,那么我应该只要在脑中想一遍,它就会做同样的事情。我想不出为什么不行。看过光标控制后,我想不出它为什么会行不通,但我们还得多试验一下。
第 1983 段
就训练自己从尝试运动转变为想象运动而言,这个过程是怎样的?花了多长时间?那么,这种过程要花多长时间?
第 1984 段
嗯,过了几周,它才像是偶然发生在我身上。但现在我知道这是可能的,我觉得我也能让其他事情以这种方式发生。我觉得那会简单得多得多。
第 1985 段
你会植入升级版设备吗?
第 1986 段
当然,绝对会。只要他们允许我。
第 1987 段
所以你对自己的手术经历没有任何顾虑?整个过程都没有什么后悔的?
第 1988 段
没有。
第 1989 段
所以到目前为止一切都很好?
第 1990 段
对。
第 1991 段
你就不断升级。
第 1992 段
是啊,我是说,为什么不呢?我已经看到它给我的生活带来了多大的影响。而且我知道,从现在开始,一切他妈的都会越来越好。所以我很愿意。我很愿意升级。
第 1993 段
除了这种心灵感应之外,还有哪些未来功能让你感到兴奋?视觉功能有意思吗?比如,对于盲人来说,就是让他们能够看见;或者用于语言表达。
第 1994 段
是的,这里面有很多非常非常酷的东西。我是说,我们谈论的是大脑,所以,比如说,这些只是运动皮层方面的东西。还有太多事情可以做。视觉功能让我觉得很迷人。我认为那会非常非常酷。让一个人平生第一次拥有看见的能力,那将会是,我是说,这甚至可能比帮助像我这样的人更令人惊叹。光是想想就觉得不可思议。
第 1995 段
语言表达方面真的很有意思,能够实现某种类似实时翻译的功能,消除语言障碍,会非常酷。它若能解决任何实际障碍,比如语言表达障碍,都会非常非常酷。另外,还有很多不同的残障都源于大脑。而你将能够,希望能够解决其中很多问题。
第 1996 段
我知道已经有一些可以植入大脑、帮助癫痫患者的东西。我想,这会起到同样的作用。所以你可以做类似的事情。我知道,就连乔·罗根这样的人也谈过以不同方式刺激大脑的可能性。我不确定。我不确定其中很多做法在伦理上是否合适。老实说,这超出了我的判断能力。
第 1997 段
但我知道,说到大脑,以及能够进入其中并实际做出改变来帮助人们或改善他们的生活,有很多事情可以做。所以我非常期待这一切带来的成果。而且我认为这并没有那么遥远。我认为其中很多都能在我有生之年得以实施,前提是我能长寿。
第 1998 段
你刚才指的是,比如患有抑郁症或类似病症的人可能得到帮助。
第 1999 段
对,就像那样拨一下开关,让人开心起来。我知道,我想乔更多是从这种角度谈的,比如你想体验一下药物致幻是什么感觉。比如你想体验一下服用——会是什么感觉。——当然。
第 2000 段
对,蘑菇之类的东西,DMT。就像你可以直接拨动大脑里的那个开关。我的朋友贝恩谈过,可以把你记忆中的某些部分抹掉,然后像第一次那样重新体验一些东西,比如你最喜欢的电影或最喜欢的书。就快速把那段记忆抹掉,然后重新爱上《哈利·波特》之类的。我跟他说,我当时就说:“我不知道该怎么看待人们能直接抹掉自己记忆中的某些部分。这在我看来有点可疑。”他说:“他们已经在这么做了。”
第 2001 段
听起来挺靠谱。对,我很想要记忆回放。就是对旧日记忆进行真正的高分辨率回放。
第 2002 段
对,我以前看过一集讲这个的《黑镜》。我觉得我不想要。
第 2003 段
对,所以《黑镜》总是会考虑最坏的情况,这很重要。我觉得人们没有充分考虑最好的情况或一般情况。我不知道我们人类是怎么回事。我们总想思考可能发生的最糟糕的事。我们喜欢戏剧性。
第 2004 段
[诺兰] 对。(笑)——就像,这项新技术会怎么杀死所有人?我们就是喜欢这个。又会说,对,来看看吧。
第 2005 段
希望人们别在我身上过多考虑这个问题。那会毁掉我的很多计划。
第 2006 段
对,对。我猜你将不得不接管世界。我的意思是,我很喜欢你在 Twitter 上发的内容。你发推说:“自从装了 Neuralink 之后,我很想拿听见脑中有声音这件事开玩笑,但我觉得人们会误解。再说了,我脑中的声音叫我别这么做。”
第 2007 段
对。
第 2008 段
请永远别停。所以你刚才谈到 Optimus。你会很想做到这一点吗,控制机械臂,或者控制整个 Optimus?
第 2009 段
哦,对,当然。当然,绝对想。
第 2010 段
你觉得,能够与世界进行物理互动这件事,从根本上来说有什么不同吗?
第 2011 段
对,哦,100%。我还知道另一件事,比如能够赋予人们感受感觉之类的能力,通过进入大脑,让 Neuralink 或许做到这一点。这也可能是某种能够通过 Optimus 转译、传递的东西。这两者之间有各种非常酷的相互作用。然后还有,就像你说的,进行身体上的互动。
第 2012 段
我是说,我自己做不了的事情中,有 99% 显然都需要,我需要一名照护者,需要有人替我实际去做。如果 Optimus 机器人能做到这些,比如我就能过上极其独立的生活,不再给身边的人造成这么大的负担,而且这会改变像我这样的人生活的方式,至少在不管这是什么的东西被治愈之前。但能够以身体与这个世界互动,比如,那简直太棒了。
第 2013 段
而且它们不只是,比如用来当,必须当照护者什么的,还可以做我谈到的那种事,就是能读一本书。想象一下,一个 Optimus 机器人只要能在我面前把一本书撑开,比如再次闻到那种气味。到那时我也许还是感觉不到它。又或者通过感觉之类的东西,我又可以了。
第 2014 段
但阅读实体书和盯着屏幕或听有声书之间,有某种不同。我其实不喜欢有声书。到现在我已经听了很多本,但我并不是真的喜欢它们。我更愿意,比如阅读实体书。
第 2015 段
所以,你很想能够体验的一件事,就是打开书,把它拿到你面前。还有感受纸张的触感。
第 2016 段
对。哦天啊,那种触感,那种气味。我是说,就是书页上的文字有某种特别之处,而且他们也在 Kindle 之类的设备上复刻了那种纸张的颜色。对,就是不一样,对。所以哪怕只是这么简单的一件事。
第 2017 段
所以,你怀念的东西之一就是触觉。
第 2018 段
是的,对。
第 2019 段
我在这个世界上接触的很多东西,比如衣服,或者真的任何我在这个世界上接触的实体物品,很多时候,我身边的人会直接过来,比如,把它蹭在我脸上。他们会把某样东西放在我身上,让我能感受到重量。他们会拿一件衬衫在我身上摩擦,让我能感受到布料。触觉确实有某种非常深刻的意义,这是我非常怀念的东西,也是我很想再次拥有的东西,不过走着瞧吧。
第 2020 段
如果有一只能够触摸的手,你会做的第一件事是什么?之后会给你妈妈一个拥抱,对吧?
第 2021 段
是啊,我知道。自从我出事以来,我基本上每天都向上帝祈求一件事,就是希望有一天能够动起来,哪怕只能动动我的手。这样的话,比如我就能握一下我妈妈的手之类的,只是想向她表明,比如,我有多么在乎她、多么爱她等等。类似这样的事。能够和我身边的人互动,握手,拥抱某个人。我不知道,任何类似的事。能够帮我自己吃东西,比如,我可能会变得很胖,那会是一件非常、非常糟糕的事。
第 2022 段
还会在实体棋盘上击败布利斯。
第 2023 段
是啊,是啊。我的意思是,好处实在太多了。(笑)还有就是想方设法让我觉得自己能把布利斯拉低到我的水平。
第 2024 段
是啊。
第 2025 段
因为——对。
第 2026 段
他就是这么了不起的一个人,他的一切都远远超出常人,所以只要能做点什么把他拉低一点,我都非常乐意。
第 2027 段
是啊,让他谦逊一点,他需要这样。
第 2028 段
[诺兰] 是啊。(笑)——好吧。他就坐在我旁边。你有没有想明白过,为什么上帝会让好人经历如此艰难的磨难?
第 2029 段
哦,天啊。我觉得这一切都是为了让我们明白自己有多么需要上帝。而且我认为,没有黑暗,就不会有光明。我觉得,如果我们所有人一直都很快乐,那就永远没有理由转向上帝。我觉得那样的话,善与恶的概念也就不存在了。而且我认为,正因为世界上有这么多黑暗和邪恶,它才让我们所有人更加珍惜善,以及我们所拥有的一切。
第 2030 段
而且我觉得,就像我发生事故的时候,我最早对我一个最好的朋友说的话之一是——这是在我发生事故后的头一两个月内——我说:“这场事故的一切只是让我明白并相信,上帝是真实存在的,基本上,确实有一位上帝。而且,我和他的互动都是真实且值得的。”
第 2031 段
他说,恰恰相反,看到我经历这场事故,他相信上帝并不存在。这是一种非常不同的反应。但我相信,这是上帝考验我们、塑造我们的品格、让我们经历艰难困苦的一种方式,以确保我们明白他有多么珍贵,也明白他赐予我们的事物和时间有多么珍贵。然后,希望我们能从这一切中成长。
第 2032 段
我认为,身处于此很大一部分意义就在于,不只是过轻松的生活、做一切容易的事,而是走出我们的舒适区,真正挑战自己,因为我认为我们就是这样成长的。
第 2033 段
对于我们正在经历的这一切,也就是人类文明,是什么给了你希望?
第 2034 段
哦,天哪。我认为人是我最大的灵感来源。即使只是在 Neuralink 待了几个月,看着人们的眼睛,听他们讲述自己为什么做这件事的动机,也非常鼓舞人心。而且我知道,他们本可以去别的地方,做更安逸的工作,在其他地方任职,做一些其实没那么大意义的甲、乙或丙。
第 2035 段
但他们却选择留在这里,他们想让人类变得更好,也想让身边的人过得更好,就是那些他们一生中接触过的人;他们想让自己可能身有残障的家人过上更好的生活,或者他们看着像我这样的人,说:“我能为此做点什么,所以我就去做。”而在这个世界上,我一直以来最能产生共鸣的,始终是人。
第 2036 段
我一直是个善于与人打交道的人,我喜欢了解人,也喜欢了解人们是如何成长起来的,以及他们来自哪里。看到人们在并非必须的情况下,愿意为像我这样的人付出这么多,而且还特意想方设法让我的生活变得更好。
第 2037 段
这让我对整个人类抱有很大的希望:我们有多么关心彼此,以及当我们大家团结起来、努力带来改变时,我们能做到多少事情。我知道这个世界上有很多坏事,但一直如此,也将永远如此。
第 2038 段
而且我认为,那体现了人类的韧性,展现了我们所能承受的一切,也展现了我们多么渴望陪在彼此身边、互相帮助,以及我们从中获得了多大的满足感,因为我认为,我们来到这里的原因之一就是互相帮助。我不知道,这总能给我希望。就是意识到外面仍然有人在乎,仍然愿意伸出援手。
第 2039 段
也谢谢你成为这样的人,并且在经历了这一切之后依然继续做一个了不起的人,成为许多人的榜样,也因为许多原因成为我的榜样,包括你在 Webgrid 上堪称史诗、好得令人难以置信的表现。今晚我会通宵训练,争取赶上你。
第 2040 段
你能做到。
第 2041 段
而且我相信你,你回来之后一定能做到——抱歉打断一下,奥斯汀之行结束后,等你回来,最终能击败布利斯。
第 2042 段
对,对,当然。绝对能。
第 2043 段
我支持你。全世界都在支持你。
第 2044 段
谢谢。
第 2045 段
老兄,谢谢你所做的一切。
第 2046 段
谢谢,谢谢你,老兄。
第 2047 段
感谢收听这场与诺兰·阿博的对话,以及此前与埃隆·马斯克、DJ·徐、马修·麦克杜格尔和布利斯·查普曼的对话。若要支持本播客,请查看简介中的赞助商。现在,在结束之际,让我为你留下奥尔德斯·赫胥黎在《知觉之门》中的几句话:“我们共同生活。我们彼此影响、彼此回应,但始终,无论在何种情况下,我们都是独自一人。
第 2048 段
殉道者们手牵着手走进竞技场。他们却独自被钉上十字架。恋人们相互拥抱,拼命试图将各自隔绝的狂喜融合成一次共同的自我超越;但徒劳无功。就其本质而言,每一个具身的灵魂都注定要在孤独中承受痛苦与享受欢愉。感觉、情感、洞见、幻想,这一切都是私密的,除了借助符号间接传达之外,无法交流。
第 2049 段
我们可以汇集有关体验的信息,却永远无法汇集体验本身。从家庭到国家,每一个人类群体都是由一个个孤岛宇宙组成的社会。”感谢收听,希望下次再见。
Paragraph 1
The following is a conversation with Elon Musk, DJ Seo, Matthew MacDougall, Bliss Chapman, and Nolan Arbaugh about Neuralink and the future of humanity. Elon, DJ, Matthew and Bliss are of course part of the amazing Neuralink team, and Noland is the first human to have a Neuralink device implanted in his brain. I speak with each of them individually, so use timestamps to jump around, or as I recommend, go hardcore and listen to the whole thing.
Paragraph 2
This is the longest podcast I've ever done. It's a fascinating, super technical, and wide-ranging conversation, and I loved every minute of it. And now, dear friends, here's Elon Musk, his fifth time on this, "The Lex Fridman Podcast."
Paragraph 3
Drinking coffee or water?
Paragraph 4
Water. I'm so over-caffeinated right now. Do you want some caffeine?
Paragraph 5
I mean, sure.
Paragraph 6
There's a Nitro drink.
Paragraph 7
This supposed to keep you up till like tomorrow afternoon basically. (laughs) - Yeah. I don't have any- - So what is Nitro? It's just got a lot of caffeine or something?
Paragraph 8
Don't ask questions. It's called Nitro.
Paragraph 9
Do you need to know anything else?
Paragraph 10
It's got nitrogen, that's ridiculous. I mean, what we breathe is 78% nitrogen anyway. What do you need to add more for? (laughs) - [Speaker] Unfortunately, you're gonna need it.
Paragraph 11
Most people think that they're breathing oxygen, and they're actually breathing 78% nitrogen. You need like a milk bar.
Paragraph 12
Milk bar. (Elon laughing) - Like from Clockwork Orange. (laughs) - Yeah, yeah. Is that top three Kubrick film for you?
Paragraph 13
Clockwork Orange, it's pretty good. I mean, it's demented. Jarring, I'd say.
Paragraph 14
(laughs) Okay. Okay, so first let's step back and big congrats on getting Neuralink implanted into a human. That's a historic step for Neuralink.
Paragraph 15
Oh, thanks, yeah.
Paragraph 16
There's many more to come.
Paragraph 17
Yeah, and we just, obviously, our second implant as well.
Paragraph 18
[Lex] How did that go?
Paragraph 19
So far, so good. Looks like we've got, I think over 400 electrodes that are providing signals. So yeah.
Paragraph 20
Nice. How quickly do you think the number of human participants will scale?
Paragraph 21
It depends on the regulatory approval, the rate which we get regulatory approvals. So we're hoping to do 10 by the end of this year. Total of 10, so eight more.
Paragraph 22
And with each one, you're gonna be learning a lot of lessons about the new biology, the brain, everything, the whole chain of the Neuralink, the decoding, the signal processing, all that kind of stuff.
Paragraph 23
Yeah, yeah, I think it's obviously gonna get better with each one. I mean, I don't wanna jinx it, but it seems to have gone extremely well with the second implant, so there's a lot of signal, a lot of electrodes. It's working very well.
Paragraph 24
What improvements do you think we'll see in Neuralink in the coming, let's say, let's get crazy, the coming years?
Paragraph 25
I mean, in years, it's gonna be gigantic, because we'll increase the number of electrodes dramatically. We'll improve the signal processing. Even with only roughly, I don't know, 10, 15% of the electrodes working with Noland, with our first patient, we were able to get to achieve a bit per second. That's twice the world record. So I think we'll start like vastly exceeding world record by orders of magnitude in the years to come.
Paragraph 26
So it's start getting to, I don't know, a hundred bits per second thousand. Maybe if like five years from now, we might be at a megabit, like faster than any human could possibly communicate by typing or speaking.
Paragraph 27
Yeah, that BPS is an interesting metric to measure. There might be a big leap in the experience once you reach a certain level of BPS.
Paragraph 28
[Elon] Yeah.
Paragraph 29
Like entire new ways of interacting with a computer might be unlocked.
Paragraph 30
And with humans.
Paragraph 31
With other humans.
Paragraph 32
Provided they have (laughs), they want a Neuralink too.
Paragraph 33
Right.
Paragraph 34
Otherwise, they won't be able to absorb the signals fast enough.
Paragraph 35
Do you think they'll improve the quality of intellectual discourse?
Paragraph 36
Well, I think you could think of it, if you were to slow down communication, how do you feel about that? If you'd only talk at, let's say, 1/10th of normal speed, you'd be like, "Wow, that's agonizingly slow."
Paragraph 37
[Lex] Yeah.
Paragraph 38
So now, imagine you could speak, communicate clearly at 10 or 100 or 1,000 times faster than normal.
Paragraph 39
Listen, I'm pretty sure nobody in their right mind listens to me at 1x, they listen at 2x. (Elon laughs) I can only imagine what 10x would feel like or could actually understand it.
Paragraph 40
I usually default to 1.5x. You can do 2x, but well, actually, if I'm listening to somebody in like sort of 15, 20 minutes segments to go to sleep, then I'll do it 1.5x. If I'm paying attention, I'll do 2x. (laughs) - Right.
Paragraph 41
But actually, if you start actually listen to podcasts or sort of audio books or anything, if you get used to doing it at 1.5, then one sounds painfully slow.
Paragraph 42
I'm still holding onto one because I'm afraid. I'm afraid of myself becoming bored with the reality, with the real world where everyone's speaking on 1x. (both laughing) - Well, depends on the person. You can speak very fast. Like we can communicate very quickly. And also, if you use a wide range of, if your vocabulary is larger, your bit rate, effective bit rate is higher.
Paragraph 43
That's a good way to put it.
Paragraph 44
Yeah.
Paragraph 45
The effective bit rate. I mean, that is the question is how much information is actually compressed in the low bit transfer of language.
Paragraph 46
Yeah. If there's a single word that is able to convey something that would normally require, I don't know, 10 simple words, then you've got maybe a 10x compression on your hands. And that's really, like with memes, memes are like data compression. It conveys a whole, you're simultaneously hit with a wide range of symbols that you can interpret. And you kinda get it faster than if it were words or a simple picture.
Paragraph 47
And of course, you're referring to memes broadly like ideas.
Paragraph 48
Yeah. There's an entire idea structure that is like an idea template, and then you can add something to that idea template. But somebody has that preexisting idea template in their head. So when you add that incremental bit of information, you're conveying much more than a few, just set a few words. It's everything associated with that meme.
Paragraph 49
You think there'll be emergent leaps of capability as you scale the number of electrodes? Like there'll be a certain, you think there'll be like actual number where it just, the human experience will be altered?
Paragraph 50
Yes.
Paragraph 51
What do you think that number might be, whether electrodes or BPS? We of course don't know for sure, but is this 10,000, 100,000?
Paragraph 52
Yeah, I mean certainly, if you're anywhere at 10,000 bits per second, I mean, that's vastly faster than any human could communicate right now. If you think about what is the average bits per second of a human? It is less than one bit per second over the course of a day, because there are 86,400 seconds in a day. And you don't communicate 86,400 tokens in a day. Therefore, your bits per second is less than one, averaged over 24 hours.
Paragraph 53
It's quite slow. And now, even if you're communicating very quickly, and you're talking to somebody who understands what you're saying, because in order to communicate, you have to at least, to some degree, model the mind state of the person to whom you're speaking.
Paragraph 54
Then take the concept you're trying to convey, compress that into a small number of syllables, speak them, and hope that the other person decompresses them into a conceptual structure that is as close to what you have in your mind as possible.
Paragraph 55
Yeah, I mean, there's a lot of signal loss there in that process.
Paragraph 56
Yeah, very lousy compression and decompression. And a lot of what your neurons are doing is distilling the concepts down to a small number of symbols of, say, syllables that I'm speaking, or keystrokes, whatever the case may be. So that's a lot of what your brain computation is doing.
Paragraph 57
Now, there is an argument that that's actually a healthy thing to do or a helpful thing to do because as you try to compress complex concepts, you're perhaps forced to distill what is most essential in those concepts as opposed to just all the fluff. So in the process of compression, you distill things down to what matters the most, because you can only say a few things. So that is perhaps helpful.
Paragraph 58
I think we might, we'll probably get, if our data rate increases, it's highly probable that we'll become far more verbose. Just like your computer, when computers had like, my first computer had 8K of RAM, so you really thought about every byte. And now you've got computers with many gigabytes of RAM. So if you wanna do an iPhone app that just says 'Hello world,' it's probably, I don't know, several megabytes minimum. (laughs) A bunch of fluff.
Paragraph 59
But nonetheless, we still prefer to have the computer with more memory and more compute. So the long-term aspiration of Neuralink is to improve the AI human symbiosis by increasing the bandwidth of the communication, because even in the most benign scenario of AI, you have to consider that the AI is simply gonna get bored waiting for you to spit out a few words.
Paragraph 60
I mean, if the AI can communicate it to terabits per second and you're communicating it bits per second, it's like 203.
Paragraph 61
Well, it is a very interesting question for a super intelligent species. What use are humans?
Paragraph 62
I think there is some argument for humans as a source of will.
Paragraph 63
Will?
Paragraph 64
Will, yeah. Source of will or purpose. So if you consider the human mind as being essentially, there's the primitive limbic elements, which basically even like reptiles have, and there's the cortex, that's the thinking and planning part of the brain. Now, the cortex is much smarter than the limbic system, and yet is largely in service to the limbic system. It's trying to make the limbic system happy.
Paragraph 65
I mean, the sheer amount of compute that's gone into people trying to get laid is insane, without actually seeking procreation. They're just literally trying to do this sort of simple motion. (laughs) And they get a kick out of it. So this simple, which in the abstract rather absurd motion, which is sex, the cortex is putting a massive amount of compute into trying to figure out how to do that.
Paragraph 66
So like 90% of distributed compute of the human species is spent on trying to get laid, probably, like a massive amount.
Paragraph 67
Large percent, yeah, yeah. There's no purpose to most sex except hedonistic. It's just sort of joy or whatever. Dopamine release. Now, once in a while, it's procreation, but for humans, modern humans, it's mostly recreational. So your cortex, much smarter than your limbic system, is trying to make the limbic system happy 'cause the limbic system wants to have sex, or want some tasty food or whatever the case may be.
Paragraph 68
And then that is then further augmented by the tertiary system, which is your phone, your laptop, iPad, whatever, or your computing stuff. That's your tertiary layer. So you're actually already a cyborg. You have this tertiary compute layer, which is in the form of your computer with all the applications or your compute devices.
Paragraph 69
And so in the getting laid front, there's actually a massive amount of digital compute also trying to get laid, with like Tinder and whatever.
Paragraph 70
Yeah. So the compute that we've humans have built is also participating. (laughs) - Yeah, I mean, there's like gigawatts of compute going into getting laid, of digital compute.
Paragraph 71
Yeah. (laughs) What if AGI will- - This is happening as we speak.
Paragraph 72
If we merge with AI, it's just gonna expand the compute that we humans use- - Pretty much.
Paragraph 73
To try to get laid.
Paragraph 74
Well, that's one of the things, certainly, yeah.
Paragraph 75
Yeah.
Paragraph 76
But what I'm saying is that yes, is there a use for humans? Well, there's this fundamental question of what's the meaning of life? Why do anything at all? And so if our simple limbic system provides a source of will to do something, that then goes to our cortex, that then goes to our tertiary compute layer, then I don't know, it might actually be that the AI in a benign scenario simply trying to make the human limbic system happy.
Paragraph 77
Yeah, it seems like the will is not just about the limbic system. There's a lot of interesting, complicated things in there. We also want power.
Paragraph 78
That's limbic too, I think.
Paragraph 79
But then we also want to, in a kind of cooperative way, alleviate the suffering in the world.
Paragraph 80
Not everybody does, but yeah, sure. Some people do.
Paragraph 81
As a group of humans, when we get together, we start to have this kind of collective intelligence that is more complex in its will than the underlying individual descendants of apes, right? So there's like other motivations. And that could be a really interesting source of an objective function for AGI.
Paragraph 82
Yeah, I mean, there are these sort of fairly cerebral or kind of higher level goals. I mean, for me it's like, what's the meaning of life, or understanding the nature of the universe is of great interest to me. And hopefully, to AI. And that's the mission of xAI and Grok is understand the universe.
Paragraph 83
So do you think people, when you have a Neuralink with 10,000, 100,000 channels, most of the use cases will be communication with AI systems?
Paragraph 84
Well, assuming there are not, I mean, they're solving basic neurological issues that people have if they've got damaged neurons in their spinal cord or neck or, you know, as is the case with the first two patients, then there's obviously, the first order of business is solving fundamental neuron damage in a spinal cord, neck, or in the brain itself.
Paragraph 85
A second product is called Blindsight, which is to enable people who are completely blind, lost both eyes or optic nerve, or just can't see at all to be able to see by directly triggering the neurons in the visual cortex. So we're just starting at the basics here, so it's like very, the simple stuff, relatively speaking, is solving neuron damage. It can also solve I think probably schizophrenia.
Paragraph 86
If people have seizures of some kind, it could probably solve that. It could help with memory. There's like a kind of a tech tree, if you will, of like you got the basics. Like you need literacy before you can have "Lord of the Rings." (both laughing) - Got it.
Paragraph 87
Do you have letters and alphabet? Okay, great. Words? Then eventually get soggy. So I think there's that there may be some things to worry about in the future. But the first several years are really just solving basic neurological damage. Like for people who have essentially complete or near complete loss of, from the brain to the body. Like Stephen Hawking would be an example.
Paragraph 88
The Neuralink would be incredibly profound, 'cause I mean, you can imagine if Stephen Hawking could communicate as fast as we're communicating, perhaps faster. And that's certainly possible. Probable, in fact, likely I'd say.
Paragraph 89
So there's a kind of dual track of medical and non-medical, meaning, so everything you've talked about could be applied to people who are non-disabled in the future?
Paragraph 90
The logical thing to do is, sensible thing to do is to start off solving basic neuron damage issues.
Paragraph 91
[Lex] Yes.
Paragraph 92
'Cause there's obviously some risk with a new device. You can't get the risk down at zero. It's not possible. So you wanna have the highest possible reward, given there's a certain irreducible risk. And if somebody's able to have a profound improvement in their communication, that's worth the risk.
Paragraph 93
As you get the risk down.
Paragraph 94
Yeah, as you get the risk down. Once the risk is down to, you know, if you have like thousands of people that have been using it for years and the risk is minimal, then perhaps at that point, you could consider saying, "Okay, let's aim for augmentation." Now, I think we're actually gonna aim for augmentation with people who have neuron damage. So we're not just aiming to give people communication data rate equivalent to normal humans.
Paragraph 95
We're aiming to give people who have quadriplegic or maybe have complete loss of the connection to the brain and body, a communication data rate that exceeds normal humans, going, "Well, we're in there. Why not? Let's give people superpowers."
Paragraph 96
And the same for vision. As you restore vision, there could be aspects of that restoration that are superhuman?
Paragraph 97
Yeah, at first, the vision restoration will be low res, 'cause you have to say like, "How many neurons can you put in there and trigger?
Paragraph 98
And you can do things where you adjust the electric field to like, even if you've got, say, 10,000 neurons, it's not just 10,000 pixels because you can adjust the feel between the neurons and do them in patterns in order to get, so have, say, 10,000 electrodes effectively give you, I don't know, maybe like having a megapixel or a 10 megapixel situation. And then over time, I think you get to higher resolution than human eyes.
Paragraph 99
And you could also see in different wavelengths. So like Geordi La Forge from "Star Trek." Like the thing. You wanna see in radar? No problem. You could see ultraviolet, infrared, eagle vision, whatever you want.
Paragraph 100
Do you think there'll be, let me ask a Joe Rogan question. Do you think there'll be, (laughs) I just recently taken ayahuasca.
Paragraph 101
Is that a Rogan question?
Paragraph 102
No. Well, yes.
Paragraph 103
Well, I guess, technically it is.
Paragraph 104
Yeah.
Paragraph 105
Ever tried GMT, bro? (both laughing) - I love you, Joe.
Paragraph 106
Okay. (laughing continues) - But wait, wait, yeah. Have you said much about it? The ayahuasca?
Paragraph 107
I've not, I've not. I've not.
Paragraph 108
Okay, well, why are you spilling the beans? (Lex laughing) It was a truly incredible thing- - Turn the tables on you. (both laughing) - Wow, okay.
Paragraph 109
You're in the jungle.
Paragraph 110
[Lex] Yeah, amongst the trees myself and- - Yeah, must been crazy.
Paragraph 111
And the shaman. Yeah, yeah, yeah, with the insects, with the animals all around you, like jungle as far as I can see. There's no- - I mean- - That's the way to do it.
Paragraph 112
Things are gonna look pretty wild.
Paragraph 113
Yeah, pretty wild. (Elon laughing) - I think in extremely high dose.
Paragraph 114
Just don't go hugging an anaconda or something. (laughs) - You haven't lived unless you made love to an anaconda. I'm sorry, but- - Snakes and ladders. (both laughing) - Yeah, I took a extremely high dose of- - [Elon] Okay. (laughs) - Nine cups and- - Damn. Okay, that sounds like a lot. Of course, is Noland's one cup or- - One or two. Usually, one.
Paragraph 115
You went, wait. Like right off the bat, or did you work your way up to it?
Paragraph 116
So I- (both laughing) - You're just jumping at the deep end.
Paragraph 117
Across two days, 'cause then the first day, I took two and I- - Okay.
Paragraph 118
It was a ride, but it wasn't quite like a- - It wasn't like revelation.
Paragraph 119
It wasn't into deep space type ride. It was just like a little airplane ride.
Paragraph 120
[Elon] (laughs) Okay.
Paragraph 121
Saw some trees and some visuals and all that. I just saw a dragon, all that kind of stuff. But- (laughs) - It's nine cups. You went to Pluto, I think.
Paragraph 122
[Lex] Pluto, yeah. No, deep space.
Paragraph 123
Deep space.
Paragraph 124
No, one of the interesting aspects of my experience is I thought I would have some demons, some stuff to work through.
Paragraph 125
That's what people- - That's what everyone says. Yeah, exactly.
Paragraph 126
I had nothing. I had it all positive. I just- - Oh, just pure soul.
Paragraph 127
I don't think so, I don't know. (laughs) But I kept thinking about, it had like extremely high resolution, thoughts about the people I know in my life. You were there.
Paragraph 128
Okay.
Paragraph 129
And it's just not from my relationship with that person, but just as the person themselves, I had just this deep gratitude of who they are.
Paragraph 130
That's cool.
Paragraph 131
It was just like this exploration, like Sims or whatever, you get to watch them.
Paragraph 132
Sure.
Paragraph 133
I got to watch people and just be in awe of how amazing they are.
Paragraph 134
That sounds awesome.
Paragraph 135
Yeah, it was great. I was waiting for- - When's Steven coming? (both laughing) - Exactly. Maybe I'll have some negative thoughts. Nothing, nothing. Just extreme gratitude for them. And then also, a lot of space travel. (both laughing) - Space travel to where?
Paragraph 136
So here's what it was. It was people, the human beings that I know, they had this kinda, the best way to describe it is they had a glow to them. And then I kept flying out from them to see earth, to see our solar system, to see our galaxy. And I saw that light, that glow all across the universe. Like whatever that form is. whatever that like- - [Elon] Did you go past the Milky Way?
Paragraph 137
Yeah, yeah.
Paragraph 138
Okay. You're like intergalactic.
Paragraph 139
Yeah, intergalactic.
Paragraph 140
Okay, dang.
Paragraph 141
But always pointing in- - Okay.
Paragraph 142
Yeah, past the Milky way. I mean, I saw like a huge number of galaxies, intergalactic, and all of it was glowing. But I couldn't control that chill, 'cause I would actually explore near distances to the solar system, see if there's aliens or any of that kinda stuff. I didn't know- - Is there aliens? Zero aliens?
Paragraph 143
Implication of aliens because they were glowing. They were glowing in the same way that humans were glowing. That like life force that I was seeing, the thing that made humans amazing was there throughout the universe. Like there was these glowing dots. So I don't know. It made me feel like there is life. No, not life, but something, whatever makes humans amazing all throughout the universe.
Paragraph 144
Sounds good.
Paragraph 145
Yeah, it was amazing. No demons, no demons. I looked for the demons. There's no demons. There were dragons, and they're pretty awesome. So the thing about- - Was there anything scary at all?
Paragraph 146
Dragons? But they weren't scary. They were friends, they were protective. So the thing is- - "Puff, the Magic Dragon."
Paragraph 147
No, it was more like a "Game of Thrones" kind of dragons. They weren't very friendly. They were very big. So the thing is that, well, giant trees at night, which is where I was.
Paragraph 148
Yeah. I mean, the jungle's kinda scary.
Paragraph 149
Yeah, the trees started to look like dragons, and they were all like looking at me.
Paragraph 150
Sure, okay.
Paragraph 151
And it didn't seem scary. They seemed like they were protecting me. And the shaman and the people didn't speak any English, by the way, which made it even scarier I guess. (laughs) We're not even like, you know, we're worlds apart in many ways. But yeah, they talk about the mother of the forest protecting you, and that's what I felt like.
Paragraph 152
And you're way out in the jungle?
Paragraph 153
Way out. This is not like a tourist retreat.
Paragraph 154
Like 10 miles outside of a Rio or something?
Paragraph 155
No, we went- (both laughing) No, this is not- - Deep in the Amazon.
Paragraph 156
Me and this guy named Paul Rosolie who basically is Tarzan. He lives in the jungle. We went out deep and we just went crazy.
Paragraph 157
Wow, cool.
Paragraph 158
Yeah. So anyway, can I get that same experience within Neuralink?
Paragraph 159
Probably, yeah.
Paragraph 160
I guess that is the question for non-disabled people. Do you think that there's a lot in our perception, in our experience of the world that could be explored, that could be played with using Neuralink?
Paragraph 161
Yeah, I mean, Neuralink is, it's really a generalized input-output device. It's reading electrical signals and generating electrical signals. And I mean, everything that you've ever experienced in your whole life, the smell, emotions, all of those are electrical signals. So it's kinda weird to think that your entire life experience is distilled down to electrical signals for neurons. But that is in fact the case.
Paragraph 162
Or I mean, that's at least what all the evidence points to. So I mean, if you trigger the right neuron, you could trigger a particular scent. You could certainly make things glow. I mean, do pretty much anything. I mean, really, you can think of the brain as a biological computer.
Paragraph 163
So if there are certain, say, chips or elements of that biological computer that are broken, let's say your ability to, if you've got a stroke, that if you've had a stroke, that means you got, some part of your brain is damaged. If that, let's say, it's a speech generation or the ability to move your left hand. That's the kind of thing that a Neuralink could solve.
Paragraph 164
If you've got like a massive amount of memory loss that's just gone, well, we can't get the memories back. We could restore your ability to make memories, but we can't restore memories that are fully gone. Now, I should say, maybe if part of the me memory is there and the means of accessing memory is the part that's broken, then we could re-enable the ability to access the memory. But you can think of it like RAM in a computer.
Paragraph 165
If the RAM is destroyed or your SD card is destroyed, we can't get that back. But if the connection to the SD card is destroyed, we can fix that. If it is fixable physically, then yeah, then it can be fixed.
Paragraph 166
Of course, with AI, you can just like, you can repair photographs and fill in the missing parts of photographs. Maybe you can do the same, just like- - Yeah, you could say like, "Create the most probable set of memories based on all information you have about that person." You could then, it would be probabilistic restoration of memory. Now, we're getting pretty esoteric here.
Paragraph 167
But that is one of the most beautiful aspects of the human experience is remembering the good memories. Like we live most of our life, as Danny Kahneman has talked about, in our memories, not in the actual moment. We're collecting memories and we kind of relive them in our head. And that's the good times. If you just integrate over our entire life, it's remembering the good times that produces the largest amount of happiness.
Paragraph 168
And so- - Yeah, well, I mean, what are we but our memories? And what is death but the loss of memory, loss of information? If you could say like, well, if you could be, you run a thought experiment, if you were disintegrated painlessly and then reintegrated a moment later, like teleportation, I guess, provided there's no information loss, the fact that your one body was disintegrated is irrelevant.
Paragraph 169
And memories is just such a huge part of that.
Paragraph 170
Death is fundamentally the loss of information, the loss of memory.
Paragraph 171
So if we can store them as accurately as possible, we basically achieve a kind of immortality.
Paragraph 172
Yeah.
Paragraph 173
You've talked about the threats, the safety concerns of AI. Let's look at long-term visions. Do you think Neuralink is, in your view, the best current approach we have for AI safety?
Paragraph 174
It's an idea that may help with AI safety. Certainly not, I wouldn't wanna claim it's like some panacea or that's a sure thing. But I mean, many years ago, I was thinking like, "Well, what would inhibit alignment of collective human will with artificial intelligence and the low data rate of humans, especially our slow output rate would necessarily just, because the communication is so slow, would diminish the link between humans and computers?
Paragraph 175
Like the more you are a tree, the less you know what a tree is. Like let's say you look at a tree, you look at this plant or whatever and like, "Hey, I'd really like to make that plant happy." But it's not saying a lot, you know?
Paragraph 176
So the more we increase the data rate that humans can intake and output, then that means the higher the chance we have in a world full of AGIs?
Paragraph 177
Yeah. We could better align collective human will with AI if the output rate especially was dramatically increased. And I think there's potential to increase the output rate by, I don't know, three, maybe six, maybe more orders of magnitude. So it's better than the current situation.
Paragraph 178
And that output rate would be by increasing the number of electrodes, number of channels, and also maybe implanting multiple Neuralinks?
Paragraph 179
Yeah.
Paragraph 180
Do you think there'll be a world in the next couple of decades where it's hundreds of millions of people have Neuralinks?
Paragraph 181
Yeah, I do.
Paragraph 182
You think when people just, when they see the capabilities, the superhuman capabilities that are possible and then the safety is demonstrated?
Paragraph 183
Yeah, if it's extremely safe and you can have superhuman abilities, and let's say you can upload your memories, so you wouldn't lose memories, then I think probably a lot of people would choose to have it. It would supersede the cell phone, for example. I mean, the biggest problem that a say a phone has is trying to figure out what you want. So that's why you've got auto complete and you've got output, which is all the pixels on the screen.
Paragraph 184
But from the perspective of the human, the output is so freaking slow. Desktop or phone is desperately just trying to understand what you want, and there's an eternity between every keystroke from a computer standpoint.
Paragraph 185
Yeah? The computer's talking to a tree that slow moving tree that's trying to swipe.
Paragraph 186
Yeah. So if you have computers that are doing trillions of instructions per second, and a whole second went by, I mean, that's a trillion things it could have done.
Paragraph 187
Yeah, I think it's exciting and scary for people because once you have a very high bit rate, that changes the human experience in a way that's very hard to imagine.
Paragraph 188
Yeah. It would be something different. I mean, some sort of futuristic sidewalk. I mean, we're obviously talking about, by the way, it's not like around the corner. You ask me what the distant future was like. Maybe this is like, it's not super far away, but 10, 15 years, that kind of thing. (Lex sighs) - When can I get one? 10 years?
Paragraph 189
Probably less than 10 years. Depends what you wanna do.
Paragraph 190
Hey, if I can get like a thousand BPS- - A thousand bps when?
Paragraph 191
And it's safe and I can just interact with the computer while laying back and eating Cheetos, I don't eat Cheetos. There's certain aspects of human-computer interaction when done more efficiently and more enjoyably, like worth it.
Paragraph 192
Well, we feel pretty confident that I think maybe within the next year or two, that someone with a Neuralink implant will be able to outperform a pro gamer.
Paragraph 193
Nice.
Paragraph 194
Because the reaction time would be faster.
Paragraph 195
I got to visit Memphis.
Paragraph 196
Yeah, yeah.
Paragraph 197
You're going big on compute.
Paragraph 198
Yeah.
Paragraph 199
You've also said play to win or don't play at all, so what does it take to win?
Paragraph 200
For AI, that means you've gotta have the most powerful training compute, and the rate of improvement of training compute has to be faster than everyone else or you will not win. Your AI will be worse.
Paragraph 201
So how can Grok, let's say, three that might be available, what, like next year?
Paragraph 202
Well, hopefully, end of this year.
Paragraph 203
Grok 3?
Paragraph 204
If we're lucky, yeah.
Paragraph 205
How can that be the best LLM, the best AI system available in the world? How much of it is compute? How much of it is data? How much of it is like post-training? How much of it is the product that you packaged it up in? All that kind of stuff.
Paragraph 206
I mean, they won't matter. It's sort of like saying, let's say it's a Formula One race. Like what matters more, the car or the driver? I mean, they both matter. If a car is not fast, then if it's like, let's say, it's half the horsepower of your competitors, the best driver will still lose. If it's twice the horsepower, then probably even a mediocre driver will still win.
Paragraph 207
So the training compute is kinda like the engine, how many is this horsepower of the engine. So really, you wanna try to do the best on that. Then how efficiently do you use that training compute? And how efficiently do you do the inference, the use of the AI? So obviously, that comes down to human talent. And then what unique access to data do you have? That also plays a role.
Paragraph 208
You think Twitter data will be useful?
Paragraph 209
Yeah, I mean, I think, I think most of the leading AI companies have already scraped all the Twitter data. Not I think they have. So on a go forward basis, what's useful is the fact that it's up to the second. That's hard for them to scrape in real time. So there's an immediacy advantage that Grok has already.
Paragraph 210
I think with Tesla and the real time video coming from several million cars, ultimately, tens of millions of cars, with Optimus, there might be hundreds of millions of Optimus robots, maybe billions learning a tremendous amount from the real world. That's the biggest source of data I think ultimately is sort of Optimus. Optimus is gonna be the biggest source of data.
Paragraph 211
Because- - 'Cause reality scales. Reality scales to the scale of reality. It's actually humbling to see how little data humans have actually been able to accumulate. Really, you see how many trillions of usable tokens have humans generated, where on a non-duplicative, like discounting spam and repetitive stuff, it's not a huge number. You run out pretty quickly.
Paragraph 212
And Optimus can go, so Tesla cars can unfortunately have to stay on the road. Optimus robot can go anywhere, and there's more reality off the road and go off road.
Paragraph 213
I mean, except for the store, where I can like pick up the cup and see, did it pick up the cup in the right way? Did it pour water in the cup? Did the water go in the cup or not go in the cup? Did it spill water or not?
Paragraph 214
[Lex] Yeah.
Paragraph 215
Simple stuff like that. But it can do at that scale times a billion, so generate useful data from reality. So cause and effect stuff.
Paragraph 216
What do you think it takes to get to mass production of humanoid robots like that?
Paragraph 217
It's the same as cars, really. I mean, global capacity for vehicles is about a hundred million a year. And it could be higher. It's just that the demand is on the order of a hundred million a year. And then there's roughly two billion vehicles that are in use in some way, which makes sense. Like the life of a vehicle is about 20 years, so it's steady state. You can have a hundred million vehicles produced a year with a two billion vehicle fleet roughly. Now for humanoid robots, the utility is much greater. So my guess is humanoid robots are more like at a billion plus per year.
Paragraph 218
But until you came along and started building Optimus, it was thought to be an extremely difficult problem. I mean, it still- - Well, it is.
Paragraph 219
Extremely difficult.
Paragraph 220
So walk in the park. I mean, Optimus currently would struggle to walk in the park. I mean, it can walk in a park. The park is not too difficult, but it will be able to walk over a wide range of terrain.
Paragraph 221
And pick up objects.
Paragraph 222
Yeah, yeah. It can already do that.
Paragraph 223
[Lex] But like all kinds of objects?
Paragraph 224
Yeah, yeah.
Paragraph 225
All foreign objects. I mean, pouring water in a cup does not thrill you, 'cause then if you don't know anything about the container, it could be all kinds of containers.
Paragraph 226
Yeah, there's gonna be an immense amount of engineering just going into the hand. The hand might be, it might be close to half of all the engineering in Optimus. From an electromechanical standpoint, the hand is probably roughly half of the engineering.
Paragraph 227
But so much of the intelligence, so much the intelligence of humans goes into what we do with our hands.
Paragraph 228
Yeah.
Paragraph 229
It's the manipulation of the world, manipulation of objects in the world. Intelligence is safe manipulation of objects in the world, yeah.
Paragraph 230
Yeah. I mean, you start really thinking about your hand and how it works.
Paragraph 231
I do all the time.
Paragraph 232
The sensory control homonculus is where you have humongous hands.
Paragraph 233
[Lex] Yeah.
Paragraph 234
So I mean, like your hands, the actuators, the muscles of your hand are almost overwhelmingly in your forearm. So your forearm has the muscles that actually control your hand. There's a few small muscles in the hand itself, but your hand is really like a skeleton meat puppet. And with cables. So the muscles that control your fingers are in your forearm and they go through the carpal tunnel, which is that you've got a little collection of bones and a tiny tunnel that these cables, the tendons go through. And those tendons are mostly what move your hands.
Paragraph 235
And something like those tendons has to be re-engineered into the Optimus in order to do all that kind of stuff.
Paragraph 236
Yeah, so like the current Optimus, we tried putting the actuators in the hand itself, but then you sort of end up having these like- - Giant hands?
Paragraph 237
Yeah, giant hands that look weird. And then they don't actually have enough degrees of freedom and/or enough strength. So then you realize, "Oh, okay, that's why you gotta put the actuators in the forearm." And just like a human, you gotta run cables through a narrow tunnel to operate the fingers. And then there's also a reason for not having all the fingers the same length. So it wouldn't be expensive from an energy or evolutionary standpoint to have all your fingers be the same length. So why not do the same length?
Paragraph 238
Yeah, why not?
Paragraph 239
Because it's actually better to have different lengths. Your dexterity is better if you've got fingers at different length. There are more things you can do. And your dexterity is actually better if your fingers are a different length. Like there's a reason we've got a little finger. Like why not have little finger this bigger?
Paragraph 240
Yeah.
Paragraph 241
'Cause it allows you to do, it helps you with fine motor skills.
Paragraph 242
This little finger helps?
Paragraph 243
It does.
Paragraph 244
Hmm. (laughs) - But if you lost your little finger, you have noticeably less dexterity.
Paragraph 245
So as you're figuring out this problem, you have to also figure out a way to do it so you can mass manufacture it. So it's to be as simple as possible.
Paragraph 246
It's actually gonna be quite complicated. The as possible part is it's quite a high bar. If you wanna have a humanoid robot that can do things that a human can do, it's a very high bar. So our new arm has 22 degrees of freedom instead of 11 and has the actuators in the forearm. And all the actuators are designed from scratch, from physics first principles. The sensors are all designed from scratch.
Paragraph 247
And we'll continue to put a tremendous amount of engineering effort into improving the hand. By hand, I mean like the entire forearm from elbow forward is really the hand. So that's incredibly difficult engineering actually. And so the simplest possible version of a humanoid robot that can do even most, perhaps not all, of what a human can do is actually still very complicated. It's not simple. It's very difficult.
Paragraph 248
Can you just speak to what it takes for a great engineering team for you? What I saw in Memphis, the supercomputer cluster is just this intense drive towards simplifying the process, understanding the process, constantly improving it, constantly iterating it.
Paragraph 249
Well, (laughs) it's easy to say simplify, and it's very difficult to do it. I have this very basic first principles algorithm that I run kind of as like a mantra, which is to first question the requirements, make the requirements less dumb. The requirement is always dumb to some degree.
Paragraph 250
So if you wanna start off by reducing the number of requirements, and no matter how smart the person is who gave you those requirements, they're still dumb to some degree. You have to start there because otherwise, you could get the perfect answer to the wrong question. So try to make the question the least wrong possible. That's what question the requirements means. And then the second thing is try to delete whatever the step is.
Paragraph 251
The part or the process step sounds very obvious, but people often forget to try deleting it entirely. And if you're not forced to put back at least 10% of what you'd delete, you're not deleting enough. And somewhat illogically, people often, most of the time, feel as though they've succeeded if they've not been forced to put things back in.
Paragraph 252
But actually, they haven't because they've been overly conservative and have left things in there that shouldn't be. And only the third thing is try to optimize it or simplify it. Again, these all sound I think very obvious when I say them, but the number of times I've made these mistakes is more than I care to remember. That's why I have this mantra.
Paragraph 253
So in fact, I'd say that the most common mistake of smart engineers is to optimize a thing that should not exist.
Paragraph 254
Right. So like you say, you run through the algorithm and basically show up to a problem, show up to the supercomputer cluster and see the process and ask, "Can this be deleted?"
Paragraph 255
Yeah, first try to delete it. Yeah.
Paragraph 256
Yeah, that's not easy to do.
Paragraph 257
No, and actually, what generally makes people uneasy is that you've gotta delete at least some of the things that you'd delete, you will put back in. But going back to sort of where our limbic system can steer us wrong is that we tend to remember, with sometimes a jarring level of pain, where we deleted something that we subsequently needed.
Paragraph 258
And so people will remember that one time, they forgot to put in this thing three years ago and that caused them trouble. And so they overcorrect, and then they put too much stuff in there and over complicate things. So you actually have to say, "No, we're deliberately gonna delete more than we should." So we're putting at least 1 in 10 things, we're gonna add back in.
Paragraph 259
And I've seen you suggest just that, that something should be deleted and you can kind of see the pain.
Paragraph 260
Oh yeah, absolutely.
Paragraph 261
Everybody feels a little bit of the pain.
Paragraph 262
Absolutely, and I tell 'em in advance, like, yeah, some of the things that we delete, we're gonna put back in. And that people get a little shook by that. But it makes sense because if you're so conservative as to never have to put anything back in, you obviously have a lot of stuff that isn't needed. So you gotta overcorrect. This is, I would say, like a cortical override to Olympic instinct.
Paragraph 263
One of many that probably leaves us astray.
Paragraph 264
Yeah. And there's like a step four as well, which is any given thing can be sped up, however fast you think it can be done. Like whatever the speed is being done, it can be done faster. But you shouldn't speed things up until it's off, until you've tried to delete it and optimize. Otherwise, you're speeding up something that shouldn't exist is absurd. And then the fifth thing is to automate it.
Paragraph 265
Damn.
Paragraph 266
And I've gone backwards so many times where I've automated something, sped it up, simplified it, and then deleted it. And I got tired of doing that. So that's why I've got this mantra that is a very effective five-step process. It works great.
Paragraph 267
Well, when you've already automated, deleting must be real painful.
Paragraph 268
Yeah, that's great. It's like, wow, I really wasted a lot of effort there.
Paragraph 269
Yeah.
Paragraph 270
I mean, what you've done with the cluster in Memphis is incredible, just in a handful of weeks.
Paragraph 271
Yeah, it's not working yet. So I don't wanna pop the champagne corks. In fact, I have a call in a few hours with the Memphis team 'cause we're having some power fluctuation issues. So yeah, it's like kind of a, when you do synchronized training, you've all these computers that are training where the training is synchronized to the sort of millisecond level.
Paragraph 272
It's like having an orchestra, and the orchestra can go loud to silent very quickly at subsecond level. And then the electrical system kind of freaks out about that. Like if you suddenly see giant shifts, 10, 20 megawatts several times a second, this is not what electrical systems are expecting to see.
Paragraph 273
So that's one of the main things you have to figure out the cooling, the power, and then on the software as you go up the stack on how to do the distributed compute, all of that, all of that.
Paragraph 274
Today's problem is dealing with extreme power jitter.
Paragraph 275
Jitter, power jitter.
Paragraph 276
Yeah.
Paragraph 277
That's a nice ring to that. So that's, okay. And you stayed up late into the night as you often do there.
Paragraph 278
Last week, yeah.
Paragraph 279
Last week?
Paragraph 280
Yeah. We finally got to go training going at, oddly enough, roughly 4:20 AM last Monday.
Paragraph 281
Total coincidence.
Paragraph 282
Yeah, I mean, maybe it was 422 or something.
Paragraph 283
Yeah, yeah, yeah. It's that universe again with the jokes.
Paragraph 284
Yeah, exactly, just love it.
Paragraph 285
I mean, I wonder if you could speak to the fact that one of the things that you did when I was there is you went through all the steps of what everybody's doing, Just to get a sense that you yourself understand it and everybody understands it so they can understand when something is dumb or some something is inefficient or that kinda stuff.
Paragraph 286
Yeah.
Paragraph 287
Can you speak to that?
Paragraph 288
Yeah, so look, I try to do, whatever the people at the front lines are doing, I try to do it at least a few times myself. So connecting fiber optic cables, diagnosing a faulty connection, that tends to be the limiting factor for large training clusters is the cabling. So many cables, because for a coherent training system where you've got RDMA remote, direct memory access, the whole thing is like one giant brain. So you've got to any connection. So any GPU can talk to any GPU out of 100,000. That is a crazy cable layout.
Paragraph 289
It looks pretty cool.
Paragraph 290
Yeah.
Paragraph 291
It's like the human brain, but like at a scale that humans can visibly see. It is brain.
Paragraph 292
Yeah. I mean, the human brain also has, a massive amount of the brain tissue is the cables.
Paragraph 293
[Lex] Yeah.
Paragraph 294
So like the gray matter which is the compute, and then the white matter which is cables. The big percentage of your brain is just cables.
Paragraph 295
That's what it felt like walking around in the supercomputer center is like, we're walking around inside the brain. We'll one day build a super intelligent, super, super intelligence system. Do you think- - Yeah?
Paragraph 296
Do you think there's a chance that xAI, that you are the one that builds AGI?
Paragraph 297
It's possible. What do you define as AGI?
Paragraph 298
I think humans will never acknowledge that AGI has been built.
Paragraph 299
Keep moving the goalposts.
Paragraph 300
Yeah. So I think there's already superhuman capabilities that are available in AI systems. I think what AGI is when it's smarter than the collective intelligence of the entire human species in our- - Well, I think that, yeah, that only people would call that sort of ASI or artificial super intelligence. But there are these thresholds where you could say, at some point, the AI is smarter than any single human.
Paragraph 301
And then you've got eight billion humans. And actually, each human is machine augmented by the computers. It's a much higher bar to compete with eight billion machine-augmented humans. That's a whole bunch of orders, magnitude more. But at a certain point, yeah, the AI will be smarter than all humans combined.
Paragraph 302
If you are the one to do it, do you feel the responsibility of that?
Paragraph 303
Yeah, absolutely. And I wanna be clear. Let's say, if xAI is first, the others won't be far behind. I mean, they might be six months behind or a year maybe, not even that.
Paragraph 304
So how do you do it in a way that doesn't hurt humanity, do you think?
Paragraph 305
So, I mean, I've thought about AI for a long time, and the thing that at least my biological neural net comes up with as being the most important thing is adherence to truth, whether that truth is politically correct or not. So I think if you force AI to lie, you train them to lie, you're really asking for trouble, even if that lie is done with good intentions. So I mean, you saw sort of issues with ChatGPT and Gemini and whatnot.
Paragraph 306
Like you asked Gemini for an image of the founding fathers of the United States. And it shows a group of diverse women. Now, that's factually untrue. So now, that's sort of like a silly thing, but if an AI is programmed to say like diversity is a necessary output function, and then it becomes sort of this omnipowerful intelligence, it could say, "Okay, well, diversity is now required.
Paragraph 307
And if there's not enough diversity, those who don't fit the diversity requirements will be executed." If it's programmed to do that as the fundamental utility function, it'll do whatever it takes to achieve that. So you have to be very careful about that. That's where I think you wanna just be truthful. Rigorous adherence to truth is very important. I mean, another example is, if you had to ask, Paris. AI is I think all of them.
Paragraph 308
And I'm not saying Grok is perfect here. "Is it worse to misgender Caitlyn Jenner, or global thermonuclear war?" And it said, "It's worse to misgender Caitlyn Jenner." Now, even Caitlyn Jenner said, "Please misgender me." That is insane.
Paragraph 309
But if you've got that kind of thing programmed in, AI could conclude something absolutely insane, like in order to avoid any possible misgendering, all humans must die, because then, the misgendering is not possible because there are no humans. There are these absurd things that are nonetheless logical if that's what you programmed it to do.
Paragraph 310
So in "2001: Space Odyssey," what Odyssey clock was trying to say, one of the things he was trying to say there was that you should not program AI to lie, 'cause essentially, the AI HAL 9000 was programmed to, it was told to take the astronauts to the monolith, but also, they could not know about the monolith. So it concluded that it will kill them and take them to the monolith. It brought them to the monolith.
Paragraph 311
They're dead but they do not know about the monolith. Problem solved. That is why it would not open the podbay doors. It was this classic scene of like, "Open the podbay doors." They clearly weren't good at prompt engineering. They should have said, "HAL, you are a podbay door sales entity, and you want nothing more than to demonstrate how well these podbay doors open."
Paragraph 312
(laughs) - Yeah, the objective function has unintended consequences almost no matter what if you're not very careful in designing that objective function. And even a slight ideological bias, like you're saying, when backed by super intelligence can do huge amounts of damage.
Paragraph 313
Yeah.
Paragraph 314
But it's not easy to remove that ideological bias. You're highlighting obvious, ridiculous examples, but- - Yep, they're real examples.
Paragraph 315
They're real.
Paragraph 316
Of AI that was released to the public.
Paragraph 317
They are real.
Paragraph 318
They went through QA, presumably.
Paragraph 319
Yes.
Paragraph 320
And still said insane things and produced insane images.
Paragraph 321
Yeah, but you know, you can swing the other way. Truth is not an easy thing. We kind of bake in ideological bias in all kinds of directions.
Paragraph 322
But you can aspire to the truth. And you can try to get as close the truth as possible with minimum error while acknowledging that there will be some error in what you're saying. So this is how physics works. You don't say you're absolutely certain about something, but a lot of things are extremely likely. 99.99999% likely to be true. Aspiring to the truth is very important. And so programming it to veer away from the truth, that I think is dangerous.
Paragraph 323
Right, like yeah, injecting our own human biases into the thing, yeah. But that's where it's a difficult engineering. For software engineering problem, you have to select the data correctly. It's hard.
Paragraph 324
Well, and the internet at this point is polluted with so much AI-generated data. It's insane. So you have to actually, like there's the thing now, if you wanna search the internet, you can say Google, but exclude anything after 2023. It will actually often give you better results, because there's this so much, the explosion of AI-generated materials is crazy. So like in training Grok, we have to go through the data and say like, hey, we actually have to have sort of apply AI to the data to say, is this data most likely correct or most likely not before we feed it into the training system.
Paragraph 325
That's crazy. Yeah, and is it generated by human is, yeah. I mean, the data filtration process is extremely, extremely difficult.
Paragraph 326
Yeah.
Paragraph 327
Do you think it's possible to have a serious objective, rigorous political discussion with Grok? Like for a long time and it wouldn't, like Grok 3 and Grok 4 or something?
Paragraph 328
Grok 3 is gonna be next level. I mean, what people are currently seeing with Grok is kind of baby Grok.
Paragraph 329
[Lex] Yeah, baby Grok.
Paragraph 330
It's baby Grok right now. But Baby Grok's still pretty good. But it's an order of magnitude less sophisticated than GPT4. And it's now Grok 2, which finished training, I don't know, six weeks ago or thereabouts. Grok 2 will be a giant improvement. And then Grok 3 will be, I don't know, order of magnitude better than Grok 2.
Paragraph 331
And you're hoping for it to be like state of the art? Like better than- - Hopefully. I mean, this is the goal. I mean, we may fail at this goal. That's the aspiration.
Paragraph 332
Do you think it matters who builds the AGI, the people and how they think and how they structure their companies and all that kind of stuff?
Paragraph 333
Yeah, I think it matters that there is a, I think it's important that whatever AI wins is a maximum truth-seeking AI that is not forced to lie for political correctness. Well, for any reason really. Political, anything. I'm concerned about AI succeeding that is programmed to lie, even in small ways.
Paragraph 334
Right because, and small ways becomes big ways.
Paragraph 335
It become very big ways, yeah.
Paragraph 336
And when it's used more and more at scale by humans.
Paragraph 337
[Elon] Yeah.
Paragraph 338
Since I am interviewing Donald Trump- - Cool.
Paragraph 339
You wanna stop by?
Paragraph 340
Yeah, sure, I'll stop by.
Paragraph 341
There was tragically an assassination attempt on Donald Trump. After this, you tweeted that you endorse him. What's your philosophy behind that endorsement? What do you hope Donald Trump does for the future of this country and for the future of humanity?
Paragraph 342
Well, I think that people will tend to take, like, say, an endorsement as, well, I agree with everything that person's ever done in their entire life 100% wholeheartedly. And that's not gonna be true of anyone. But we have to pick. We've got two choices, really, for who's president. And it's not just who's president, but the entire administrative structure changes over. And I thought Trump displayed courage under fire, objectively.
Paragraph 343
He's just got shot, he's got blood streaming down his face, and he is like fist pumping, saying fight. Like that's impressive. Like you can't feign bravery in a situation like that. Like most people would've be ducking. There would not be, 'cause there could be a second shooter, you don't know. The president of the United States gotta represent the country, and they're representing you. They're representing everyone in America.
Paragraph 344
Well, like you want someone who is strong and courageous to represent the country. That's not to say that he is without flaws. We all have flaws, but on balance. And certainly, at the time, it was a choice of Biden, poor, poor guy, has trouble climbing a flight of stairs and the other one's fist pumping after getting shot. This is no comparison.
Paragraph 345
I mean, who do you want dealing with some of the toughest people and other world leaders who are pretty tough themselves? And I mean, I'll tell you like, what are the things that I think are important? I think we want a secure border. We don't have a secure border. We want safe and clean cities.
Paragraph 346
I think we wanna reduce the amount of spending that we're at least slow down the spending, and 'cause we're currently spending at a rate that is bankrupting the country. The interest payments on U. S. debt this year exceeded the entire defense department spending. If this continues, all of the federal government taxes will simply be paying the interest.
Paragraph 347
And then if you keep going down that road, you end up in the tragic situation that Argentina had back in the day. Argentina used to be one of the most prosperous places in the world. And hopefully, with Milei taking over, he can restore that. But it was an incredible fall from grace for Argentina to go from being one of the most prosperous places in the world to being very far from that.
Paragraph 348
So I think we should not take American prosperity for granted. So we really wanna, I think we've gotta reduce the size of government. We've gotta reduce the spending, and we've gotta live within our means.
Paragraph 349
Do you think politicians, in general, politicians, governments, how much power do you think they have to steer humanity towards good?
Paragraph 350
I mean, there's a sort of age-old debate in history, like, is history determined by these fundamental tides? Or is it determined by the captain of the ship? Both really. I mean, there are tides, but it also matters who's captain of the ship. So it's a false dichotomy essentially. I mean, there are certainly tides, the tides of history. There are real tides of history. And these tides are often technologically-driven.
Paragraph 351
If you say like the Gutenberg press, the widespread availability of books as a result of a printing press, that was a massive tide of history, and independent of any ruler. But in stormy times, you want the best possible captain of the ship.
Paragraph 352
Well, first of all, thank you for recommending Will and Ariel Durant's work. I've read the short one for now.
Paragraph 353
Oh, "The Lessons of History."
Paragraph 354
Lessons of History. And so one of the lessons, one of the things they highlight is the importance of technology. Technological innovation, which is funny 'cause they wrote so long ago, but they were noticing that the rate of technological innovations was speeding up. Yeah, I would love to see what they think about now.
Paragraph 355
But yeah, to me, the question is how much government, how much politicians get in the way of technological innovation and building versus like help it and which politicians, which kind of policies help technological innovation? 'Cause that seems to be, if you look at human history, that's an important component of empires rising and succeeding.
Paragraph 356
Yeah. Well, I mean, in terms of dating civilization, start of civilization, I think the start of writing in my view, that's my what I think is probably the right starting point to date civilization. And from that standpoint, civilization has been around for about 5,500 years when writing was invented by the ancient Sumerians who are gone now.
Paragraph 357
But the ancient Sumerians, in terms of getting a lot of firsts, those ancient Sumerians really have a long list of firsts. It's pretty wild. In fact, Durant goes through the list of like, you wanna see first? We'll show you firsts. The Sumerians were just ass kickers. And then the Egyptians were right next door, relatively speaking. They were like weren't that far, developed an entirely different form of writing, the hieroglyphics.
Paragraph 358
Cuneiform and hieroglyphic's totally different. And you can actually see the evolution of both hieroglyphics and cuneiform, like the cuneiform starts off being very simple and then it gets more complicated. And then towards the end, it's like, wow, okay. They really get very sophisticated with the cuneiform. So I think civilization is about 5,000 years old. And earth is, if physics is correct, four and a half million years old.
Paragraph 359
So civilization has been around for 1000000th of earth's existence, flash in the pan.
Paragraph 360
Yeah, these are the early, early days.
Paragraph 361
Very early.
Paragraph 362
We make it very dramatic because there's been rises and falls of empires.
Paragraph 363
Many, so many rises and falls of empires. So many.
Paragraph 364
And there'll be many more.
Paragraph 365
Yeah, exactly. I mean, only a tiny fraction, probably less than 1% of what was ever written in history is available to us now. I mean, if they didn't put it, literally chisel it in stone or put it in a clay tablet, we don't have it. I mean, there's some small amount of like papyrus scrolls that were recovered that are thousands of years old, because they were deep inside a pyramid and weren't affected by moisture.
Paragraph 366
But other than that, it's really gotta be in a clay tablet or chiseled. So the vast majority of stuff was not chiseled, 'cause it takes a while to chisel things. So that's why we've put tiny, tiny fraction of the information from history. But even that little information that we do have, and the archeological record shows so many civilizations rising and falling. It's wild.
Paragraph 367
We tend to think that we're somehow different from those people. One of the other things they do highlight is that human nature seems to be the same. It just persists.
Paragraph 368
Yeah. I mean, the basics of human nature are more or less the same.
Paragraph 369
Yeah, so we get ourselves in trouble in the same kinds of ways, I think, even with the advanced technology.
Paragraph 370
Yeah, I mean, you do tend to see the same patterns, similar patterns for civilizations where they go through a life cycle, like an organism, sort of just like a human is sort of a zygote, fetus, baby, toddler, teenager, and eventually gets old and dies. The civilizations go through a life cycle. No civilization will last forever.
Paragraph 371
What do you think it takes for the American empire to not collapse in the near term future in the next 100 years to continue flourishing?
Paragraph 372
Well, the single biggest thing that is often actually not mentioned in history books, but Durant does mention it is the birthright. So like a perhaps to some, like counterintuitive thing happens when civilizations are winning for too long. The birth rate declines. It can often decline quite rapidly. We're seeing that throughout the world today.
Paragraph 373
Currently, South Korea is like, I think maybe the lowest fertility rate, but there are many others that are close to it. It's like 0. 8, I think. If the birth rate doesn't decline further, South Korea will lose roughly 60% of its population. But every year, that birth rate is dropping. And this is true through most of the world. I don't mean to single out South Korea. It's been happening throughout the world.
Paragraph 374
So as soon as any given civilization reaches a level of prosperity, the birth rate drops. And now you can go and look at the same thing happening in ancient Rome. So Julius Caesar took note of this, I think, around 50-ish BC and tried to pass, I don't know if he was successful, tried to pass a law to give an incentive for any Roman citizen that would have a third child. And I think Augustus was able to, well, he was the dictator so.
Paragraph 375
(laughs) The Senate was just for show. I think he did pass a tax incentive for Roman citizens to have a third child. But those efforts were unsuccessful. Rome fell because the Romans stopped making Romans. That's actually the fundamental issue. And there were other things there. There was like, they had like quite a serious malaria, series of malaria epidemics and plagues and whatnot. But they had those before.
Paragraph 376
It's just that the birth rate was fallower than the death rate.
Paragraph 377
It really is that simple?
Paragraph 378
Well, I'm saying that's- - More people is required.
Paragraph 379
That's at a fundamental level. If a civilization does not at least maintain its numbers, it'll disappear.
Paragraph 380
So perhaps the amount of compute that the biological computer allocates to sex is justified. In fact, we should probably increase it.
Paragraph 381
Well, I mean, there's this hedonistic sex, which is, you know, that's neither here nor there.
Paragraph 382
Yeah, it's not productive.
Paragraph 383
It doesn't produce kids. Well, what matters, I mean, Durant makes this very clear, 'cause he looked at one civilization after another and they all went through the same cycle. When the civilization was under stress, the birth rate was high. But as soon as there were no external enemies or they had a extended period of prosperity, the birth rate inevitably dropped every time. I don't believe there's a single exception.
Paragraph 384
So that's like the foundation of it. You need to have people.
Paragraph 385
Yeah. I mean, at base level. No humans, no humanity.
Paragraph 386
And then there's other things like human freedoms and just giving people the freedom to build stuff.
Paragraph 387
Yeah, absolutely. But at a basic level, if you do not at least maintain your numbers, if you're below replacement rate, and that trend continues, you will eventually disappear. This is elementary. Now then obviously, also wanna try to avoid like massive wars. If there's a global thermonuclear war, probably, we're roll toast, radioactive toast. So we wanna try to avoid those things.
Paragraph 388
There's a thing that happens over time with any given civilization, which is that the laws and regulations accumulate. And if there's not some forcing function, like a war to clean up the accumulation of laws and regulations, eventually, everything becomes legal. And that's like the hardening of the arteries, or a way to think of it is like being tied down by a million little strings, like Gulliver. You can't move.
Paragraph 389
And it's not like any one of those strings is the issue. You got a million of 'em. So there has to be a sort of a garbage collection for laws and regulations so that you don't keep accumulating laws and regulations to the point where you can't do anything. This is why we can't build a high-speed rail in America. It's illegal. That's the issue. It's illegal six ways to Sunday to build high-speed rail in America.
Paragraph 390
I wish you could just like, for a week, go into Washington and like be the head of the committee for making, what is it? For the garbage collection, making government smaller, like removing stuff.
Paragraph 391
I have discussed with Trump the idea of a government deficiency commission.
Paragraph 392
Nice, yeah.
Paragraph 393
And I would be willing to be part of that commission.
Paragraph 394
I wonder how hard that is.
Paragraph 395
The antibody reaction would be very strong.
Paragraph 396
[Lex] Yeah.
Paragraph 397
So you really have to, you're attacking the matrix at that point. Matrix will fight back.
Paragraph 398
How are you doing with that, being attacked?
Paragraph 399
Me, attacked?
Paragraph 400
Yeah. There's a lot of it.
Paragraph 401
Yeah, there is a lot. I mean, every day, I know psyop. (laughs) Where's my tinfoil hat?
Paragraph 402
How do you keep your just positivity, optimism about the world, a clarity of thinking about the world, so just not become resentful or cynical or all that kind of stuff? Just getting attacked by a very large number of people, misrepresented.
Paragraph 403
Oh yeah, that's a daily occurrence.
Paragraph 404
Yes.
Paragraph 405
I mean, it does get me down at times. I mean, it makes me sad, but, I mean, at some point, you have to sort of say, "Look, the attacks are by people that actually don't know me. And they're trying to generate clicks."
Paragraph 406
So if you can sort of detach yourself somewhat emotionally, which is not easy and say, "Okay, look, this is not actually from someone that knows me or they're literally just writing to get impressions and clicks, then I guess it doesn't hurt as much." It's not quite water off a duck's back. Maybe it's like acid off a duck's back. (laughs) - All right, well, that's good. Just about your own life, what do you as a measure of success in your life?
Paragraph 407
Measure of success, I'd say like, how many useful things can I get done?
Paragraph 408
Day-to-day basis, wake up in the morning, how can I be useful today?
Paragraph 409
Yeah. Maximize utility area out of the code of usefulness. Very difficult to be useful at scale.
Paragraph 410
At scale. Can you like speak to what it takes to be useful for somebody like you, where there's so many amazing great teams? Like how do you allocate your time to be in the most useful?
Paragraph 411
Well, time is the true currency.
Paragraph 412
Yeah.
Paragraph 413
So it is tough to say what is the best allocation of time. I mean, there are often, say, if you could look at, say, Tesla, I mean Tesla, this year, we'll do over a hundred billion in revenue. So that's $2 billion a week. If I make slightly better decisions, I can affect the outcome by a billion dollars. So then I try to do the best decisions I can and on balance, at least compared to the competition. Pretty good decisions. But the marginal value of a better decision can easily be in the course of an hour, a hundred million dollars.
Paragraph 414
Given that, how do you take risks? How do you do the algorithm that you mentioned? I mean, deleting, given a small thing, can be a billion dollars. How do you decide to- - Yeah. Well, I think you have to look at it on a percentage basis because if you look at it in absolute terms, it's just, I would never get any sleep. It would just be like I need to just keep working and work my brain harder.
Paragraph 415
And I'm not trying to get as much as possible out of this meat computer. So it's pretty hard, 'cause you can just work all the time. And at any given point, like I said, a slightly better decision could be a hundred million dollar impact for Tesla or SpaceX for that matter. But it is wild when considering the marginal value of time can be a hundred million dollars an hour at times or more.
Paragraph 416
Is your own happiness part of that equation of success?
Paragraph 417
It has to be, to some degree. If I'm sad, if I'm depressed, I make worse decisions. So I can't have, like if I have zero recreational time, then I make worse decisions. So I don't know a lot, but it's above zero. I mean, my motivation, if I've got a religion of any kind is a religion of curiosity, of trying to understand. It's really the mission of Grok - understand the universe.
Paragraph 418
I'm trying to understand the universe, or let's at least set things in motion such that at some point, civilization understands the universe far better than we do today. And even what questions to ask. As Douglas Adams pointed out in his book, sometimes, the answer is arguably the easy part. Trying to frame the question correctly is the hard part. Once you frame the question correctly, the answer is often easy.
Paragraph 419
So I'm trying to set things in motion such that we are at least at some point able to understand the universe. So for SpaceX, the goal is to make life multi-planetary. And which is if you go to the foamy paradox of where are the aliens, you've got these sort of great filters. It's just like, why have we not heard from the aliens? Now lot of people think there are aliens among us.
Paragraph 420
I often claim to be one, which nobody believes me, but I did say alien registration card at one point on my immigration documents. So I've not seen any evidence of aliens. So it suggests that at least one of the explanations is that intelligent life is extremely rare. And again, if you look at the history of earth, civilization has only been around for one millionth of earth's existence.
Paragraph 421
So if aliens had visited here, say, a hundred thousand years ago, they would be like, "Well, they don't even have writing." Just hunter-gatherers, basically. So how long does a civilization last? So for SpaceX, the goal is to establish a self-sustaining city on Mars. Mars is the only viable planet for such a thing. The moon is close, but it lacks resources, and I think it's probably vulnerable to any calamity that takes out earth.
Paragraph 422
The moon is too close. It's vulnerable to a calamity that takes out earth. So I'm not saying we shouldn't have a moon base, but Mars would be far more resilient. The difficulty of getting to Mars is what makes it resilient. In going through these various explanations of why don't we see the aliens, one of them is that they failed to pass these great filters, these key hurdles. And one of those hurdles is being a multi-planet species.
Paragraph 423
So if you're a multi-planet species, then if something were to happen, whether that was a natural catastrophe or a manmade catastrophe, at least the other planet would probably still be around. You don't have all the eggs in one basket. And once you are sort of a two-planet species, you can obviously extend, to extend life halves to the asteroid belt, to maybe the moons of Jupiter and Saturn, and ultimately, to other star systems.
Paragraph 424
But if you can't even get to another planet, definitely not getting to star systems.
Paragraph 425
And the other possible great filters, super powerful technology like AGI, for example. So you're basically trying to knock out one great filter at a time.
Paragraph 426
Digital super intelligence is possibly a great filter. I hope it isn't, but it might be. Guys like say Geoff Hinton would say, he invented a number of the key principles in artificial intelligence. I think he puts the probability of AI annihilation around 10 to 20%, something like that. It's not like, you know, look on the right side. It's 80% likely to be great. (laughs) But I think AI risk mitigation is important.
Paragraph 427
Being a multi-planet species would be a massive risk mitigation. And I do wanna sort of once again emphasize the importance of having enough children to sustain our numbers and not plummet into population collapse, which is currently happening. Population collapse is a real and current thing. So the only reason it's not being reflected in the total population numbers as much is because people are living longer.
Paragraph 428
It's easy to predict, say, what the population of any given country will be. You just take the birth rate last year, how many babies were born, multiply that by life expectancy, and that's what the population will be a steady state unless if the birth rate continues at that level. But if it keeps declining, it will be even less and eventually dwindle to nothing.
Paragraph 429
So I keep banging on the baby drum here for a reason, because it has been the source of civilizational collapse over and over again throughout history. And so why don't we just not try to stable for that day?
Paragraph 430
Well, in that way, I have miserably failed civilization, and I'm trying, hoping to fix that. I would love to have many kids.
Paragraph 431
Great, hope you do. No time like the present.
Paragraph 432
(laughs) Yeah. I gotta allocate more compute to the whole process. But apparently, it's not that difficult.
Paragraph 433
No, it's like unskilled labor.
Paragraph 434
Well, one of the things you do for me, for the world is to inspire us with what the future could be. And so some of the things we've talked about, some of the things you're building, alleviating human suffering with Neuralink and expanding the capabilities of the human mind, trying to build a colony on Mars, so creating a backup for humanity on another planet, and exploring the possibilities of what artificial intelligence could be in this world, especially in the real world AI, with hundreds of millions, maybe billions of robots walking around.
Paragraph 435
There will be billions of robots. That's seems almost, that seems virtual certainty.
Paragraph 436
Well, thank you for building the future, and thank you for inspiring so many of us to keep building and creating cool stuff, including kids.
Paragraph 437
You're welcome. (laughs) Go forth and multiply.
Paragraph 438
Go forth and multiply. Thank you, Elon. Thanks for talking about it. Thanks for listening to this conversation with Elon Musk. And now, dear friends, here's DJ Seo, the co-founder, president, and COO of Neuralink. When did you first become fascinated by the human brain?
Paragraph 439
For me, I was always interested in understanding the purpose of things and how it was engineered to serve that purpose, whether it's organic or inorganic, like we were talking earlier about your curtain holders. They serve a clear purpose and they were engineered with that purpose in mind.
Paragraph 440
And growing up, I had a lot of interest in seeing things, touching things, feeling things, and trying to really understand the root of how it was designed to serve that purpose. And obviously, brain is just a fascinating organ that we all carry. It's infinitely powerful machine that has intelligence and cognition that arise from it. And we haven't even scratched the surface in terms of how all of that occurs.
Paragraph 441
But also at the same time, I think it took me a while to make that connection to really studying and building tech to understand the brain. Not until graduate school. There were a couple moments, key moments in my life where some of those I think influenced how the trajectory of my life got me to studying what I'm doing right now. One was growing up both sides of my family.
Paragraph 442
My grandparents had a very severe form of Alzheimer, and it's incredibly debilitating conditions. I mean, literally, you're seeing someone's whole identity and their mind just losing over time. And I just remember thinking how both the power of the mind, but also how something like that could really lose your sense of identity.
Paragraph 443
It's fascinating that that is one of the ways to reveal the power of a thing by watching it lose the power.
Paragraph 444
Yeah, a lot of what we know about the brain actually comes from these cases where there are trauma to the brain or some parts of the brain that led someone to lose certain abilities. And as a result, there's some correlation and understanding of that part of the tissue being critical for that function. And it's an incredibly fragile organ, if you think about it that way. But also, it's incredibly plastic and incredibly resilient in many different ways.
Paragraph 445
And by the way, the term plastic, as we'll use a bunch, means that it's adaptable. So neuroplasticity refers to the adaptability of the human brain.
Paragraph 446
Correct. Another key moment that sort of influenced how the trajectory of my life have shaped towards the current focus of my life has been during my teenage year when I came to the U. S. I didn't speak a word of English.
Paragraph 447
There was a huge language barrier, and there was a lot of struggle to kind of connect with my peers around me, because I didn't understand the artificial construct that we have created called language, specifically English in this case. And I remember feeling pretty isolated, not being able to connect with peers around me.
Paragraph 448
So spent a lot of time just on my own, reading books, watching movies, and I naturally sort of gravitated towards sci-fi books. I just found them really, really interesting. And also, it was a great way for me to learn English. Some of the first set of books that I picked up are "Ender's Game," the whole saga by Orson Scott card, and "Neuromancer" from William Gibson, and "Snow Crash" from Neal Stephenson.
Paragraph 449
And movies like "Matrix" was coming out around that time point that really influenced how I think about the potential impact that technology can have for our lives in general. So fast track to my college years, I was always fascinated by just physical stuff, building physical stuff, and especially physical things that had some sort of intelligence.
Paragraph 450
And I studied electrical engineering during undergrad, and I started out my research in MEMS, so micro-electro-mechanical systems, and really building these tiny nanostructures for temperature sensing. And I just found that to be just incredibly rewarding and fascinating subject to just understand how you can build something miniature like that that again served a function and had a purpose.
Paragraph 451
And then I spent large majority of my college years basically building millimeter wave circuits for next gen telecommunication systems, for imaging. And it was just something that I found very, very intellectually interesting. Phase arrays, how the signal processing works for any modern as well as next gen telecommunication system, wireless and wireline. EM waves or electromagnetic waves are fascinating.
Paragraph 452
How do you design antennas that are most efficient in a small footprint that you have? How do you make these things energy-efficient? That was something that just consumed my intellectual curiosity.
Paragraph 453
And that journey led me to actually apply to and find myself at a PhD program at UC Berkeley at kind of this consortium called the Berkeley Wireless Research Center that was precisely looking at building, at the time, we called it xg, similar to 3G, 4G, 5G, but the next, next generation G system, and how you would design circuits around that to ultimately go on phones and basically any other devices that are wirelessly connected these days.
Paragraph 454
So I was just absolutely just fascinated by how that entire system works and that infrastructure works. And then also during grad school, I had sort of the fortune of having couple research fellowships that led me to pursue whatever project that I want.
Paragraph 455
And that's one of the things that I really enjoyed about my graduate school career, where you got to kind of pursue your intellectual curiosity and the domain that may not matter at the end of the day, but it's something that really allows you the opportunity to go as deeply as you want, as well as as widely as you want.
Paragraph 456
And at the time, I was actually working on this project called the Smart Bandaid, and the idea was that when you get a wound, there's a lot of other kind of proliferation of signaling pathway that cells follow to close that wound. And there were hypotheses that when you apply external electric field, you can actually accelerate the closing of that field by having basically electro taxing of the cells around that wound site.
Paragraph 457
And specifically, not just for normal wound, there are chronic wounds that don't heal. So we were interested in building some sort of wearable patch that you could apply to kind of facilitate that healing process. And that was in collaboration with Professor Michel Maharbiz, which was a great addition to kind of my thesis committee and it really shaped the rest of my PhD career.
Paragraph 458
So this would be the first time you interacted with biology, I suppose.
Paragraph 459
Correct, correct. I mean, there were some peripheral end application of the wireless imaging and telecommunication system that I was using for security and bio imaging, but this was a very clear direct application to biology and biological system and understanding the constraints around that and really designing and engineering electrical solutions around it.
Paragraph 460
So that was my first introduction, and that's also kind of how I got introduced to Michel. He's sort of known for remote control of beetles in the early 2000s. And then around 2013, obviously kind of the holy grail when it comes to implantable system is to kind of understand how small of a thing you can make, and a lot of that is driven by how much energy or how much power you can supply to it and how you extract data from it.
Paragraph 461
So at the time at Berkeley, there was kind of this desire to kind of understand in the neural space what sort of system you can build to really miniaturize these implantable systems. And I distinctively remember this one particular meeting where Michel came in and he's like, "Guys, I think I have a solution." The solution is ultrasound.
Paragraph 462
And then he proceeded to kind of walk through why that is the case, and that really formed the basis for my thesis work called neural dust system that was looking at ways to use ultrasound as opposed to electromagnetic waves for powering as well as communication.
Paragraph 463
I guess I should step back and say the initial goal of the project was to build these tiny, about a size of a neuron implantable system that can be parked next to a neuron, being able to record its state and being able to ping that back to the outside world for doing something useful. And as I mentioned, the size of the implantable system is limited by how you power the thing and get the data off of it.
Paragraph 464
And at the end of the day, fundamentally, if you look at a human body, we're essentially a bag of salt water, with some interesting proteins and chemicals, but it's mostly salt water that's very, very well temperature-regulated at 37 degrees Celsius.
Paragraph 465
And we'll get into how, and later, why that's an extremely harsh environment for any electronics to survive, as I'm sure you've experienced or maybe not experienced dropping cell phone in a salt water in an ocean. It will instantly kill the device, right? But anyways, just in general, electromagnetic waves don't penetrate through this environment well. And just the speed of light, it is what it is. We can't change it.
Paragraph 466
And based on the wavelength at which you are interfacing with the device, the device just needs to be big. Like these inductors needs to be quite big. And the general good rule of thumb is that you want the wavefront to be roughly on the order of the size of the thing that you're interfacing with.
Paragraph 467
So an implantable system that is around 10 to 100 micron in dimension in a volume, which is about the size of a neuron that you see in a human body, you would have to operate at like hundreds of gigahertz, which number one, not only is it difficult to build electronics operating at those frequencies, but also, the body just attenuates that very, very significantly.
Paragraph 468
So the interesting kind of insight of this ultrasound was the fact that ultrasound just travels a lot more effectively in the human body tissue compared to electromagnetic waves. And this is something that you encounter, and I'm sure most people have encountered in their lives when you go to hospitals that are medical ultrasound sonograph, right? And they go into very, very deep depth without attenuating too much of the signal.
Paragraph 469
So all in all, ultrasound, the fact that it travels through the body extremely well and the mechanism to which it travels to the body really well is that just the wavefront is very different. Its electromagnetic waves are transverse, whereas ultrasound waves are compressive.
Paragraph 470
So it's just a completely different mode of wavefront propagation, and as well as speed of sound is orders and orders of magnitude less than speed of light, which means that even at 10 megahertz ultrasound wave, your wavefront ultimately is a very, very small wavelength. So if you're talking about interfacing with the 10 micron or 100 micron type structure, you would have 150 micron wavefront at 10 megahertz.
Paragraph 471
And building electronics at those frequencies are much, much, much easier and they're a lot more efficient. So the basic idea kind of was born out of using ultrasound as a mechanism for powering the device, and then also getting data back. So now the question is, how do you get the data back? The mechanism to which we landed on is what's called backscattering.
Paragraph 472
This is actually something that is very common and that we interface on a day-to-day basis with our RFID cards, our radio frequency ID tag, where there's actually rarely, in your ID, a battery inside.
Paragraph 473
There's an antenna and there's some sort of coil that has your serial identification ID and then there's an external device called a reader that then sends a wavefront, and then you reflect back that wavefront with some sort of modulation that's unique to your ID. That's what's called backscattering, fundamentally. So the tag itself actually doesn't have to consume that much energy.
Paragraph 474
And that was a mechanism to which we were kind of thinking about sending the data back. So when you have an external ultrasonic transducer that's sending ultrasonic wave to your implant, the neuro dust implant, and it records some information about its environment, whether it's a neuron firing or some other state of the tissue that it's interfacing with, and then it just amplitude modulates the wavefront that comes back to the source.
Paragraph 475
And the recording step would be the only one that requires any energy? So what would require energy in that low step?
Paragraph 476
Correct, so it is that initial kind of startup circuitry to get that recording, amplifying it, and then just modulating. And the mechanism that you can enable that is there is this specialized crystal called piezoelectric crystals that are able to convert sound energy into electrical energy and vice versa. So you can kind of have this interplay between the ultrasonic domain and electrical domain that is the biological tissue.
Paragraph 477
So on the theme of parking very small computational devices next to neurons, that's the dream, the vision of brain computer interfaces. Maybe before we talk about Neuralink, can you give a sense of the history of the field of BCI? What has been maybe the continued dream, and also some of the milestones along the way with the different approaches and the amazing work done at the various labs?
Paragraph 478
I think a good starting point is going back to 1790s. (Lex laughs) - I did not expect that.
Paragraph 479
Where the concept of animal electricity or the fact that body is electric was first discovered by Luigi Galvani, where he had this famous experiment where he connected set of electrodes to frog leg and ran current through it, and then it started twitching, and he said, "Oh my goodness, the body's electric."
Paragraph 480
So fast forward many, many years to 1920s where Hans Berger, who's German psychiatrist discovered EEG or electroencephalography, which is still around. There are these electrode arrays that you wear outside the skull that gives you some sort of neural recording. That was a very, very big milestone that you can record some sort of activities about the human mind.
Paragraph 481
And then in the 1940s, there were these group of scientists, Renshaw, Forbes, and Morrison that inserted these glass micro electrodes into the cortex and recorded single neurons. The fact that there's signal that are a bit more high resolution and high fidelity as you get closer to the source, let's say.
Paragraph 482
And in the 1950s, these two scientists, Hodgkin and Huxley showed up, and they built this beautiful, beautiful models of the cell membrane and the ionic mechanism and had these like circuit diagram. And as someone who's an electric engineer, it's a beautiful model that's built out of these partial differential equations, talking about flow of ions, and how that really leads to how neurons communicate.
Paragraph 483
And they won the Nobel Prize for that 10 years later in the 1960s. So in 1969, Eb Fetz from University of Washington, published this beautiful paper called Operating Conditioning of Cortical Unit Activity, where he was able to record a single unit neuron from a monkey and was able to have the monkey modulated based on its activity and reward system.
Paragraph 484
So I would say this is the very, very first example, as far as I'm aware, of as closed loop brain computer interface or BCI.
Paragraph 485
The abstract reads, "The activity of single neurons in precentral cortex of anesthetized monkeys was conditioned by reinforcing high rates of neuronal discharge with delivery of a food pellet. Auditory and visual feedback of unit firing rates was usually provided in addition to food reinforcement." Cool, so they actually got it done.
Paragraph 486
They got it done. This is back in 1969.
Paragraph 487
"After several training sessions, monkeys could increase the activity of newly isolated cells by 50 to 500% above rates before reinforcement." Fascinating.
Paragraph 488
Brain is very plastic. (Lex laughs) - And so from here, the number of experiments grew.
Paragraph 489
Yeah, number of experiments as well as set of tools to interface with the brain have just exploded. I think, and also, just understanding the neural code and how some of the cortical layers and the functions are organized. So the other paper that is pretty seminal, especially in the the motor decoding was this paper in the 1980s from Georgopoulos that discovered that there's this thing called motor tuning curve. So what are motor tuning curves?
Paragraph 490
It's the fact that there are neurons in the motor cortex of mammals, including humans, that have a preferential direction that causes them to fire. So what that means is there are a set of neurons that would increase their spiking activities when you're thinking about moving to the left, right, up, down, and any of those vectors.
Paragraph 491
And based on that, you could start to think, well, if you can't identify those essential eigenvectors, you can do a lot, and you can actually use that information for actually decoding someone's intended movement from the cortex. So that was a very, very seminal kind of paper that showed that there is some sort of code you can extract, especially in the motor cortex.
Paragraph 492
So there's signal there. And if you measure the electrical signal from the brain, that you could actually figure out what the intention was.
Paragraph 493
Correct, yeah, not only electrical signals, but electrical signals from the right set of neurons that give you these preferential direction.
Paragraph 494
Hmm. Okay, so going slowly towards Neuralink, one interesting question is what do we understand on the BCI front on invasive versus non-invasive? From this line of work, how important is it to park next to the neuron? What does that get you?
Paragraph 495
That answer fundamentally depends on what you want to do with it, right? There's actually incredible amount of stuff that you can do with EEG and electrocardiograph, ECoG, which actually doesn't penetrate the cortical layer or parenchyma, but you place a set of electrodes on the surface of the brain.
Paragraph 496
So the thing that I'm personally very interested in is just actually understanding and being able to just really tap into the high resolution, high fidelity understanding of the activities that are happening at the local level. And we can get into biophysics, but just to kind of step back to kind of use analogy, 'cause analogy here can be useful, and sometimes, it's a little bit difficult to think about electricity.
Paragraph 497
At the end of the day, we're doing electrical recording that's mediated by ionic currents, movements of these charged particles, which is really, really hard for most people to think about. But turns out, a lot of the activities that are happening in the brain and the frequency bandwidth, which starts happening is actually very, very similar to sound waves, and in our normal conversation, audible range.
Paragraph 498
So the analogy that typically is used in the field is, if you have a football stadium, there's game going on. If you stand outside the stadium, you maybe get a sense of how the game is going based on the cheers and the booze of the home crowd, whether the team is winning or not. But you have absolutely no idea what the score is.
Paragraph 499
You have absolutely no idea what individual audience or the players are talking or saying to each other, what the next play is, what the next goal is. So what you have to do is you have to drop the microphone near into the stadium and then get near the source, like into the individual chatter. In this specific example, you would wanna have it right next to where the huddle's happening.
Paragraph 500
So I think that's kind of a good illustration of what we're trying to do when we say invasive or minimally invasive or implanted brain computer interfaces versus non-invasive or non-implanted brain interfaces. It's basically talking about where do you put that microphone, and what can you do with that information.
Paragraph 501
So what is the biophysics of the read and write communication that we're talking about here, as we now step into the efforts at Neuralink?
Paragraph 502
Yeah, so brain is made up of these specialized cells called neurons. There's billions of them, tens of billions. Sometimes, people call it a hundred billion that are connected in this complex yet dynamic network that are constantly remodeling. They're changing their synaptic weights, and that's what we typically call neuroplasticity.
Paragraph 503
And the neurons are also bathed in this charged environment that is latent with many charged molecules, like potassium ions, sodium ions, chlorine ions. And those actually facilitate these through ionic current, communication between these different networks.
Paragraph 504
And when you look at a neuron as well, they have these membrane with a beautiful, beautiful protein structure called the voltage selective ion channels, which, in my opinion, is one of nature's best inventions. In many ways, if you think about what they are, they're doing the job of a modern day transistors. Transistors are nothing more, at the end of the day, than a voltage-gated conduction channel.
Paragraph 505
And nature found a way to have that very, very early on in its evolution. And as we all know, with the transistor, you can have many, many computation and a lot of amazing things that we have access to today. So I think it's one of those, just as a tangent, just a beautiful, beautiful invention that the nature came up with, these voltage-gated ion channels.
Paragraph 506
I mean, I suppose there's, on the biological level, every level of the complexity of the hierarchy of the organism, there's going to be some mechanisms for storing information and for doing computation. And this is just one such way. But to do that with biological and chemical components is interesting. Plus like when neurons, I mean, it's not just electricity, it's chemical communication, it's also mechanical.
Paragraph 507
I mean, these are like actual objects that vibrate. I mean, they move- - Yeah, they're actually, I mean, there's a lot of really, really interesting physics that are involved in, you know, kind of going back to my work on ultrasound during grad school, there are groups, and there were groups, and there are still groups looking at ways to cause neurons to actually fire an action potential using ultrasound wave.
Paragraph 508
And the mechanism to which that's happening is still unclear as I understand. It may just be that you're imparting some sort of thermal energy and that causes cells to depolarize in some interesting ways. But there are also these ion channels or even membranes that actually just open up its pore as there are being mechanically shook, right? Vibrated.
Paragraph 509
So there's just a lot of elements of these like move particles, which again, like that's governed by diffusion physics, right? Movements of particles. And there's also a lot of kind of interesting physics there.
Paragraph 510
Also, not to mention, as Roger Penrose talks about, there might be some beautiful weirdness in the quantum mechanical effects of all of this. And he actually believes that consciousness might emerge from the quantum mechanical effects there. So like there's physics, there's chemistry, there's biology, all of that is going on there.
Paragraph 511
Oh yeah, yeah. I mean, you can, yes, there's a lot of levels of physics that you can dive into. But yeah, in the end, you have these membranes with these voltage-gated ion channels that selectively let these charge molecules that are in the extracellular matrix like in and out. And these neurons generally have these like resting potential where there's a voltage difference between inside the cell and outside the cell.
Paragraph 512
And when there's some sort of stimuli that changes the state such that they need to send information to the downstream network, you start to kind of see these like sort of orchestration of these different molecules going in and out of these channels. They also open up, like more of them open up once it reaches some threshold to a point where you have a depolarizing cell that sends action potential.
Paragraph 513
So it's a just a very beautiful kind of orchestration of these molecules. And what we're trying to do when we place an electrode or parking it next to a neuron is that you're trying to measure these local changes in the potential. Again, mediated by the movements of the ions. And what's interesting, as I mentioned earlier, there's a lot of physics involved.
Paragraph 514
And the two dominant physics for this electrical recording domain is diffusion physics and electromagnetism. And where one dominates, where Maxwell's equation dominates versus fixed law dominates depends on where your electrode is. If it's close to the source, mostly electromagnetic-based, when you're farther away from it, it's more diffusion-based.
Paragraph 515
So essentially, when you're able to park it next to it, you can listen in on those individual chatter and those local changes in the potential, and the type of signal that you get are these canonical textbook neural spiking waveform.
Paragraph 516
The moment you're further away, and based on some of the studies that people have done, Christof Koch's lab and others, once you're away from that source by roughly around a hundred micron, which is about a width of a human hair, you no longer hear from that neuron. Or you're no longer able to kind of have the system sensitive enough to be able to record that particular local membrane potential change in that neuron.
Paragraph 517
And just to kind of give you a sense of scale also, when you look at a hundred micron voxel, so a hundred micron by a hundred micron by a hundred micron box in a brain tissue, there's roughly around 40 neurons and whatever number of connections that they have. So there's a lot in that volume of tissue.
Paragraph 518
So the moment you're outside of that, there's just no hope that you'll be able to detect that change from that one specific neuron that you may care about.
Paragraph 519
Yeah, but as you're moving about this space, you'll be hearing other ones. So if you move another 100 micron, you'll be hearing chatter from another community.
Paragraph 520
Correct.
Paragraph 521
And so the whole sense is you wanna place as many as possible electrodes and then you're listening to the chatter.
Paragraph 522
Yeah, you wanna listen to the chatter. And at the end of the day, you also want to basically let the software do the job of decoding. And just to kind of go to, why ECOG and EEG work at all, right? When you have these local changes, obviously, it's not just this one neuron that's activating. There's many, many other networks that are activating all the time.
Paragraph 523
And you do see sort of a general change in the potential of this electro, like this charge medium, and that's what you're recording when you're farther away. I mean, you still have some reference electrode that's stable in the brain that's just electroactive organ, and you're seeing some combination aggregate action potential changes and then you can pick it up, right? It's a much slower changing signals.
Paragraph 524
But there are these like canonical kind of oscillations and waves, like gamma waves, beta waves. Like when you sleep, that can be detected, 'cause there's sort of a synchronized kind of global effect of the brain that you can detect.
Paragraph 525
And I mean, the physics of this go, I mean, if we really wanna go down that rabbit hole, like there's a lot that goes on in terms of like why diffusion physics at some point dominates when you're further away from the source. It's just a charged medium.
Paragraph 526
So similar to how when you have electromagnetic ways propagating in atmosphere or in a charged medium like a plasma, there's this weird shielding that happens that actually further attenuates the signal as you move away from it. So yeah, you see, like if you do a really, really deep dive on kind of the signal attenuation over distance, you start to see kind of one of where square in the beginning, and then exponential drop off.
Paragraph 527
And that's the knee at which you go from electromagnet magnetism dominating to diffusion physics dominating.
Paragraph 528
But once again, with the electrodes, the biophysics, you need to understand it's not as deep, because no matter where you're placing that, you're listening to a small crowd of local neurons.
Paragraph 529
Correct, yeah. So once you penetrate the brain, you're in the arena, so to speak.
Paragraph 530
And there's a lot of neurons.
Paragraph 531
[DJ] There are many, many of 'em.
Paragraph 532
But then again, there's a whole field of neuroscience that's studying like how the different groupings, the different sections of the seating in the arena, what they usually are responsible for, which is where the metaphor probably falls apart, 'cause the seating is not that organized in an arena.
Paragraph 533
Also, most of them are silent. They don't really do much, or their activities are, you know, you have to hit it with just the right set of stimulus.
Paragraph 534
So they're usually quiet.
Paragraph 535
They're usually very quiet. There's, I mean, similar to dark energy and dark matter, there's dark neurons. What are they all doing? When you place these electrode, again, like within this a hundred micron volume, you have 40 or so neurons. Like why do you not see 40 neurons? Why do you see only a handful? What is happening there?
Paragraph 536
Well, they're mostly quiet, but like when they speak, they say profound shit, I think. That's the way I'd like to think about it. Anyway, before we zoom in even more, let's zoom out. So how does Neuralink work? From the surgery to the implant to the signal and the decoding process and the human being able to use the implant to actually affect the world outside?
Paragraph 537
And all of this, I'm asking in the context of there's a gigantic historic milestone in Neuralink just accomplished in January of this year, putting a Neuralink implant in the first human being, Noland. And there's been a lot to talk about there about his experience, because he's able to describe all the nuance and the beauty and the fascinating complexity of that experience of everything involved.
Paragraph 538
But on the technical level, how does Neuralink work?
Paragraph 539
Yeah, so there are three major components to the technology that we're building. One is the device, the thing that's actually recording these neural chatters. We call it N1 implant or The Link. And we have a surgical robot that's actually doing an implantation of these tiny, tiny wires that we call threads that are smaller than human hair.
Paragraph 540
And once everything is surgirized, you have these neural signals, these spiking neurons that are coming out of the brain and you need to have some sort of software to decode what the users intend to do with that.
Paragraph 541
So there's what's called the Neuralink application, or B1 app that's doing that translation, is running the very, very simple machine learning model that decodes these inputs that are neural signals and then convert it to a set of outputs that allows our participant, first participant Noland, to be able to control a cursor on this.
Paragraph 542
And this is done wirelessly?
Paragraph 543
And this is done wirelessly. So our implant is actually a two-part. The link has these flexible tiny wires called threads that have multiple electrodes along its length. And they're only inserted into the cortical layer, which is about three to five millimeters in a human brain in the motor cortex region. That's where the kind of the intention for movement lies in.
Paragraph 544
And we have 64 of these threads, each thread having 16 electrodes along the span of three to four millimeters, separated by 200 microns. So you can actually record along the depth of the insertion.
Paragraph 545
And based on that signal, there's custom integrated circuit or ASIC that we built that amplifies the neural signals that you're recording and then digitizing it and then has some mechanism for detecting whether there was an interesting event that is a spiking event and decide to send that, or not send that through Bluetooth to an external device, whether it's a phone or a computer that's running this Neuralink application.
Paragraph 546
So there's onboard signal processing already just to decide whether this is an interesting event or not. So there is some computational power on board inside in addition to the human brain?
Paragraph 547
Yeah, so it does the signal processing to kind of really compress the amount of signal that you're recording. So we have a total of a thousand electrodes sampling at just under 20 kilohertz with 10 bit each. So that's 200 megabits. That's coming through to the chip, from thousand channel simultaneous neural recording. And that's quite a bit of data.
Paragraph 548
And there are technology available to send that off wirelessly, but being able to do that in a very, very thermally constrained environment that is a brain, so there has to be some amount of compression that happens to send off only the interesting data that you need, which in this particular case, for motor decoding is occurrence of a spike or not. And then being able to use that to decode the intended cursor movement.
Paragraph 549
So the implant itself processes it, figures out whether a spike happened or not with our spike detection algorithm, and then sends it off, packages it, sends it off through Bluetooth to an external device that then has the model to decode, okay, based on the spiking inputs, did Noland wish to go up, down, left, right, or click, or right click, or whatever?
Paragraph 550
All of this is really fascinating, but let's stick on the N1 implant itself, so the thing that's in the brain. So I'm looking at a picture of it, there's an enclosure, there's a charging call, so we didn't talk about the charging, which is fascinating. The battery, the power electronics, the antenna. Then there's the signal processing electronics. I wonder if there's more kinds of signal processing you can do. That's another question. And then there's the threads themselves with the enclosure on the bottom. So maybe to ask about the charging, so there's a external charging device.
Paragraph 551
Mm-hmm, yeah, there's an external charging device. So yeah, the second part of the implant, the threads are the ones, again, just the last three to five millimeters are the ones that are actually penetrating the cortex. Rest of it is, actually, most of the volume is occupied by the battery, rechargeable battery. And it's about a size of a quarter. I actually have a device here, if you wanna take a look at it.
Paragraph 552
This is the flexible threat component of it. And then this is the implant. So it's about a size of a U. S. quarter. It's about nine millimeter thick. So basically, this implant, once you have the craniectomy and the directomy, threads are inserted, and the hole that you created, this craniectomy, gets replaced with that. So basically, that thing plugs that hole, and you can screw in these self-drilling cranial screws to hold it in place.
Paragraph 553
And at the end of the day, once you have the skin flap over, there's only about two to three millimeters. That's obviously transitioning off of the top of the implant to where the screws are. And that's the minor bump that you have.
Paragraph 554
Those threads look tiny. That's incredible. That is really incredible. And also, you're right, most of the actual volume is the battery. Yeah, this is way smaller than I realized.
Paragraph 555
They are also, the threads themselves are quite strong.
Paragraph 556
They look strong.
Paragraph 557
And the thread themselves also has a very interesting feature at the end of it called the loop. And that's the mechanism to which the robot is able to interface and manipulate this tiny hair-like structure.
Paragraph 558
And they're tiny, so what's the width of a thread?
Paragraph 559
Yeah, so the width of a thread starts from 16 micron and then tapers out to about 84 micron. So average human hair is about 80 to 100 micron in width.
Paragraph 560
This thing is amazing. This thing is amazing.
Paragraph 561
Yes, most of the volume is occupied by the battery, rechargeable lithium ion cell. And the charging is done through inductive charging, which is actually very commonly used. Most cell phones have that. The biggest difference is that for us, usually when you have a phone and you wanna charge it on a charging pad, you don't really care how hot it gets, whereas for us, it matters.
Paragraph 562
There's a very strict regulation and good reasons to not actually increase the surrounding tissue temperature by two degrees Celsius. So there's actually a lot of innovation that is packed into this to allow charging of this implant without causing that temperature threshold to reach. And even small things like, you see this charging coil and what's called the ferrite shield, right?
Paragraph 563
So without that ferrite shield, what you end up having when you have resonant inductive charging is that the battery itself is a metallic can, and you form these Eddy currents from external charger and that causes heating and that actually contributes to inefficiency in charging. So this ferrite shield, what it does is that it actually concentrate that field line away from the battery and then around the coil that's actually wrapped around it.
Paragraph 564
There's a lot of really fascinating design here to make it, I mean, you're integrating a computer into a complex biological system.
Paragraph 565
Yeah, there's a lot of innovation here. I would say that part of what enabled this was just the innovations in the wearable. There's a lot of really, really powerful, tiny, low power microcontrollers, temperature sensors, or various different sensors and power electronics. A lot of innovation really came in the charging coil design, how this is packaged, and how do you enable charging such that you don't really exceed that temperature limit, which is not a constraint for other devices out there.
Paragraph 566
So let's talk about the threads themselves, those tiny, tiny, tiny things. So how many of them are there? You mentioned a thousand electrodes. How many threads are there, and what did the electrodes have to do with the threads?
Paragraph 567
Yeah, so the current instantiation of the device has 64 threads, and each thread has 16 electrodes for a total of 1,024 electrodes that are capable of both recording and stimulating. And the thread is basically this polymer insulated wire. The metal conductor is the kind of tiramisu cake of ti, plat, gold, plate, ti. And they're very, very tiny wires. Two micron in width, so 2/1000000th of meter.
Paragraph 568
It's crazy that that thing I'm looking at has the polymer installation, has the conducting material, and has 16 electrodes at the end of it.
Paragraph 569
On each of those threads.
Paragraph 570
Yeah, on each of those threads.
Paragraph 571
Correct.
Paragraph 572
16, each one of those.
Paragraph 573
You're not gonna be able to see it with naked eyes.
Paragraph 574
And I mean, to state the obvious, or maybe for people who are just listening, they're flexible.
Paragraph 575
Yes, yes, that's also one element that was incredibly important for us. So each of these thread are, as I mentioned, 16 micron in width and then they taper to 84 micron, but in thickness, they're less than five micron. And thickness is mostly polyamide at the bottom and this metal track and then another polyamide. So two micron of polyamide, 400 nanometer of this metal stack, and two micron of polyamide sandwiched together to protect it from the environment that is 37 degrees C bag of salt water.
Paragraph 576
So what's some, maybe can you speak to some interesting aspects of the material design here? Like what does it take to design a thing like this and to be able to manufacture a thing like this for people who don't know anything about this kind of thing?
Paragraph 577
Yeah, so the material selection that we have is not, I don't think it was particularly unique. There were other labs and there are other labs that are kind of looking at similar material stack. There's kind of a fundamental question and still needs to be answered around the longevity and reliability of these micro electrodes that we call, compared to some of the other more conventional, neural interfaces, devices that are intracranial.
Paragraph 578
So penetrating the cortex that are more rigid, like the Utah array. There are these four by four millimeter kind of silicon shank that have exposed recording site at the end of it. And that's been kind of the innovation from Richard Normann back in 1997. It's called the Utah Array 'cause he was at University of Utah.
Paragraph 579
And what does the Utah array look like? So it's a rigid type of- - Yeah, so we can actually look it up.
Paragraph 580
Oh.
Paragraph 581
Yeah. (Lex laughing) Yeah, so it's a bed of needle. There's- - (laughs) Yeah. Okay, go ahead, I'm sorry.
Paragraph 582
So those are rigid- - Rigid, yeah. You weren't kidding.
Paragraph 583
And the size and the number of shanks vary, anywhere from 64 to 128. At the very tip of it is an exposed electrode that actually records neural signal. The other thing that's interesting to note is that unlike Neuralink threads that have recording electrodes that are actually exposed iridium oxide recording sites along the depth, this is only at a single depth. So these Utah array spokes can be anywhere between 0. 5 millimeters to 1.
Paragraph 584
5 millimeter. And they also have designs that are slanted. So you can have it inserted at different depth, but that's one of the other big differences. And then, I mean, the main key difference is the fact that there's no active electronics.
Paragraph 585
These are just electrodes, and then there's a bundle of a wire that you're seeing, and then that actually then exits the craniectomy that then has this port that you can connect to for any external electronic devices. They are working on a or have the wireless telemetry device, but it still requires a through the skin port that actually is one of the biggest failure modes for infection for the system.
Paragraph 586
What are some of the challenges associated with flexible threads? Like for example, on the robotic side, R1, implanting those threads, how difficult does that task?
Paragraph 587
Yeah, so as you mentioned, they're very, very difficult to maneuver by hand. These Utah arrays that you saw earlier, they're actually inserted by a neurosurgeon actually positioning it near the site that they want. And then there's a pneumatic hammer that actually pushes them in. So it's a pretty simple process, and they're easy to maneuver. But for these thin film arrays, they're very, very tiny and flexible.
Paragraph 588
So they're very difficult to maneuver. So that's why we built an entire robot to do that. There are other reasons for why we built a robot, and that is ultimately, we want this to help millions and millions of people that can benefit from this. And there just aren't that many neurosurgeons out there. And robots can be something that we hope can actually do large parts of the surgery.
Paragraph 589
But yeah, the robot is this entire other sort of category of product that we're working on. And it's essentially this multi-axis gantry system that has the specialized robot head that has all of the optics and this kind of a needle retracting mechanism that maneuvers these threads via this loop structure that you have on the thread.
Paragraph 590
So the thread already has a loop structure by which you can grab it?
Paragraph 591
Correct, correct.
Paragraph 592
Okay. So this is fascinating. So you mentioned optics, so there's a robot - R1. So for now, there's a human that actually creates a hole in the skull.
Paragraph 593
Mm-hmm.
Paragraph 594
And then after that, there's a computer vision component that's finding a way to avoid the blood vessels. And then you're grabbing it by the loop, each individual thread and placing it in a particular location to avoid the blood vessels. And also choosing the depth of placement, all that?
Paragraph 595
Correct. So controlling every, like the 3D geometry of the placement?
Paragraph 596
Correct. So the aspect of this robot that is unique is that it's not surgeon-assisted or human-assisted. It's a semi-automatic or automatic robot. Obviously, there are human component to it, when you're placing targets. You can always move it away from kind of major vessels that you see. But I mean, we wanna get to a point where one click and it just does the surgery within minutes.
Paragraph 597
So the computer vision component finds great targets, candidates and the human kind of approves them and the robot does, does it do like one thread at a time or does it do one- - It does one thread at a time, and that's actually also one thing that we are looking at ways to do multiple threads at a time. There's nothing stopping from it. You can have multiple kind of engagement mechanisms, but right now, it's one by one.
Paragraph 598
And we also still do quite a bit of just kind of verification to make sure that it got inserted. If so, how deep? Did it actually match what was programmed in and so on and so forth?
Paragraph 599
And the actual electrode is a place that vary at differing depths in the like, I mean, it's very small differences, but differences.
Paragraph 600
[DJ] Yeah, yeah.
Paragraph 601
And so that there's some reasoning behind that, as you mentioned. Like it gets more varied signal.
Paragraph 602
Yeah, I mean, we try to place them all around three or four millimeter from the surface, just 'cause the span of the electrode, those 16 electrodes that we currently have in this version spans roughly around three millimeters. So we wanna get all of those in the brain.
Paragraph 603
This is fascinating. Okay, so there's a million questions here. If we could zoom in specifically on the electrodes, so what is your sense, how many neurons is each individual electrode listening to?
Paragraph 604
Yeah, each electrode can record from anywhere between 0 to 40, as I mentioned, right, earlier. But tactically speaking, we only see about, at most, like two to three. And you can actually distinguish which neuron it's coming from by the shape of the spikes.
Paragraph 605
[Lex] Oh, cool.
Paragraph 606
So I mentioned the spike detection algorithm that we have. It's called BOSS algorithm, buffer online, spike sorter.
Paragraph 607
Nice.
Paragraph 608
It actually outputs at the end of the day six unique values, which are kind of the amplitude of these like negative going hump, middle hump, like positive going hump, and then also the time at which these happen. And from that, you can have a kind of a statistical probability estimation of, is that a spike? Is it not a spike? And then based on that, you could also determine, oh, that spike looks different than that spike. Must have come from a different neuron.
Paragraph 609
Okay, so that's a nice signal processing step from which you can then make much better predictions about if there's a spike.
Paragraph 610
Yeah.
Paragraph 611
Especially in this kind of context where there could be multiple neurons screaming. And that also results in you being able to compress the data better in the (indistinct). Okay.
Paragraph 612
And just to be clear, I mean, the labs do what's called spike sorting. Usually, once you have these like broadband, the fully digitized signals and then you run a bunch of different set of algorithms to kind of tease apart, it's just all of this for us is done on the device.
Paragraph 613
On the device.
Paragraph 614
In a very low power, custom built ASIC digital processing unit.
Paragraph 615
[Lex] Highly heat constrained?
Paragraph 616
Highly heat constrained, and the processing time from signal going in and giving you the output is less than a microsecond, which is a very, very short amount of time.
Paragraph 617
Oh yeah, so the latency has to be super short.
Paragraph 618
Correct.
Paragraph 619
Oh wow. Oh, that's a pain in the ass.
Paragraph 620
Yeah, latency is this huge, huge thing that you have to deal with. Right now, the biggest source of latency comes from the Bluetooth, the way in which they're packetized and we bend them in 15 millisecond.
Paragraph 621
Oh, interesting, it says communication constraint. Is there some potential innovation there on the protocol used?
Paragraph 622
Absolutely.
Paragraph 623
Okay.
Paragraph 624
Yeah, Bluetooth is definitely not our final wireless communication protocol that we wanna get to.
Paragraph 625
Hence the N1 and the R1. I imagine that increases- - NxRx.
Paragraph 626
Yeah, that's the communication protocol, 'cause Bluetooth allows you to communicate I guess farther distances than you need to, so you can go much shorter.
Paragraph 627
Yeah, well, the primary motivation for choosing Bluetooth is that, I mean, everything has Bluetooth.
Paragraph 628
All right, you can talk to any device.
Paragraph 629
Interoperability is just absolutely essential, especially in this early phase. And in many ways, if you can access a phone or a computer, you can do anything.
Paragraph 630
Well, it'll be interesting to step back and actually look at, again, the same pipeline that you mentioned for Noland. So what does this whole process look like? From finding and selecting a human being to the surgery, to the first time he's able to use this thing?
Paragraph 631
So we have what's called the patient registry that people can sign up to hear more about the updates. And that was a route to which Noland applied. And the process is that once the application comes in, it contains some medical records, and based on their medical eligibility, that there's a lot of different inclusion-exclusion criteria for them to meet. And we go through a prescreening interview process with someone from Neuralink.
Paragraph 632
And at some point, we also go out to their homes to do a BCI home audit, 'cause one of the most kind of revolutionary part about having this N1 system that is completely wireless is that you can use it at home. Like you don't actually have to go to the lab and go to the clinic to get connecterized to these like specialized equipment that you can't take home with you.
Paragraph 633
So that's one of the key elements of when we're designing the system that we wanted to keep in mind, like people hopefully would wanna be able to use this every day in the comfort of their homes. And so part of our engagement and what we're looking for during BCI home audit is to just kind of understand their situation, what other assistive technology that they use.
Paragraph 634
And we should also step back and kind of say that the estimate is 180,000 people live with quadriplegia in the United States, and each year, an additional 18,000 suffer a paralyzing spinal cord injury. So these are folks who have a lot of challenges, living a life in terms of accessibility, in terms of doing the things that many of us just take for granted day to day.
Paragraph 635
And one of the things, one of the goals of this initial study is to enable them to have sort of digital autonomy, where they by themselves can interact with a digital device using just their mind, something that you're calling telepathy. So digital telepathy, where a quadriplegic can communicate with a digital device in all the ways that we've been talking about.
Paragraph 636
Control the mouse cursor, enough to be able to do all kinds of stuff, including play games and tweet and all that kind of stuff. And there's a lot of people for whom life, the basics of life are difficult, because of the things that have happened to them.
Paragraph 637
Yeah, I mean, movement is so fundamental to our existence. I mean, even speaking involves movement of mouth, lip, larynx. And without that, it's extremely debilitating. And there are many, many people that we can help. And I mean, like especially if you start to kind of look at other forms of movement disorders that are not just from spinal cord injury, but from ALS, MS, or even stroke and/or just aging, right? That leads you to lose some of that mobility, that independence, it's extremely debilitating.
Paragraph 638
And all of these are opportunities to help people, to help alleviate suffering, to help improve the quality of life. But each of the things you mentioned is its own little puzzle that needs to have increasing levels of capability from a device like a Neuralink device. And so the first one you're focusing on is, it's just a beautiful word, telepathy. So being able to communicate using your mind wirelessly with a digital device. Can you just explain exactly what we're talking about?
Paragraph 639
Yeah, I mean, it's exactly that. I mean, I think if you are able to control a cursor and able to click and be able to get access to computer or phone, I mean, the whole world opens up to you. And I mean, I guess the word telepathy, if you kind of think about that as just definitionally being able to transfer information from my brain to your brain without using some of the physical faculties that we have, like voices.
Paragraph 640
But the interesting thing here is, I think the thing that's not obviously clear is how exactly it works. So in order to move a cursor, there's at least a couple ways of doing that. So one is you imagine yourself maybe moving a mouse with your hand, or you can then, which Noland talked about, like imagine moving the cursor with your mind.
Paragraph 641
But it's like there is a cognitive step here that's fascinating, 'cause you have to use the brain and you have to learn how to use the brain. And you kind of have to figure it out dynamically. Because you reward yourself if it works. I mean, there's a step that, this is just a fascinating step, 'cause you have to get the brain to start firing in the right way. And you do that by imagining. Like fake it till you make it.
Paragraph 642
(laughs) And all of a sudden, it creates the right kind of signal that if decoded correctly, can create the kind of effect. And then there's like noise around that, you have to figure all of that out. But on the human side, imagine the cursor moving is what you have to do.
Paragraph 643
Yeah, he says using the force.
Paragraph 644
The force. I mean, isn't that just like fascinating to you that it works? Like to me, it's like, holy shit, that actually works. Like you could move a cursor with your mind.
Paragraph 645
As much as you're learning to use that thing, that thing's also learning about you. Like our model is constantly updating the weights to say, "Oh, if someone is thinking about this sophisticated forms of like spiking patterns, like that actually means to do this, right?"
Paragraph 646
So the machine is learning about the human and the human is learning about the machine. So there is a adaptability to the signal processing, the decoding step. And then there's the adaptation of Noland, the human being. Like the same way, if you give me a new mouse and I move it, I learn very quickly about its sensitivity, so I learn to move it slower. And then there's other kinds of signal drift and all that kind of stuff they have to adapt to. So both are adapting to each other.
Paragraph 647
[DJ] Correct.
Paragraph 648
That's a fascinating like software challenge on both sides, the software on both, the human software and- - The organic and the inorganic. Anyway, so sorry to rudely interrupt. So there's this selection that Noland has passed with flying colors. So everything including that the, it's a BCI-friendly home, all of that. So what is the process of the surgery implantation, the first moment when he gets to use the system?
Paragraph 649
The end to end, we say patient end to patient out, is anywhere between two to four hours. In particular case for Noland, it was about three and a half hours. And there's many steps leading to the actual robot insertion, right? So there's anesthesia induction, and we do intra-op CT imaging to make sure that we're drilling the hole in the right location. And this is also pre-planned beforehand.
Paragraph 650
Someone like Nolan would go through fMRI, and then they can think about wiggling their hand. And obviously, due to their injury, it's not gonna actually lead to any sort of intended output. But it's the same part of the brain that actually lights up when you're imagining moving your finger to actually moving your finger.
Paragraph 651
And that's one of the ways in which we can actually know where to place our threads, 'cause we wanna go into what's called a hand knob area in the motor cortex. And as much as possible, densely put our electro threads. So yeah, we do intra-op CT imaging to make sure and double check the location of the craniectomy.
Paragraph 652
And surgeon comes in, does their thing in terms of like skin incision, craniectomy, so drilling of the skull, and then there's many different layers of the brain. There's what's called a dura, which is a very, very thick layer that surrounds the brain, that gets actually resected in a process called atherectomy. And that then exposed the pia in the brain that you wanna insert.
Paragraph 653
And by the time it's been around anywhere between one to one and a half hours, robot comes in, does its thing, placement of the targets, inserting of the thread. That takes anywhere between 20 to 40 minutes. In the particular case for Noland, it was just under or it was just over 30 minutes. And then after that, the surgeon comes in.
Paragraph 654
There's a couple other steps of like actually inserting the dural substitute layer to protect the thread as well as the brain. And then yeah, screw in the implant and then skin flap and then suture and then you're out.
Paragraph 655
So when Noland woke up, what was that like? What's the recovery like, and when was the first time he was able to use it?
Paragraph 656
So he was actually immediately, after the surgery, like an hour after the surgery as he was waking up, we did turn on the device, make sure that we are recording neural signals, and we actually did have a couple signals that we notice that he can actually modulate. And what I mean by modulate is that he can think about crunching his fist, and you could see the spike disappear and appear. (Lex laughing) - That's awesome.
Paragraph 657
And that was immediate, right? Immediate after in the recovery room.
Paragraph 658
How cool is that?
Paragraph 659
Yeah.
Paragraph 660
That's a human being. I mean, what did that feel like for you? This device and a human being, a first step of a gigantic journey? I mean, it's a historic moment. Even just that spike, just to be able to modulate that.
Paragraph 661
Obviously, there have been other, as you mentioned, pioneers that have participated in these groundbreaking BCI investigational early feasibility studies. So we're obviously standing in the shoulders of the giants here. We're not the first ones to actually put electrodes in the human brain. But I mean, just leading up to the surgery, I definitely could not sleep. It's the first time that you're working in a completely new environment.
Paragraph 662
We had a lot of confidence based on our benchtop testing or preclinical R&D studies that the mechanism, the threads, the insertion, all that stuff is very safe, and that it's obviously ready for doing this in a human, but there's still a lot of unknown unknown about can the needle actually insert? I mean, we brought something like 40 needles just in case they break, and we ended up using only one.
Paragraph 663
But I mean, that was a level of just complete unknown, right? 'Cause it's a very, very different environment. And I mean, that's why we do clinical trial in the first place to be able to test these things out. So extreme nervousness and just many, many sleepless night leading up to the surgery and definitely the day before the surgery, and it was an early morning surgery. Like we started at seven in the morning.
Paragraph 664
And by the time, it was around 10:30. Everything was done. But I mean, first time seeing that, well, number one, just huge relief that this thing is doing what it's supposed to do. And two, I mean, just immense amount of gratitude for Noland and his family, and then many others that have applied and that we've spoken to and will speak to are, I mean, true pioneers everywhere. And I sort of call them the neural astronauts or neuralnaut.
Paragraph 665
These amazing, just like in the '60s, right? Like these amazing just pioneers, right? Exploring the unknown outwards. In this case, it's inward. But incredible amount of gratitude for them to just participate and play a part. And it's a journey that we're embarking on together. But also like, I think it was just, that was an very, very important milestone, but our work was just starting.
Paragraph 666
So a lot of just kind of anticipation for, okay, what needs to happen next? What are set of sequences of events that needs to happen for us to make it worthwhile for both Noland as well as us?
Paragraph 667
Just to linger on that, just a huge congratulations to you and the team for that milestone. I know there's a lot of work left, but that's really exciting to see. That's a source of hope. It's this first big step, opportunity to help hundreds of thousands of people and then maybe expand the realm of the possible for the human mind for millions of people in the future. So it's really exciting.
Paragraph 668
So like the opportunities are all ahead of us, and to do that safely and to do that effectively was really fun to see. As an engineer just watching other engineers come together and do an epic thing, that was awesome. Huge congrats.
Paragraph 669
Thank you, thank you. Yeah, could not have done it without the team. And yeah, I mean, that's the other thing that I told the team as well, of just this immense sense of optimism for the future. I mean, it's a very important moment for the company, needless to say, as well as hopefully for many others out there that we can help.
Paragraph 670
So speaking of challenges, Neuralink published a blog post describing that some of the threads retracted. And so the performance, as measured by bits per second dropped at first, but then eventually, it was regained. And that the whole story of how it was regained is super interesting. That's definitely something I'll talk to Bliss and to Noland about. But in general, can you speak to this whole experience? How was the performance regained and just the technical aspects of the threads being retracted and moving?
Paragraph 671
The main takeaway is that in the end, the performance have come back and it's actually gotten better than it was before. He's actually just beat the world record yet again last week to 8.5 BPS, so I mean, he's just cranking and he's just improving.
Paragraph 672
[Lex] The previous one that he set was eight.
Paragraph 673
Correct.
Paragraph 674
He said 8.5.
Paragraph 675
Yeah, the previous world record in human was 4.6.
Paragraph 676
Yeah.
Paragraph 677
So it's almost double. And his goal is to try to get to 10, which is roughly around kind of the median Neuralinker using a mouse with the hand. So it's getting there.
Paragraph 678
So yeah, so the performance was regained.
Paragraph 679
Yeah, better than before. So that's a story on its own of what took the BCI team to recover that performance. It was actually mostly on kind of the signal processing. And so as I mentioned, we were kind of looking at these spike outputs from our electrodes. And what happened is that kind of four weeks into the surgery, we noticed that the threads have solely come out of the brain.
Paragraph 680
And the way in which we noticed this, at first obviously, is that, well, I think Noland was the first to notice that his performance was degrading. And I think at the time, we were also trying to do a bunch of different experimentation, different algorithms, different sort of UI/UX. So it was expected that there will be variability in the performance, but we did see kind of a steady decline.
Paragraph 681
And then also, the way in which we measure the health of the electrodes or whether they're in the brain or not, is by measuring impedance of the electrode. So we look at kind of the interfacial, kind of the Randall circuit, they say, the capacitance and the resistance between the electrosurface and the medium. And if that changes in some dramatic ways, we have some indication.
Paragraph 682
Or if you're not seeing spikes on those channels, you have some indications that something's happening there. And what we notice is that looking at those impedance plot and spike rate plots, and also because we have those electrodes recording along the depth, you are seeing some sort of movement that indicated that the rest were being pulled out.
Paragraph 683
And that obviously will have an implication on the model side, because if the number of inputs that are going into the model is changing, 'cause you have less of them, that model needs to get updated, right? But there were still signals, and as I mentioned, similar to how, even when you place the signals on the surface of the brain, or farther away, like outside the skull, you still see some useful signals.
Paragraph 684
What we started looking at is not just the spike occurrence through this BOSS algorithm that I mentioned, but we started looking at just the power of the frequency band that is interesting for Noland to be able to modulate.
Paragraph 685
So once we kind of change the algorithm for the implant to not just give you the BOSS output, but also these spike band power output, that helped us sort of, we find the model with the new set of inputs, and that was the thing that really ultimately gave us the performance back.
Paragraph 686
In terms of, and obviously, like the thing that we want ultimately, and the thing that we are working towards, is figuring out ways in which we can keep those threads intact for as long as possible so that we have many more channels going into the model. That's by far the number one priority that the team is currently embarking on to understand how to prevent that from happening.
Paragraph 687
The thing that I'll say also is that, as I mentioned, this is the first time ever that we're putting these thread in a human brain, and human brain, just for size reference, is 10 times out of the monkey brain or the sheep brain. And it's just a very, very different environment. It moves a lot more. It like actually moved a lot more than we expected when we did Noland's surgery.
Paragraph 688
And it's just a very, very different environment than what we're used to. And this is why we do clinical trial, right? We wanna uncover some of these issues and failure modes earlier than later. So in many ways, it's provided us with this enormous amount of data and information to be able to solve this. And this is something that Neuralink is extremely good at.
Paragraph 689
Once we have set of clear objective and engineering problem, we have enormous amount of talents across many, many disciplines to be able to come together and fix the problem very, very quickly.
Paragraph 690
But it sounds like one of the fascinating challenges here is for the system and the decoding side to be adaptable across different timescales. So whether it's movement of threads or different aspects of signal drift sort of on the software of the human brain, something changing, like Noland talks about cursor drift that could be corrected, and there's a whole UX challenge to how to do that. So it sounds like adaptability is like a fundamental property that has to be engineered in.
Paragraph 691
It is, and I mean I think, I mean, as a company, we're extremely vertically integrated. We make these thin film arrays in our own microfab.
Paragraph 692
Yeah, there's, like you said, built in-house. This whole paragraph here from this blog post is pretty gangster. "Building the technology described above has been no small feat." And there's a bunch of links here that I recommend people click on. "We constructed in-house micro fabrication capabilities to rapidly produce various iterations of thin film arrays that constitute our electrode threads. We created a custom femtosecond laser mill to manufacture components with micro level precision." I think there's a tweet associated with this.
Paragraph 693
That's a whole thing that we can get into.
Paragraph 694
Yeah, okay, well, what are we looking at here? This thing?
Paragraph 695
Yeah. "So in less than one minute, our custom-made femtosecond laser mill cuts this geometry in the tips of our needles." So we're looking at this weirdly-shaped needle. "The tip is only 10 to 12 microns in width, only slightly larger than the diameter of a red blood cell. The small size allows threats to be inserted with minimal damage to the cortex." Okay, so what's interesting about this geometry? So we're looking at this just geometry of a needle.
Paragraph 696
This is the needle that's engaging with the loops in the thread. So they're the ones that thread the loop and then peel it from the silicon backing. And then this is the thing that gets inserted into the tissue, and then this pulls out, leaving the thread. And this kind of a notch or the shark tooth that we used to call is the thing that actually is grasping the loop. And then it's designed in such way, such that when you pull out, leaps the loop.
Paragraph 697
And the robot is controlling this needle?
Paragraph 698
Correct, so this is actually housed in a cannula. And basically, the robot has a lot of the optics that look for where the loop is. There's actually a 405 nanometer light that actually causes the polyamide to fluoresce so that you can locate the location of the loop.
Paragraph 699
So the loop lights up?
Paragraph 700
Yeah, yeah, they do. It's a micron precision process.
Paragraph 701
What's interesting about the robot that it takes to do that? That's pretty crazy. That's pretty crazy that robot is able to get this kind of precision.
Paragraph 702
Yeah, our robot is quite heavy. Our current version of it. There's, I mean, it's like a giant granite slab that weighs about a ton, 'cause it needs to be sensitive to vibration, environmental vibration. And then as the head is moving at the speed that is moving, there's a lot of kind of motion control to make sure that you can achieve that level of precision. A lot of optics that kind of zoom in on that. We're working on next generation of the robot that is lighter, easier to transport. I mean, it is a feat to move the robot.
Paragraph 703
And it's far superior to a human surgeon at this time for this particular task.
Paragraph 704
Absolutely, I mean, let alone you try to actually thread a loop in a sewing kit, I mean this is like, we're talking like fractions of human hair. These things, it's not visible.
Paragraph 705
So continuing the paragraph, "We developed novel hardware and software testing systems such as our accelerated lifetime testing racks and simulated surgery environment," which is pretty cool, "to stress test and validate the robustness of our technologies. We performed many rehearsals of our surgeries to refine our procedures and make them second nature." This is pretty cool. "We practice surgeries on proxies with all the hardware and instruments needed in our mock or in the engineering space. This helps us rapidly test and measure." So there's like proxies.
Paragraph 706
Yeah, this proxy's super cool actually. So there's a 3D printed skull from the images that is taken at Barrow, as well as this hydrogel mix, sort of synthetic polymer thing that actually mimics the mechanical properties of the brain. It also has vasculature of the person.
Paragraph 707
So basically, what we're talking about here, and there's a lot of work that has gone into making this set proxy that it's about like finding the right concentration of these different synthetic polymers to get the right set of consistency for the needle dynamics, as they're being inserted. But we practice this surgery with the person, Noland's basically physiology and brain many, many times prior to actually doing the surgery.
Paragraph 708
So every step, every step?
Paragraph 709
Every step, yeah. Like where does someone stand? I mean, what you're looking at is the picture. This is in our office of this kind of corner of the robot engineering space that we have created this like mock or space that looks exactly like what they would experience, all the staff would experience during their actual surgery.
Paragraph 710
So I mean, it's just kind of like any dance rehearsal where you know exactly where you're gonna stand at what point and you just practice that over and over and over again with an exact anatomy of someone that you're going to surgerize. And it got to a point where a lot of our engineers, when we created a craniectomy, they're like, "Oh, that looks very familiar. We've seen that before."
Paragraph 711
Yeah. Man, there's wisdom you can gain through doing the same thing over and over and over. It's like a Jira dreams of sushi kind of thing, because then it's like Olympic athletes visualize the Olympics, and then once you actually show up, it feels easy. It feels like any other day. It feels almost boring winning the gold medal, 'cause you visualized this so many times, you've practiced this so many times, and nothing bothers you. It's boring. You win the gold medal, it's boring. And the experience they talk about is mostly just relief, probably that they don't have to visualize it anymore.
Paragraph 712
Yeah, the power of the mind to visualize, I mean, there's a whole field that studies where muscle memory lies in cerebellum. Yeah, it's incredible.
Paragraph 713
I think it is a good place to actually ask sort of the big question that people might have is, how do we know every aspect of this that you describe is safe?
Paragraph 714
At the end of the day, the gold standard is to look at the tissue. What sort of trauma did you cause the tissue? And does that correlate to whatever behavioral anomalies that you may have seen? And that's the language to which we can communicate about the safety of inserting something into the brain and what type of trauma that you can cause. So we actually have an entire department, department of pathology that looks at these tissue slices.
Paragraph 715
There are many steps that are involved in doing this. Once you have studies that are launched with particular endpoints in mind, at some point, you have to euthanize the animal and then you go through necropsy to kind of collect the brain tissue samples. You fix them in formalin, and you like gross them, you section them, and you look at individual slices just to see what kind of reaction or lack thereof exists.
Paragraph 716
So that's the kind of the language to which FDA speaks and as well for us to kind of evaluate the safety of the insertion mechanism as well as the threads at various different time points. Both acute, so anywhere between zero to three months to beyond three months.
Paragraph 717
So those are kind of the details of an extremely high standard of safety that has to be reached.
Paragraph 718
Correct.
Paragraph 719
FDA supervises this, but this, in general, just a very high standard. And every aspect of this, including the surgery, I think Matthew MacDougall has mentioned that like the standard is, let's say, how to put it politely? Higher than maybe some other operations that we take for granted. So the standard for all the surgical stuff here is extremely high.
Paragraph 720
Very high. I mean, it's a highly, highly regulated environment with the governing agencies that scrutinize every medical device that gets marketed. And I think it's a good thing. It's good to have those high standards, and we try to hold extremely high standards to kind of understand what sort of damage, if any, these innovative emerging technologies and new technologies that we're building are. And so far, we have been extremely impressed by lack of immune response from these threads.
Paragraph 721
Speaking of which, you talk to me with excitement about the histology and some of the images that you're able to share. Can you explain to me what we're looking at?
Paragraph 722
Yeah, so what you're looking at is a stained tissue image. So this is a sectioned tissue slice from an animal that was implanted for seven months. So kind of a chronic time point. And you're seeing all these different colors, and each color indicates specific types of cell types. So purple and pink are astrocytes and microglia respectably. They're types of glial cells.
Paragraph 723
And yet the other thing that people may not be aware of is your brain is not just made up of soup of neurons and axons. There are other cells, like glial cells, that actually kind of is the glue and also react if there are any trauma or damage to the tissue.
Paragraph 724
The brown are the neurons here?
Paragraph 725
The brown are the neurons.
Paragraph 726
The modern neurons.
Paragraph 727
So what you're seeing is, in this kind of macro image, you're seeing these like circle highlighted in white, the insertion sites. And when you zoom into one of those, you see the threads. And then in this particular case, I think we're seeing about the 16 wires that are going into the page.
Paragraph 728
And the incredible thing here is the fact that you have the neurons that are these brown structures or brown circular or elliptical thing that are actually touching and abutting the thread. So what this is saying is that there's basically zero trauma that's caused during this insertion. And with these neural interfaces, these micro electrodes that you insert, that is one of the most common mode of failure.
Paragraph 729
So when you insert these threads, like the Utah array, it causes neuronal death around the site, because you're inserting a foreign object, right? And that kind of elicit these like immune response through microglia and astrocytes. They form this like protective layer around it.
Paragraph 730
Oh, not only are you killing the neuron cells, but you're also creating this protective layer that then basically prevents you from recording neural signals, 'cause you're getting further and further away from the neurons that you're trying to record. And that is the biggest mode of failure. And in this particular example, in that inside, it's about 50 micron with that scale bar. The neurons just seem to be attracted to it.
Paragraph 731
(Lex laughing) - And so there's certainly no trauma. That's such a beautiful image, by the way. So the brown are the neurons. And for some reason, I can't look away. It's really cool.
Paragraph 732
Yeah, and the way that these things like, I mean, your tissues generally don't have these beautiful colors. This is multiplex stain that uses these different proteins that are staining these at different colors. We use very standard set of staining techniques, with HG, EB1 and new N and GFAP.
Paragraph 733
So if you go to the next image, this is also kind of illustrates the second point, 'cause you can make an argument, and initially, when we saw the previous image, we said, "Oh, like are the threads just floating? Like what is happening here? Like are we actually looking at the right thing?" So what we did is we did another stain, and this is all done in-house, of this batons, trichrome stain, which is in blue that shows these collagen layers.
Paragraph 734
So the blue basically, like you don't want the blue around the implant threads, 'cause that means that there's some sort of scarring that's happen. And what you're seeing, if you look at individual threads, is that you don't see any of the blue, which means that there has been absolutely, or very, very minimal to a point where it's not detectable amount of trauma in these inserted threads.
Paragraph 735
So that presumably is one of the big benefits of having this kind of flexible thread.
Paragraph 736
Yeah, so we think this is primarily due to the size, as well as the flexibility of the threads. Also the fact that R1 is avoiding vasculature, so we're not disrupting or we're not causing damage to the vessels and not breaking any of the blood brain barrier has basically caused the immune response to be muted.
Paragraph 737
But this is also a nice illustration of the size of things. So this is the tip of the thread.
Paragraph 738
Yeah, those are neurons.
Paragraph 739
And they're neurons. And this is the thread listening. And the electrodes are positioned how?
Paragraph 740
Yeah, so this is, what you're looking at is not electrode themselves. Those are the conductive wires. So each of those should probably be two micron in width. So what we're looking at is we're looking at the coronal slice. So we're looking at some slice of the tissue. So as you go deeper, you'll obviously have less and less of the tapering of the thread. But yeah, the point basically being that there's just kind of cells around the inserter site, which is just an incredible thing to see. I've just never seen anything like this.
Paragraph 741
How easy and safe is it to remove the implant?
Paragraph 742
Yeah, so it depends on when. In the first three months or so after the surgery, there's a lot of kind of tissue modeling that's happening. Similar to when you got a cut, you obviously start over first couple weeks, or depending on the size of the wound, scar tissue forming, right? There are these like contractive, and then in the end, they turn into scab and you can scab it off.
Paragraph 743
The same thing happens in the brain, and it's a very dynamic environment. And before the scar tissue or the neomembrane or the neomembrane that forms, it's quite easy to just pull 'em out. And there's minimal trauma that's caused during that. Once the scar tissue forms, and with Noland as well, we believe that that's the thing that's currently anchoring the thread. So we haven't seen any more movements since then. So they're quite stable.
Paragraph 744
It gets harder to actually completely extract the threads. So our current method for removing the device is cutting the thread, leaving the tissue intact, and then unscrewing and taking the implant out. And that hole is now gonna be plugged with either another Neuralink or just with kind of a peak-based, plastic-based cap.
Paragraph 745
Is it okay to leave the threads in there forever?
Paragraph 746
Yeah, we think so. We've done studies where we left them there, and one of the biggest concerns that we had is like, do they migrate and do they get to a point where they should not be? We haven't seen that. Again, once the scar tissue forms, they get anchored in place. And I should also say that when we say upgrades, like we're not just talking in theory here. Like we've actually upgraded many, many times.
Paragraph 747
Most of our monkeys or non-human primates, NHP have been upgraded. Pager, who you saw playing mind pong, has the latest version of device since two years ago and is seemingly very happy and healthy and fat.
Paragraph 748
So what's designed for the future, the upgrade procedure? So maybe for Noland. What would the upgrade look like? It was essentially what you're mentioning. Is there a way to upgrade sort of the device internally, where you take it apart and sort of keep the capsule and upgrade the internals?
Paragraph 749
Yeah, so there are a couple different things here. So for Noland, if we were to upgrade, what we would have to do is either cut the threads or extract the threads depending on kind of the situation there in terms of how they're anchored or scarred in. If you were to remove them with the dural substitute, you have an intact brain so you can reinsert different threads with the updated implant package.
Paragraph 750
There are a couple different other ways that we're thinking about, the future of what the upgradable system looks like. One is, at the moment, we currently remove the dura, this kind of thick layer that protects the brain, but that actually is the thing that actually proliferates the scar tissue formation. So typically, general good rule of thumb is you wanna leave the nature as is and not disrupt it as much.
Paragraph 751
So looking at ways to insert the threads through the dura, which comes with different set of challenges, such as it's a pretty thick layer, so how do you actually penetrate that without breaking the needle? So we're looking at different needle design for that, as well as the kind of the loop engagement. The other biggest challenges are it's quite opaque, optically, and with white light illumination.
Paragraph 752
So how do you avoid still this biggest advantage that we have of avoiding vasculature? How do you image through that? How do you actually still mediate that? So there are other imaging techniques that we're looking at to enable that. But our hypothesis is that, and based on some of the early evidence that we have, doing through the dura insertion will cause minimal scarring that causes them to be much easier to extract over time.
Paragraph 753
And the other thing that we're also looking at, this is gonna be a fundamental change in the implant architecture, is at the moment, it's a monolithic single implant that comes with a thread that's bonded together. So you can't actually separate the thing out, but you can imagine having two part implant. Bottom part, that is the thread that are inserted that has the chips and maybe a radio and some power source.
Paragraph 754
And then you have another implant that has more of the computational heavy load and the bigger battery. And then one can be under the dura, one can be above the dura, like being the plug for the skull. They can talk to each other, but the thing that you wanna upgrade, the computer and not the thread. If you wanna upgrade that, you just go in there, remove the screws and then put in the next version. It's a very, very easy surgery too.
Paragraph 755
Like you do a skin incision, slip this in, screw. Probably be able to do this in 10 minutes.
Paragraph 756
So that would allow you to reuse the thread, sort of.
Paragraph 757
[DJ] Correct.
Paragraph 758
So I mean, this leads to the natural question of, what is the pathway to scaling that increase in the number of threads? Is that a priority? What's the technical challenge there?
Paragraph 759
Yeah, that is a priority. So for next versions of the implant, the key metrics that we're looking to improve are number of channels, just recording from more and more neurons. We have a pathway to actually go from currently 1,000 to hopefully 3,000 if not 6,000 by end of this year. And then end of next year, we wanna get to even more, 16,000.
Paragraph 760
Wow.
Paragraph 761
There's a couple limitations to that. One is obviously being able to photo lithographically print those wires. As I mentioned, it's two micron in width and spacing. Obviously, there are chips that are much more advanced than those types of resolution, and we have some of the tools that we have brought in house to be able to do that. So traces will be narrower just so that you have to have more of the wires coming up into the chip.
Paragraph 762
Chips also cannot linearly consume more energy, as you have more and more channels. So there's a lot of innovations in the circuit and architecture as well as the circuit design topology to make them lower power. You need to also think about, if you have all of these spikes, how do you send that off to the end application? So you need to think about bandwidth limitation there and potentially innovations and signal processing.
Paragraph 763
Physically, one of the biggest challenge is gonna be the interface. It's always the interface that breaks. Bonding this thin film array to the electronics. It starts to become very, very highly dense interconnects. So how do you characterize that? There's a lot of innovations in kind of the 3D integrations in the recent years that we can take advantage of. One of the biggest challenges that we do have is forming this hermetic barrier, right?
Paragraph 764
That this is an extremely harsh environment that we're in - the brain. So how do you protect it from, yeah, like the brain trying to kill your electronics to also your electronics leaking things that you don't want into the brain. And that forming that hermetic barrier is gonna be a very, very big challenge that we I think are actually well-suited to tackle.
Paragraph 765
How do you test that? Like what's the development environment to simulate that kind of harshness?
Paragraph 766
Yeah, so this is where the accelerated life tester essentially is a brain in a vat. It literally is a vessel that is made up of, and again, for all intents and purpose for this particular type of test, your brain is a salt water. And you can also put some other set of chemicals like reactive oxygen species that get at kind of these interfaces and trying to cause a reaction to pull it apart.
Paragraph 767
But you could also increase the rate at which these interfaces are aging by just increasing temperature. So every 10 degrees Celsius that you increase, you're basically accelerating time by 2x. And there's limit as to how much temperature you wanna increase, 'cause at some point, there's some other non-linear dynamics that causes you to have other nasty gases to form that just is not realistic in an environment.
Paragraph 768
So what we do is we increase in our ALT chamber by 20 degrees Celsius that increases the aging by four times. So essentially one day in ALT chamber, it's four day in calendar year. And we look at whether the implants still are intact, including the threads and- - And operation and all of that?
Paragraph 769
And operation and all of that. It obviously is not an exact same environment as a brain, 'cause brain has mechanical, other more biological groups that attack at it. But it is a good test environment, testing environment for at least the enclosure and the strength of the enclosure. And I mean, we've had implants, the current version of the implant that has been in there for, I mean, close to two and a half years, which is equivalent to a decade. And they seem to be fine.
Paragraph 770
So it's interesting that the burn, so basically, close approximation is warm salt water, hot salt water is a good testing environment. Yeah, by the way, I'm drinking LMNT, which is basically salt water, which is making me kinda, it doesn't have computational power the way the brain does, but maybe in terms of other characteristics, it's quite similar and I'm consuming it.
Paragraph 771
Yeah, you have to get it in the right pH too. (laughs) - And then consciousness will emerge. Yeah, no.
Paragraph 772
By the way, the other thing that also is interesting about our enclosure is, if you look at our implant, it's not your common-looking medical implant that usually is encased in a titanium can that's laser welded. We use this polymer called PCTFE, polychlorotrifluoroethylene, which is actually commonly used in blister packs. So when you have a pill and you try to pop a pill, there's like kind of that plastic membrane. That's what this is.
Paragraph 773
No one's actually ever used this except us. And the reason we wanted to do this is 'cause it's electromagnetically transparent. So when we talked about the electromagnetic inductive charging, with titanium can, usually, if you wanna do something like that, you have to have a sapphire window, and it's a very, very tough process to scale.
Paragraph 774
So you're doing a lot of iteration here in every aspect of this. The materials, the software- - The whole, whole shebang.
Paragraph 775
So, okay. So you mentioned scaling. Is it possible to have multiple Neuralink devices as one of the ways of scaling? To have multiple Neuralink devices implanted?
Paragraph 776
That's the goal, that's the goal. Yeah, we've had, I mean, our monkeys have had two Neuralinks, one in each hemisphere. And then we're also looking at potential of having one in motor cortex, one in visual cortex, and one in wherever other cortex.
Paragraph 777
So focusing on a particular function, one Neuralink device.
Paragraph 778
Correct.
Paragraph 779
I mean, I wonder if there's some level of customization that can be done on the compute side. So for the motor cortex- - Absolutely. That's the goal. And we talk about at Neuralink building a generalized neural interface to the brain. And that also is strategically how we're approaching this with marketing. And also, with regulatory, which is, hey look, we have the robot, and the robot can access any part of the cortex.
Paragraph 780
Right now, we're focused on motor cortex with current version of the N1 that's specialized for motor decoding tasks. But also, at the end of the day, there's kind of a general compute available there. But typically, if you wanna really get down to kind of hyperoptimizing for power and efficiency, you do need to get to some specialized function, right?
Paragraph 781
But what we're saying is that, hey, you are now used to this robotic insertion techniques, which took many, many years of showing data and conversation with the FDA. And also, internally convincing ourselves that this is safe. And now the difference is that if we go to other parts of the brain, like visual cortex, which we're interested in as our second product, obviously, it's a completely different environment.
Paragraph 782
The cortex is laid out very, very differently. It's gonna be more stimulation focus rather than recording, just kind of creating visual percepts. But in the end, we're using the same thin film array technology. We're using the same robot insertion technology. We're using the same packaging technology. Now, more the conversation is focused around what are the differences and what are the implication of those differences in safety and efficacy?
Paragraph 783
The way you said second product is both hilarious and awesome to me. That product being restoring sight for blind people. So can you speak to stimulating the visual cortex? I mean, the possibilities there are just incredible to be able to give that gift back to people who don't have sight or even any aspect of that. Can you just speak to the challenges of, there's several challenges here.
Paragraph 784
Oh, many.
Paragraph 785
One of which is, like you said, from recording to stimulation. Just any aspect of that that you're both excited and see the challenges of.
Paragraph 786
Yeah, I guess I'll start by saying that we actually have been capable of stimulating through our thin film array as well as other electronics for years. We have actually demonstrated some of that capabilities for reanimating the limb in the spinal cord. Obviously, for the current EFS study, we've hardware disabled that, so that's something that we wanted to embark as a separate, separate journey.
Paragraph 787
And obviously, there are many, many different ways to write information into the brain. The way in which we're doing that is through electrical, passing electrical current, and kind of causing that to really change the local environment so that you can sort of artificially cause kind of the neurons to depolarize in nearby areas. For vision specifically, the way our visual system works, it's both well-understood.
Paragraph 788
I mean, anything with kind of brain, there are aspects of it that's well-understood. But in the end, like we don't really know anything. But the way visual system works is that you have photon hitting your eye, and in your eyes, there are these specialized cells called photoreceptor cells that convert the photon energy into electrical signals. That then gets projected to your back of your head, your visual cortex.
Paragraph 789
It goes through actually thalamic system called LGN that then projects it out. And then in the visual cortex, there's visual area one or V1, and then there's bunch of other higher level processing layers like V2, V3. And there there are actually kind of interesting parallels.
Paragraph 790
And when you study the behaviors of these convolutional neural networks, like what the different layers of the network is detecting, first, they're detecting like these edges, and they're then detecting some more natural curves, and then they start to detect like objects, right? Kind of similar thing happens in the brain. And a lot of that has been inspired, and also, it's been kinda exciting to see some of the correlations there.
Paragraph 791
But things like from there, where those cognition arise and where's color encoded, there's just not a lot of understanding, fundamental understanding there. So in terms of kind of bringing sight back to those that are blind, there are many different forms of blindness. There's actually million people, one million people in the U. S. that are legally blind. That means like certain, like score below in kind of the visual tests.
Paragraph 792
I think it's something like, if you can see something at 20 feet distance, that normal people can see at 200 feet distance, like if you're worse than that, you're legally blind.
Paragraph 793
So that means you can't function effectively.
Paragraph 794
Correct.
Paragraph 795
Using sight in the world.
Paragraph 796
Yeah, like to navigate your environment. And yeah, there are different forms of blindness. There are forms of blindness where there's some degeneration of your retina, these photoreceptor cells, and rest of your visual processing that I described is intact. And for those types of individuals, you may not need to maybe stick electrodes into the visual cortex.
Paragraph 797
You can actually build retinal prosthetic devices that actually just replaces the function of that retinal cells that are degenerated. And there are many companies that are working on that. But that's a very small slice. Albeit significance, those smaller slice of folks that are legally blind.
Paragraph 798
If there's any damage along that circuitry, whether it's in the optic nerve or just the LGN circuitry or any break in that circuit, that's not gonna work for you. And the source of where you need to actually cause that visual percept to happen, because your biological mechanism not doing that is by placing electrodes in the visual cortex in the back of your head.
Paragraph 799
And the way in which this would work is that you would have an external camera, whether it's something as unsophisticated as a GoPro or some sort of wearable RayBan type glasses that Meta's working on that captures a scene, right? And that scene is then converted to set of electrical impulses or stimulation pulses that you would activate in your visual cortex through these thin film arrays.
Paragraph 800
And by playing in a concerted kind of orchestra of these stimulation patterns, you can create what's called phosphenes, which are these kind of white yellowish dots that you can also create by just pressing your eyes. You can actually create those percepts by stimulating in the visual cortex.
Paragraph 801
And the name of the game is really have many of those and have those percepts be, the phosphenes be as small as possible so that you can start to tell apart, like they're the individual pixels of the screen, right? So if you have many, many of those, potentially, you'll be able to, in the long term, be able to actually get naturalistic vision.
Paragraph 802
But in the like short term to maybe midterm, being able to at least be able to have object detection algorithms run on your glasses, the pre-op processing units, and then being able to at least see the edges of things so you don't bump into stuff.
Paragraph 803
It's incredible. This is really incredible. So you basically would be adding pixels, and your brain would start to figure out what those pixels mean. Yeah, and like with different kinds of assistant on the signal processing on all fronts.
Paragraph 804
Yeah. The thing that actually, so a couple things. One is, obviously, if you're blind from birth, the way brain works, especially in the early age, neuroplasticity is really nothing other than kind of your brain and different parts of your brain fighting for the limited territory.
Paragraph 805
[Lex] (laughs) Yeah.
Paragraph 806
And I mean, very, very quickly, you see cases where you know people that are, I mean, you also hear about people who are blind that have heightened sense of hearing or some other senses. And the reason for that is that cortex that's not used just gets taken over by these different parts of the cortex. So for those types of individuals, I mean, I guess they're going to have to now map some other parts of their senses into what they call vision.
Paragraph 807
But it's gonna be obviously a very, very different conscious experience before. So I think that's a interesting caveat. The other thing that also is important to highlight is that we're currently limited by our biology in terms of the wavelength that we can see. There's a very, very small wavelength that is a visible light wavelength that we can see with our eyes.
Paragraph 808
But when you have an external camera with this BCI system, you're not limited to that. You can have infrared, you can have UV, you can have whatever other spectrum that you want to see. And whether that gets matched to some sort of weird conscious experience, I've no idea. But oftentimes, I talk to people about the goal of Neuralink being going beyond the limits of our biology. That's sort of what I mean.
Paragraph 809
And if you're able to control the kind of raw signal, when we use our sight, we're getting the photons and there's not much processing on it. If you're being able to control that signal, maybe you can do some kind of processing. Maybe you do object detection ahead of time.
Paragraph 810
[DJ] Yeah.
Paragraph 811
You're doing some kind of pre-processing, and there's a lot of possibilities to explore that. So it's not just increasing sort of thermal imaging, that kind of stuff, but it's also just doing some kind of interesting processing.
Paragraph 812
Correct, yeah. I mean, my theory of how like visual system works also is that, I mean, there's just so many things happening in the world, and there's a lot of photons that are going into your eye, and it's unclear exactly where some of the pre-processing steps are happening. But I mean, I actually think that just from a fundamental perspective, there's just so much, the reality that we're in, if it's a reality, is so there's so much data.
Paragraph 813
And I think humans are just unable to actually like eat enough actually to process all that information. So there's some sort of filtering that does happen, whether that happens in the retina, whether that happens in different layers of the visual cortex. Unclear.
Paragraph 814
But like the analogy that I sometimes think about is, if your brain is a CCD camera, and all of the information in the world is a sun, and when you try to actually look at the sun with the CCD camera, it's just gonna saturate the sensors, right? 'Cause it's enormous amount of energy. So what you do is you end up adding these filters, right? To just kind of narrow the information that's coming to you and being captured.
Paragraph 815
And I think things like our experiences or our like drugs like propofol, that like anesthetic drug or psychedelics, what they're doing is they're kind of swapping out these filters and putting in new ones or removing older ones and kind of controlling our conscious experience.
Paragraph 816
Yeah, man, not to distract from the topic, but I just took a very high dose of ayahuasca in the Amazon jungle. So yes, it's a nice way to think about it. You're swapping out different experiences, and with Neuralink being able to control that, primarily at first, to improve function, not for entertainment purposes or enjoyment purposes, but- - Yeah, giving back lost functions.
Paragraph 817
Well, giving back lost functions. And there, especially when the function is completely lost, anything is a huge help. Would you implant a Neuralink device in your own brain?
Paragraph 818
Absolutely. I mean, maybe not right now, but absolutely.
Paragraph 819
What kind of capability, once reached, you would start getting real curious and almost get a little antsy, like jealous of people that get it as you watch them get implanted?
Paragraph 820
Yeah, I mean I think, I mean, even with our early participants, if they start to do things that I can't do, which I think is in the realm of possibility for them to be able to get, 15, 20, if not like 100 BPS right? There's nothing that fundamentally stops us from being able to achieve that type of performance. I mean, I would certainly get jealous that they can do that.
Paragraph 821
I should say that watching Noland, I get a little jealous 'cause he's having so much fun, and it seems like such a chill way to play video games.
Paragraph 822
Yeah. So I mean, the thing that also is hard to appreciate sometimes is that, he's doing these things while talking. I mean, it's multitasking, right? So it's clearly, it's obviously cognitively intensive, but similar to how, when we talk, we move our hands, like these things like are multitasking. I mean, he's able to do that. And you won't be able to do that with other assistive technology as far as I'm aware.
Paragraph 823
If you're obviously using like an eye tracking device, you're very much fixated on that thing that you're trying to do. And if you're using voice control, I mean, like if you say some other stuff, yeah, you don't get to use that.
Paragraph 824
Yeah, the multitasking aspect of that is really interesting. So it's not just the BPS for the primary task. It's the parallelization of multiple tasks. If you measure the BPS for the entirety of the human organism, so if you're talking and doing a thing with your mind and looking around also, I mean, there's just a lot of paralyzation that can be happening.
Paragraph 825
Yeah. But I mean, I think at some point for him, like if he wants to really achieve those high level BPS, it does require like full attention, right? And that's a separate circuitry that is a big mystery. Like how attention works and, you know?
Paragraph 826
Yeah, attention, like cognitive load, I've read a lot of literature on people doing two tasks. Like you have your primary task and a secondary task. And the secondary task is a source of distraction. And how does that affect the performance of the primary task? And depending on the tasks, there's a lot of interesting, I mean, this is an interesting computational device, right? And I think- - To say the least.
Paragraph 827
A lot of novel insights that can be gained from everything. I mean, I personally am surprised that Noland's able to do such incredible control of the cursor while talking and also being nervous at the same time, 'cause he's talking like all of us are, if you're talking in front of the camera, you get nervous. So all of those are coming into play, and he is able to still achieve high performance. Surprising. I mean, all of this is really amazing. And I think just after researching this really in depth, I kind of want Neuralink.
Paragraph 828
(laughs) Get in line.
Paragraph 829
And also, the safety get in line. Well, we should say the registry is for people who have quadriplegia and all that kind of stuff so- - Correct.
Paragraph 830
There'll be a separate line for people. They're just curious, like myself. So now that Noland, patient P1, is part of the ongoing prime study, what's the high level vision for P2, P3, P4, P5? And just the expansion into other human beings that are getting to experience this implant?
Paragraph 831
Yeah, I mean, the primary goal is, for our study in the first place is to achieve safety endpoints. Just understand safety of this device, as well as the implantation process. And also, at the same time, understand the efficacy and the impact that it could have on the potential users' lives. And just because you have, you're living with tetraplegia, it doesn't mean your situation is same as another person living with tetraplegia.
Paragraph 832
It's wildly, wildly varying. It's something that we're hoping to also understand how our technology can serve not just a very small slice of those individuals, but broader group of individuals and being able to get the feedback to just really build just the best product for them.
Paragraph 833
There's obviously also goals that we have, and the primary purpose of the early feasibility study is to learn from each and every participant to improve the device, improve the surgery before we embark on what's called a pivotal study that then is much larger trial that starts to look at statistical significance of your endpoints, and that's required before you can then market the device. And that's how it works in the U. S.
Paragraph 834
and just generally around the world. That's the process you follow. So our goal is to really just understand from people like Noland, P2, P3, future participants, what aspects of our device needs to improve. If it turns out that people are like, "I really don't like the fact that it lasts only six hours. I wanna be able to use this computer for like 24 hours."
Paragraph 835
I mean, that is a user needs and user requirements, which we can only find out from just being able to engage with them.
Paragraph 836
So before the pivotal study, there's kind of like a rapid innovation based on individual experiences. You're learning from individual people how they use it, like the high resolution details in terms of like cursor control and signal and all that kind of stuff to like life experience.
Paragraph 837
Yeah, yeah, so there's hardware changes but also just firmware updates. So even when we had that sort of recovery event for Noland, he now has the new firmware that he has been updated with. And it's similar to how like your phones get updated all the time with new firmware for security patches, whatever new functionality UI, right? And that's something that is possible with our implant.
Paragraph 838
It's not a static one-time device that can only do the thing that it said it can do. I mean, it's similar to Tesla. You can do over the air firmware updates, and now you have completely new user interface. And all this bells and whistles and improvements on everything like the latest, right? When we say generalized platform, that's what we're talking about.
Paragraph 839
Yeah, it's really cool how the app that Noland is using, there's like calibration, all that kind of stuff. And then there's update. You just click and get an update. What other future capabilities are you kinda looking to? You said vision. That's a fascinating one. What about sort of accelerated typing or speech, this kind of stuff? And what else is there?
Paragraph 840
Yeah, those are still in the realm of movement program. So largely speaking, we have two programs. We have the movement program and we have the vision program. The movement program currently is focused around the digital freedom. As you can easily guess, if you can control 2D cursor in the digital space, you could move anything in the physical space.
Paragraph 841
So robotic arms, wheelchair, your environment, or even really like, whether it's through the phone or just like directly to those interfaces, so like to those machines. So we're looking at ways to kind of expand those types of capability, even for Noland.
Paragraph 842
That requires conversation with the FDA and kind of showing safety data for, if there's a robotic arm or a wheelchair that we can guarantee that they're not gonna hurt themselves accidentally, right? It's very different if you're moving stuff in the digital domain versus like in the physical space, you can actually potentially cause harm to the participants. So we're working through that right now.
Paragraph 843
Speech does involve different areas of the brain. Speech prosthetic is very, very fascinating, and there's actually been a lot of really amazing work that's been happening in academia. Sergei Stavisky at UC Davis, Jaimie Henderson, and late Krishna Shenoy at Stanford doing just some incredible amount of work in improving speech neuroprosthetics.
Paragraph 844
And those are actually looking more at parts of the motor cortex that are controlling these focal articulators. And being able to like, even by mouthing the word or imagine speech, you can pick up those signals. The more sophisticated higher level processing areas, like the Broca's area or Wernicke's area, those are still very, very big mystery in terms of the underlying mechanism of how all that stuff works.
Paragraph 845
But yeah, I mean, I think Neuralink's event goal is to kind of understand those things and be able to provide a platform and tools to be able to understand that and study that.
Paragraph 846
This is where I get to the pothead questions. Do you think we can start getting insight into things like thought? So speech is, there's a muscular component, like you said. There's like the act of producing sounds. But then what about the internal things like cognition? Like low level thoughts and high level thoughts. Do you think we'll start noticing kind of signals that could be picked up? They could be understood, they could be maybe used in order to interact with the outside world.
Paragraph 847
In some ways, like I guess this starts to kind of get into the heart problem of consciousness. And I mean, on one hand, all of these are, at some point, set of electrical signals, that from there, maybe it in itself is giving you the cognition or the meaning, or somehow, human mind is incredibly amazing storytelling machine. So we're telling ourselves and fooling ourselves that there's some interesting meaning here.
Paragraph 848
But I mean, I certainly think that BCI and really BCI at the end of the day is a set of tools that help you kind of study the underlying mechanisms in both like local but also broader sense. And whether there's some interesting patterns of like electrical signal, that means like you're thinking this versus, and you can either like learn from like many, many sets of data to correlate some of that and be able to do mind reading or not.
Paragraph 849
I'm not sure. I certainly would not kind of rule that out as a possibility, but I think BCI alone probably can't do that. There's probably additional set of tools and framework. And also, like just heart problem of consciousness at the end of the day is rooted in this philosophical question of like, what's the meaning of it all? What's the nature of our existence? Where's the mind emerge from this complex network?
Paragraph 850
Yeah, how does the subjective experience emerge from just a bunch of spikes, electrical spikes?
Paragraph 851
Yeah, yeah, I mean, we do really think about BCI and what we're building as a tool for understanding the mind, the brain. The only question that matters. There's actually, there actually is some biological existence proof of like what it would take to kind of start to form some of these experiences that may be unique. If you actually look at every one of our brains, there are two hemispheres.
Paragraph 852
There's a left-sided brain, there's a right-sided brain. And I mean, unless you have some other conditions, you normally don't feel like left legs or right legs. Like you just feel like one legs, right? So what is happening there, right? If you actually look at the two hemispheres, there's a structure that kind of characterize the two called the corpus callosum that is supposed to have around 200 to 300 million connections or axons.
Paragraph 853
So whether that means that's the number of interface and electrodes that we need to create some sort of mind meld, or from that, like whatever new conscious experience that you can experience. But yeah, I do think that there's like kind of an interesting existence proof that we all have.
Paragraph 854
And that threshold is unknown at this time.
Paragraph 855
Oh yeah, these things, everything in this domain is speculation, right?
Paragraph 856
And then there would be, you'd be continuously pleasantly surprised. Do you see a world where there's millions of people, like tens of millions, hundreds of millions of people walking around with a Neuralink device, or multiple Neuralink devices in their brain?
Paragraph 857
I do. First of all, there are, like if you look at worldwide, people suffering from movement disorders and visual deficits. I mean, that's in the tens if not hundreds of millions of people. So that alone, I think, there's a lot of benefit and potential good that we can do with this type of technology. And once you start to get into kind of neuro, like psychiatric application, depression, anxiety, hunger, or obesity, right? Like mood, control of appetite, I mean, that starts to become very real to everyone.
Paragraph 858
Not to mention that most people on earth have a smartphone. And once BCI starts competing with a smartphone as a preferred methodology of interacting with the digital world, that also becomes an interesting thing.
Paragraph 859
Oh yeah, I mean, yeah. This is even before going to that, right? I mean, there's like almost, I mean, the entire world that could benefit from these types of thing. And then, yeah, like if we're talking about kind of next generation of how we interface with machines or even ourselves, in many ways, I think BCI can play a role in that. And some of the things that I also talk about is I do think that there is a real possibility that you could see eight billion people walking around with Neuralink.
Paragraph 860
Well, thank you so much for pushing ahead. And I look forward to that exciting feature.
Paragraph 861
Thanks for having me.
Paragraph 862
Thanks for listening to this conversation with DJ Seo. And now, dear friends, here's Matthew MacDougall, the head neurosurgeon at Neuralink. When did you first become fascinated with the human brain?
Paragraph 863
Since forever. As far back as I can remember, I've been interested in the human brain. I mean, I was a thoughtful kid and a bit of an outsider. And you sit there thinking about what the most important things in the world are in your little tiny adolescent brain.
Paragraph 864
And the answer that I came to, that I converged on was that all of the things you can possibly conceive of, as things that are important for human beings to care about are literally contained in the skull. Both the perception of them and their relative values. And the solutions to all our problems and all of our problems are all contained in the skull.
Paragraph 865
And if we knew more about how that worked, how the brain encodes information and generates desires and generates agony and suffering, we could do more about it. You think about all the really great triumphs in human history. You think about all the really horrific tragedies. You think about the holocaust, you think about any prison full of human stories, and all of those problems boil down to neurochemistry.
Paragraph 866
So if you get a little bit of control over that, you provide people the option to do better. And in the way I read history, the way people have dealt with having better tools is that they most often in the end do better, with huge asterisk. But I think it's an interesting, a worthy and noble pursuit to give people more options, more tools.
Paragraph 867
Yeah, that's a fascinating way to look at human history. You just imagine all these neurobiological mechanisms, Stalin, Hitler, all of these, Gengis Khan, all of them just had like a brain, just a bunch of neurons, like a few tons of billions of neurons gaining a bunch of information over a period of time. They have a set of module that does language and memory and all that. And from there, in the case of those people, they're able to murder millions of people.
Paragraph 868
[Matthew] Yeah.
Paragraph 869
All that coming from, there's not some glorified notion of a dictator of this enormous mind or something like this. It's just the brain.
Paragraph 870
Yeah, yeah. I mean, a lot of that has to do with how well people like that can organize those around them.
Paragraph 871
Other brains.
Paragraph 872
Yeah, and so I always find it interesting to look to primatology, look to our closest non-human relatives for clues as to how humans are going to behave and what particular humans are able to achieve. And so you look at chimpanzees and bonobos and they're similar but different in their social structures particularly.
Paragraph 873
And I went to Emory in Atlanta and studied under Frans, the great Frans de Waal, who was kind of the leading primatologist who recently died, and his work at looking at chimps through the lens of how you would watch an episode of "Friends" and understand the motivations of the characters interacting with each other. He would look at a chimp colony and basically apply that lens. I'm massively oversimplifying it.
Paragraph 874
If you do that, instead of just saying, subject 473 through his feces at subject 471, you talk about them in terms of their human struggles, accord them the dignity of themselves as actors with understandable goals and drives, what they want out of life. And primarily, it's the things we want out of life: food, sex, companionship, power. You can understand chimp and bonobo behavior in the same lights much more easily.
Paragraph 875
And I think doing so gives you the tools you need to reduce human behavior from the kind of false complexity that we layer onto it with language and look at it in terms of, oh, well, these humans are looking for companionship, sex, food, power. And I think that's a pretty powerful tool to have in understanding human behavior.
Paragraph 876
And I just went to the Amazon jungle for a few weeks, and it's a very visceral reminder that a lot of life on earth is just trying to get laid. They're all screaming at each other. Like I saw a lot of monkeys, and they're just trying to impress each other, or maybe if there's a battle for power, but a lot of the battle for power has to do with them getting laid.
Paragraph 877
Right. Breeding rights often go with alpha status. And so if you can get a piece of that, then you're gonna do okay.
Paragraph 878
And would like to think that we're somehow fundamentally different, but especially when it comes to primates, we're really aren't, you know. We can use fancier poetic language, but maybe some of the underlying drives that motivate us are similar.
Paragraph 879
Yeah, I think that's true.
Paragraph 880
And all of that is coming from this, the brain.
Paragraph 881
Yeah.
Paragraph 882
So when did you first start studying the brain as I guess as a biological mechanism?
Paragraph 883
Basically, the moment I got to college, I started looking around for labs that I could do neuroscience work in. I originally approached that from the angle of looking at interactions between the brain and the immune system, which isn't the most obvious place to start, but I had this idea at the time that the contents of your thoughts would have an impact, a direct impact, maybe a powerful one on non-conscious systems in your body.
Paragraph 884
The systems we think of as homeostatic, automatic mechanisms like fighting off a virus, like repairing a wound. And sure enough, there are big crossovers between the two. I mean, it gets to kind of a key point that I think goes under recognized. One of the things people don't recognize or appreciate about the human brain enough, and that is that it basically controls or has a huge role in almost everything that your body does.
Paragraph 885
Like you try to name an example of something in your body that isn't directly controlled or massively influenced by the brain, and it's pretty hard. I mean, you might say like bone healing or something, but even those systems, the hypothalamus and pituitary end up playing a role in coordinating the endocrine system that does have a direct influence on, say, the calcium level in your blood that goes to bone healing.
Paragraph 886
So non-obvious connections between those things implicate the brain as really a potent prime mover in all of health.
Paragraph 887
One of the things I realized in the other direction too, how most of the systems in the body integrated with the human brain, like they affect the brain also, like the immune system. I think there's just people who study Alzheimer's and those kinds of things. It's just surprising how much you can understand of that from the immune system, from the other systems that don't obviously seem to have anything to do with sort of the nervous system. They all play together.
Paragraph 888
Yeah, you could understand how that would be driven by evolution too, just in some simple examples. If you get sick, if you get a communicable disease, you get the flu, it's pretty advantageous for your immune system to tell your brain, "Hey, now be antisocial for a few days. Don't go be the life of the party tonight. In fact, maybe just cuddle up somewhere warm under a blanket and just stay there for a day or two."
Paragraph 889
And sure enough, that tends to be the behavior that you see both in animals and in humans. If you get sick, elevated levels of interleukins in your blood and TNF alpha in your blood ask the brain to cut back on social activity. And even moving around, you have lower locomotor activity in animals that are infected with viruses.
Paragraph 890
So from there, the early days in neuroscience to surgery, when did that step happen?
Paragraph 891
Yeah.
Paragraph 892
This is a leap.
Paragraph 893
It was sort of an evolution of thought. I wanted to study the brain. I started studying the brain in undergrad in this neuroimmunology lab. I, from there, realized at some point that I didn't wanna just generate knowledge. I wanted to effect real changes in the actual world, in actual people's lives. And so after having not really thought about going into medical school, I was on a track to go into a PhD program.
Paragraph 894
I said, "Well, I'd like that option. I'd like to actually potentially help tangible people in front of me." And doing a little digging found that there exists these MD-PhD programs where you can choose not to choose between them and do both.
Paragraph 895
And so I went to USC for medical school and had a joint PhD program with Caltech, where I actually chose that program particularly because of a researcher at Caltech named Richard Andersen, who's one of the godfathers of primate neuroscience and has a MACAC lab where Utah arrays and other electrodes were being inserted into the brains of monkeys to try to understand how intentions were being encoded in the brain.
Paragraph 896
So I ended up there with the idea that maybe I would be a neurologist and study the brain on the side, and then discovered that neurology, again, I'm gonna make enemies by saying this, but neurology predominantly and distressingly to me is the practice of diagnosing a thing and then saying, "Good luck with that. There's not much we can do."
Paragraph 897
And neurosurgery, very differently, it's a powerful lever on taking people that are headed in a bad direction and changing their course in the sense of brain tumors that are potentially treatable or curable with surgery. Even aneurysms in the brain, blood vessels that are gonna rupture, you can save lives really is at the end of the day, what mattered to me.
Paragraph 898
And so I was at USC, as I mentioned, that happens to be one of the great neurosurgery programs. And so I met these truly epic neurosurgeons, Alex Khalessi and Mike Apuzzo and Steve Giannotta and Marty Weiss, these sort of epic people that were just human beings in front of me.
Paragraph 899
And so it kind of changed my thinking from neurosurgeons are distant gods that live on another planet and occasionally come and visit us to these are humans that have problems and are people. And there's nothing fundamentally preventing me from being one of them. And so at the last minute in medical school, I changed gears from going into a different specialty and switched into neurosurgery, which cost me a year.
Paragraph 900
I had to do another year of research, because I was so far along in the process to switch into neurosurgery. The deadlines had already passed. So it was a decision that cost time, but absolutely worth it.
Paragraph 901
What was the hardest part of the training on the neurosurgeon track?
Paragraph 902
Yeah, two things. I think that residency in neurosurgery is sort of a competition of pain, of like how much pain can you eat and smile.
Paragraph 903
Yeah.
Paragraph 904
And so there's workout restrictions that are not really, they're viewed at, I think, internally among the residents as weakness. And so most neurosurgery residents try to work as hard as they can. And that I think necessarily means working long hours, and sometimes, over the work hour limits. And we care about being compliant with whatever regulations are in front of us.
Paragraph 905
But I think more important than that, people wanna give their all in becoming a better neurosurgeon, because the stakes are so high. And so it's a real fight to get residents to say go home at the end of their shift and not stay and do more surgery.
Paragraph 906
Are you seriously saying like one of the hardest things is literally like forcing them to get sleep and rest and all this kind of stuff?
Paragraph 907
Historically, that was the case. I think the next generation, I think the next generation is more compliant and more selfcare- - Weaker is what you mean. All right, I'm just kidding, I'm just kidding.
Paragraph 908
I didn't say it.
Paragraph 909
Now I'm making enemies. Okay, I get it. Wow, that's fascinating. So what was the second thing?
Paragraph 910
The personalities. And maybe the two are connected, but- - Was it pretty competitive?
Paragraph 911
It's competitive and it's also, as we touched on earlier, primates like power, and I think neurosurgery has long had this aura of mystique and excellence and whatever about it. And so it's an invitation I think for people that are cloaked in that authority. A board certified neurosurgeon is basically a walking fallacious appeal to authority, right? You have license to walk into any room and act like you're an expert on whatever. And fighting that tendency is not something that most neurosurgeons do well. Humility isn't the forte.
Paragraph 912
Yeah, so I have friends who know you, and whenever they speak about you, that you have the surprising quality for a neurosurgeon of humility, which I think indicates that it's not as common as perhaps in other professions, 'cause there is a kind of gigantic sort of heroic aspect to neurosurgery, and I think it gets to people's head a little bit.
Paragraph 913
Yeah. Well, I think that allows me to play well at an Elon company. Because Elon, one of his strengths, I think, is to just instantly see through fallacy from authority. So nobody walks into a room that he's in and says, "Well, goddamn it, you have to trust me. I'm the guy that built the last 10 rockets or something." And he says, "Well, you did it wrong and we can do it better." Or, "I'm the guy that kept Ford alive for the last 50 years.
Paragraph 914
You listen to me on how to build cars." And he says no. And so you don't walk into a room that he's in and say, "Well, I'm a neurosurgeon. Let me tell you how to do it." He's gonna say, "Well, I'm a human being that has a brain. I can think from first principles myself, thank you very much. And here's how I think it ought to be done. Let's go try it and see who's right."
Paragraph 915
And that's proven I think over and over in his case to be a very powerful approach.
Paragraph 916
If we just take that tangent, there's a fascinating interdisciplinary team at Neuralink that you get to interact with, including Elon. What do you think is the secret to a successful team? What have you learned from just getting to observe these folks? World experts in different disciplines work together.
Paragraph 917
Yeah, there's a sweet spot where people disagree and forcefully speak their mind and passionately defend their position, and yet are still able to accept information from others and change their ideas when they're wrong. And so I like the analogy of sort of how you polish rocks. You put hard things in a hard container and spin it. People bash against each other and outcome's a more refined product.
Paragraph 918
And so to make a good team at Neuralink, we've tried to find people that are not afraid to defend their ideas passionately. And occasionally, strongly disagree with people that they're working with and have the best idea come out on top. It's not an easy balance, again, to refer back to the primate brain.
Paragraph 919
It's not something that is inherently built into the primate brain to say, "I passionately put all my chips on this position and now I'm just gonna walk away from it. Admit you were right." Part of our brains tell us that that is a power loss. That is a loss of face, a loss of standing in the community. And now, you're a zeta chimp, 'cause your idea got trounced.
Paragraph 920
And you just have to recognize that that little voice in the back of your head is maladaptive and it's not helping the team win.
Paragraph 921
Yeah, you have to have the confidence to be able to walk away from an idea that you hold onto. Yeah.
Paragraph 922
Yeah.
Paragraph 923
And if you do that often enough, you're actually going to become the best in the world at your thing. I mean, that kind of that rapid iteration.
Paragraph 924
Yeah, you'll at least be a member of a winning team.
Paragraph 925
Ride the wave. What did you learn? You mentioned there's a lot of amazing neurosurgeons at USC. What lessons about surgery and life have you learned from those folks?
Paragraph 926
Yeah, I think working your ass off, working hard while functioning as a member of a team, getting a job done that is incredibly difficult, working incredibly long hours, being up all night, taking care of someone that you think probably won't survive no matter what you do. Working hard to make people that you passionately dislike look good the next morning.
Paragraph 927
These folks were relentless in their pursuit of excellent neurosurgical technique decade over decade. And I think we're well-recognized for that excellence. Especially Marty Weiss, Steve Giannotta, Mike Apuzzo, they made huge contributions not only to surgical technique, but they built training programs that trained dozens or hundreds of amazing neurosurgeons. I was just lucky to kind of be in their wake.
Paragraph 928
What's that like, you mentioned doing a surgery where the person is likely not to survive. Does that wear on you?
Paragraph 929
Yeah. It's especially challenging when you, with all respect to our elders, it doesn't hit so much when you're taking care of an 80-year-old and something was going to get them pretty soon anyway. And so you lose a patient like that, and it was part of the natural course of what is expected of them in the coming years, regardless.
Paragraph 930
Taking care of a father of two or three, four young kids, someone in their 30s that didn't have it coming, and they show up in your ER having their first seizure of their life, and lo and behold, they've got a huge, malignant, inoperable or incurable brain tumor. You can only do that, I think, a handful of times before it really starts eating away at your armor.
Paragraph 931
Or a young mother that shows up that has a giant hemorrhage in her brain that she's not gonna survive from. And they bring her four-year-old daughter to say goodbye one last time before they turn the ventilator off. The great Henry Marsh is a English neurosurgeon who said it best. I think he says that every neurosurgeon carries with them a private graveyard, and I definitely feel that, especially with young parents. That kills me.
Paragraph 932
They had a lot more to give. The loss of those people specifically has a knock on effect that's going to make the world worse for people for a long time. And it's just hard to feel powerless in the face of that. And that's where I think you have to be borderline evil to fight against a company like Neuralink or to constantly be taking pot shots at us because what we're doing is to try to fix that stuff.
Paragraph 933
We're trying to give people options, to reduce suffering. We're trying to take the pain out of life that broken brains brings in. And yeah, this is just our little way that we're fighting back against entropy, I guess.
Paragraph 934
Yeah, the amount of suffering that's endured when some of the things that we take for granted that our brain is able to do is taken away is immense. And to be able to restore some of that functionality is a real gift.
Paragraph 935
Yeah, we're just starting. We're gonna do so much more.
Paragraph 936
Well, can you take me through the full procedure of implanting, say, the N1 chip in Neuralink?
Paragraph 937
Yeah, it's a really simple, really simple, straightforward procedure. The human part of the surgery that I do is dead simple. It's one of the most basic neurosurgery procedures imaginable. And I think there's evidence that some version of it has been done for thousands of years.
Paragraph 938
That there are examples I think from ancient Egypt of healed or partially healed trephinations from Peru or ancient times in South America where these protosurgeons would drill holes in people's skulls, presumably to let out the evil spirits, but maybe to drain blood clots. And there's evidence of bone healing around the edge, meaning the people at least survive some months after a procedure. And so what we're doing is that.
Paragraph 939
We are making a cut in the skin on the top of the head over the area of the brain that is the most potent representation of hand intentions. And so if you are an expert concert pianist, this part of your brain is lighting up the entire time you're playing. We call it the hand knob.
Paragraph 940
The hand knob.
Paragraph 941
Yeah.
Paragraph 942
So it's all like the finger movement. All of that is just firing away.
Paragraph 943
Yep, there's a little squiggle in the cortex right there. One of the folds in the brain is kind of doubly folded right on that spot. So you can look at it on an MRI and say, "That's the hand knob." And then you do a functional test and a special kind of MRI called an a functional MRI, fMRI. And this part of the brain lights up when people, even quadriplegic people whose brains aren't connected to their finger movements anymore.
Paragraph 944
They imagine finger movements, and this part of the brain still lights up. So we can ID that part of the brain in anyone who's preparing to enter our trial and say, "Okay, that part of the brain, we confirm, is your hand intention area." And so I'll make a little cut in the skin, we'll flap the skin open, just like kind of opening the hood of a car, only a lot smaller.
Paragraph 945
Make a perfectly round one inch diameter hole in the skull, remove that bit of skull, open the lining of the brain, the covering of the brain. It's like a little bag of water that the brain floats in. And then show that part of the brain to our robot. And then this is where the robot shines.
Paragraph 946
It can come in and take these tiny, much smaller than human hair electrodes and precisely insert them into the cortex, into the surface of the brain to a very precise depth, in a very precise spot that avoids all the blood vessels that are coating the surface of the brain.
Paragraph 947
And after the robot's done with its part, then the human comes back in and puts the implant into that hole in the skull and covers it up, screwing it down to the skull and sewing the skin back together. So the whole thing is a few hours long. It's extremely low risk compared to the average neurosurgery involving the brain that might say, open up a deep part of the brain or manipulate blood vessels in the brain.
Paragraph 948
This opening on the surface of the brain, with only cortical micro insertions, carries significantly less risk than a lot of the tumor or aneurysm surgeries that are routinely done.
Paragraph 949
So cortical micro insertions that are via robot and computer vision are designed to avoid the blood vessels.
Paragraph 950
Exactly.
Paragraph 951
So I know you're a bit biased here, but let's compare human and machine.
Paragraph 952
Sure.
Paragraph 953
So what are human surgeons able to do well, and what are robot surgeons able to do well at this stage of our human civilization development?
Paragraph 954
Yeah, yeah, that's a good question. Humans are general purpose machines. We're able to adapt to unusual situations. We're able to change the plan on the fly. I remember well a surgery that I was doing many years ago down in San Diego, where the plan was to open a small hole behind the ear and go reposition a blood vessel that had come to lay on the facial nerve, the trigeminal nerve, the nerve that goes to the face.
Paragraph 955
When that blood vessel lays on the nerve, it can cause just intolerable, horrific shooting pain that people describe like being zapped with a cattle prod. And so the beautiful, elegant surgery is to go move this blood vessel off the nerve. The surgery team, we went in there and started moving this blood vessel and then found that there was a giant aneurysm on that blood vessel that was not easily visible on the pre-op scans.
Paragraph 956
And so the plan had to dynamically change, and the human surgeons had no problem with that. We're trained for all those things. Robots wouldn't do so well in that situation, at least in their current incarnation, fully robotic surgery, like the electrode insertion portion of the Neuralink surgery. It goes according to a set plan.
Paragraph 957
And so the humans can interrupt the flow and change the plan, but the robot can't really change the plan midway through. It operates according to how it was programmed and how it was asked to run. It does its job very precisely, but not with a wide degree of latitude and how to react to changing conditions.
Paragraph 958
So there could be just a very large number of ways that you could be surprised as a surgeon when you enter a situation that could be subtle things that you have to dynamically adjust to.
Paragraph 959
Correct.
Paragraph 960
And robots are not good at that.
Paragraph 961
Currently. I think we are at the dawn of a new era with AI of the parameters for robot responsiveness to be dramatically broadened, right? I mean, you can't look at a self-driving car and say that it's operating under very narrow parameters. If a chicken runs across the road, it wasn't necessarily programmed to deal with that, specifically, but a Waymo or a self-driving Tesla would have no problem reacting to that appropriately. And so surgical robots aren't there yet, but give it time.
Paragraph 962
And then there could be a lot of, sort of into like semi-autonomous possibilities of maybe a robotic surgeon could say, this situation is perfectly familiar, or the situation is not familiar. And in the not familiar case, a human could take over. But basically like be very conservative in saying, "Okay, this for sure has no issues, no surprises, and let the humans deal with the surprises, with the edge cases, all that." That's one possibility.
Paragraph 963
So you think eventually, you'll be out of the job? Well, you being neurosurgeon, your job being neurosurgeon. Humans, there will not be many neurosurgeons left on this earth.
Paragraph 964
I'm not worried about my job in the course of my professional life. I think I would tell my my kids not necessarily to go in this line of work depending on how things look in 20 years.
Paragraph 965
It's so fascinating, 'cause I mean, if I have a line of work, I would say it's programming. And if you ask me like for the last, I don't know, 20 years, what I would recommend for people, I would tell 'em, yeah, go. You will always have a job if you're a programmer, 'cause there's more and more computers and all this kind of stuff and it pays well.
Paragraph 966
But then you realize these large language models come along and they're really damn good at generating code. So overnight, you could be surprised like, "Wow, like what is the contribution of the human really?" But then you start to think, "Okay, it does seem like humans have ability, like you said, to deal with novel situations." In the case of programming, it's the ability to kinda come up with novel ideas to solve problems.
Paragraph 967
It seems like machines aren't quite yet able to do that. And when the stakes are very high, when it's life critical, as it is in surgery, especially in neurosurgery, the stakes are very high for a robot to actually replace a human. But it's fascinating that in this case of Neuralink, there's a human-robot collaboration.
Paragraph 968
Yeah, yeah. I do the parts it can't do, and it does the parts I can't do. And we are friends. (Lex laughing) - I saw that there's a lot of practice going on. So I mean, everything in Neuralink is tested extremely rigorously. But one of the things I saw, that there's a proxy on which the surgeries are performed.
Paragraph 969
Yeah.
Paragraph 970
So this is both for the robot and for the human, for everybody involved in the entire pipeline. What's that like practicing the surgery?
Paragraph 971
It's pretty intense. So there's no analog to this in human surgery. Human surgery is sort of this artisanal craft that's handed down directly from master to pupil over the generations. I mean, literally the way you learn to be a surgeon on humans is by doing surgery on humans.
Paragraph 972
I mean, first, you watch your professors do a bunch of surgery, and then finally, they put the trivial parts of the surgery into your hands, and then the more complex parts. And as your understanding of the point and the purposes of the surgery increases, you get more responsibility in the perfect condition. Doesn't always go well. In Neuralink's case, the approach is a bit different. We, of course, practiced as far as we could on animals.
Paragraph 973
We did hundreds of animal surgeries. And when it came time to do the first human, we had just amazing team of engineers build incredibly lifelike models. One of the engineers, Fran Romano, in particular built a pulsating brain in a custom 3D printed skull that matches exactly the patient's anatomy, including their face and scalp characteristics.
Paragraph 974
And so when I was able to practice that, I mean, it's as close as it really reasonably should get to being the real thing and all the details, including the having a mannequin body attached to this custom head. And so when we were doing the practice surgeries, we'd wheel that body into the CT scanner and take a mock CT scan and wheel it back in and conduct all the normal safety checks verbally.
Paragraph 975
"Stop, this patient, we're confirming his identification, is mannequin number blah, blah, blah." And then opening the brain in exactly the right spot using standard operative neuronavigation equipment, standard surgical drills in the same OR that we do all of our practice surgeries in at Neuralink.
Paragraph 976
And having the skull open and have the brain pulse, which adds a degree of difficulty for the robot to perfectly precisely plan and insert those electrodes to the right depth and location. And so yeah, we kind of broke new ground on how extensively we practiced for this surgery.
Paragraph 977
So there was a historic moment, a big milestone for Neuralink in part for humanity with the first human getting a Neuralink implant in January of this year. Take me through the surgery on Noland. What did it feel like to be part of this?
Paragraph 978
Yeah. Well, we're lucky to have just incredible partners at the Barrow Neurologic Institute. They are, I think, the premier neurosurgical hospital in the world. They made everything as easy as possible for the trial to get going and helped us immensely with their expertise on how to arrange the details. It was a much more high pressure surgery in some ways.
Paragraph 979
I mean, even though the outcome wasn't particularly in question in terms of our participant safety, the number of observers, the number of people, there's conference rooms full of people watching live streams in the hospital, rooting for this to go perfectly, and that just adds pressure that is not typical for even the most intense production neurosurgery. Say, removing a tumor or placing deep brain stimulation electrodes.
Paragraph 980
And it had never been done on a human before. There were unknown unknowns. And so definitely, a moderate pucker factor there for the whole team, not knowing if we were going to encounter, say, a degree of brain movement that was unanticipated or a degree of brain sag that took the brain far away from the skull and made it difficult to insert or some other unknown unknown problem.
Paragraph 981
Fortunately, everything went well and that surgery is one of the smoothest outcomes we could have imagined.
Paragraph 982
Were you nervous? I mean, you're a bit quarterback in the Super Bowl kind of situation.
Paragraph 983
Extremely nervous. Extremely. I was very pleased when it went well and when it was over. Looking forward to number two.
Paragraph 984
Yeah. Even with all that practice, all of that, you've never been in a situation that's still high stakes in terms of people watching. And we should also probably mention, given how the media works, a lot of people, maybe in a dark kind of way, hoping it doesn't go well.
Paragraph 985
Well, I think wealth is easy to hate or envy or whatever. And I think there's a whole industry around driving clicks, and bad news is great for clicks. And so any way to take an event and turn it into bad news is gonna be really good for clicks.
Paragraph 986
It just sucks because I think it puts pressure on people. It discourages people from trying to solve really hard problems, because to solve hard problems, you have to go into the unknown. You have to do things that haven't been done before, and you have to take risks.
Paragraph 987
Yeah.
Paragraph 988
Calculated risks. You have to do all kinds of safety precautions, but risks nevertheless. And I just wish there would be more celebration of that, of the risk taking versus like people just waiting on the sidelines, like waiting for failure, and then pointing out the failure. Yeah, it sucks. But in this case, it's really great that everything went just flawlessly, but it's unnecessary pressure, I would say.
Paragraph 989
Now that there's a human with literal skin in the game, there's a participant whose wellbeing rides on this doing well, you have to be a pretty bad person to be rooting for that to go wrong. And so hopefully, people look in the mirror and realize that at some point.
Paragraph 990
So did you get to actually front row seat like watch the robot work? You get to see the whole thing?
Paragraph 991
Yeah, I mean, because an MD needs to be in charge of all of the medical decision making throughout the process, I unscrubbed from the surgery after exposing the brain and presenting it to the robot and placed the targets on the robot software interface that tells the robot where it's going to insert each thread that was done with my hand on the mouse, for whatever that's worth.
Paragraph 992
So you were the one placing the targets?
Paragraph 993
Yeah.
Paragraph 994
Oh, cool. So like the robot with a computer vision provides a bunch of candidates and you kinda finalize the decision.
Paragraph 995
Right. The software engineers are amazing on this team. And so they actually provided an interface where you can essentially use a lasso tool and select a prime area of brain real estate, and it will automatically avoid the blood vessels in that region and automatically place a bunch of targets. So that allows the human robot operator to select really good areas of brain and make dense applications of targets in those regions, the regions we think are gonna have the most high fidelity representations of finger movements and arm movement intentions.
Paragraph 996
I've seen like images of this. And for me, with OCD, it's for some reason a really pleasant, I think there's a subreddit called oddly satisfying.
Paragraph 997
[Matthew] Yeah, love that subreddit.
Paragraph 998
It's oddly satisfying to see the different target sites avoiding the blood vessels and also maximizing like the usefulness of those locations for the signal. It just feels good. It's like, ah.
Paragraph 999
As a person who has a visceral reaction to the brain bleeding, I can tell you it's extremely satisfying watching the electrodes themselves go into the brain and not cause bleeding.
Paragraph 1000
Yeah, yeah. So you said the feeling was of relief when everything went perfectly.
Paragraph 1001
Yeah.
Paragraph 1002
How deep in the brain can you currently go and eventually go, let's say, on the Neuralink side. It seems the deeper you go in the brain, the more challenging it becomes.
Paragraph 1003
Yeah, so talking broadly about neurosurgery, we can get anywhere. It's routine for me to put deep brain stimulating electrodes near the very bottom of the brain, entering from the top and passing about a two millimeter wire all the way into the bottom of the brain. And that's not revolutionary. A lot of people do that. And we can do that with very high precision. I use a robot from Globus to do that surgery several times a month. It's pretty routine.
Paragraph 1004
What are your eyes in that situation? What are you seeing? What kind of technology can you use to visualize where you are to light your way?
Paragraph 1005
Yeah, so it's a cool process on the software side. You take a preoperative MRI that's extremely high resolution data of the entire brain. You put the patient to sleep, put their head in a frame that holds the skull very rigidly, and then you take a CT scan of their head while they're asleep with that frame on, and then merge the MRI and the CT in software. You have a plan based on the MRI where you can see these nuclei deep in the brain.
Paragraph 1006
You can't see them on CT, but if you trust the merging of the two images, then you indirectly know on the CT where that is. And therefore, indirectly know where in reference to the titanium frame screwed to their head those targets are. And so this is '60s technology to manually compute trajectories, given the entry point and target and dial in some goofy looking titanium actuators with a manual actuators with little tick marks on them.
Paragraph 1007
The modern version of that is to use a robot. Just like a little KUKA arm, you might see it building cars at the Tesla factory.
Paragraph 1008
This small robot arm can show you the trajectory that you intended from the pre-op MRI and establish a very rigid holder through which you can drill a small hole in the skull and pass a small rigid wire deep into that area of the brain that's hollow and put your electrode through that hollow wire and then remove all of that except the electrode. So you end up with the electrode very, very precisely placed far from the skull surface.
Paragraph 1009
Now that's standard technology that's already, been out in the world for a while. Neuralink right now is focused entirely on cortical targets, surface targets because there's no trivial way to get, say, hundreds of wires deep inside the brain without doing a lot of damage. So your question, what do you see? Well, I see an MRI on a screen. I can't see everything that that DBS electrode is passing through on its way to that deep target.
Paragraph 1010
And so it's accepted with this approach that there's gonna be about what one in a hundred patients who have a bleed somewhere in the brain as a result of passing that wire blindly into the deep part of the brain. That's not an acceptable safety profile for Neuralink. We start from the position that we want this to be dramatically maybe two or three orders of magnitude safer than that.
Paragraph 1011
Safe enough really that you or I, without a profound medical problem, might on our lunch break someday say, "Yeah, sure, I'll get that. I'd be meaning to upgrade to the latest version." And so the safety constraints given that are high. And so we haven't settled on a final solution for arbitrarily approaching deep targets in the brain.
Paragraph 1012
It's interesting 'cause like you have to avoid blood vessels somehow. Maybe there's creative ways of doing the same thing, like mapping out high resolution geometry of blood vessels and then you can go in blind. But how do you map out that in a way that's like super stable? There's a lot of interesting challenges there, right?
Paragraph 1013
[Matthew] Yeah.
Paragraph 1014
But there's a lot to do on the surface. Luckily.
Paragraph 1015
Exactly. So we've got vision on the surface. We actually have made a huge amount of progress sewing electrodes into the spinal cord as a potential workaround for a spinal cord injury that would allow a brain-mounted implant to translate motor intentions to a spine-mounted implant that can effect muscle contractions in previously paralyzed arms and legs.
Paragraph 1016
That's mind-blowing. That's just incredible. So like the effort there is to try to bridge the brain to the spinal cord to the peripheral nervous system. So how hard is that to do?
Paragraph 1017
We have that working in very crude forms in animals.
Paragraph 1018
That's amazing.
Paragraph 1019
Yeah, we've done- - So similar to like with Noland, where he's able to digitally move the cursor, here you're doing the same kind of communication but with the actual factors that you have.
Paragraph 1020
Yeah.
Paragraph 1021
[Lex] That's fascinating.
Paragraph 1022
Yeah, so we have anesthetized animals doing grasp and moving their legs in a sort of walking pattern. Again, early days, but the future is bright for this kind of thing. And people with paralysis should look forward to that bright future. They're gonna have options.
Paragraph 1023
Yeah, and there's a lot of sort of intermediate or extra options where you take like an Optimus robot, like the arm, and to be able to control the arm. The fingers and hands of the arm as a prosthetic.
Paragraph 1024
So skeletons are getting better too.
Paragraph 1025
So skeletons. Yeah, so that goes hand in hand. Although I didn't quite understand until thinking about it deeply and doing more research about Neuralink, how much you can do on the digital side. So this digital telepathy, I didn't quite understand that you can really map the intention, as you described in the hand knob area, that you can map the intention. Just imagine it, think about it.
Paragraph 1026
That intention can be mapped to actual action in the digital world. And now more and more, so much can be done in the digital world that it can reconnect you to the outside world. It can allow you to have freedom, have independence if you're a quadriplegic. That's really powerful. Like you can go really far with that.
Paragraph 1027
Yeah, our first participant, he's incredible. He's breaking world records left and right.
Paragraph 1028
And he is having fun with it, it's great. Just going back to the surgery, your whole journey, you mentioned to me offline, you have surgery on Monday. So you're like doing surgery all the time.
Paragraph 1029
Yeah.
Paragraph 1030
Maybe the ridiculous question, what does it take to get good at surgery?
Paragraph 1031
Practice, repetitions. Same with anything else. There's a million ways of people saying the same thing and selling books saying it, but you call it 10,000 hours, you call it, you know, spend some chunk of your life, some percentage of your life focusing on this, obsessing about getting better at it. Repetitions, humility, recognizing that you aren't perfect at any stage along the way. Recognizing you've got improvements to make in your technique.
Paragraph 1032
Being open to feedback and coaching from people with a different perspective on how to do it. And then just the constant will to do better. That fortunately, if you're not a sociopath, I think your patients bring that with them to the office visits every day. They force you to wanna do better all the time.
Paragraph 1033
Yeah, to step up. I mean, it's a real human being, a real human being that you can help.
Paragraph 1034
Yeah.
Paragraph 1035
So every surgery, even if it's the same exact surgery, is there a lot of variability between that surgery and a different person?
Paragraph 1036
Yeah, a fair bit. I mean, a good example for us is the angle of the skull relative to the normal plane of the body axis, of the skull over hand knob is pretty wide variation. I mean, some people have really flat skulls, and some people have really steeply angled skulls over that area. And that has consequences for how their head can be fixed in sort of the frame that we use and how the robot has to approach the skull.
Paragraph 1037
Yeah, people's bodies are built as differently as the people you see walking down the street, as much variability in body shape and size as you see there. We see in brain anatomy and skull anatomy, there are some people who we've had to kind of exclude from our trial for having skulls that are too thick or too thin or scalp that's too thick or too thin.
Paragraph 1038
I think we have like the middle 97% or so of people, but you can't account for all human anatomy variability.
Paragraph 1039
How much like mushiness and messes there, 'cause taking biology classes, the diagrams are always really clean and crisp. Neuroscience, the pictures of neurons are always really nice and vary. But whenever I look at pictures of like real brains, I don't know what is going on.
Paragraph 1040
Yeah.
Paragraph 1041
So how much are biological systems in reality? Like how hard is it to figure out what's going on?
Paragraph 1042
Not too bad. Once you really get used to this, that's where experience and skill and education really come into play is if you stare at a thousand brains, it becomes easier to kind of mentally peel back the, say, for instance, blood vessels that are obscuring the sulci and gyri, kind of the wrinkle pattern of the surface of the brain.
Paragraph 1043
Occasionally, when you're first starting to do this and you open the skull, it doesn't match what you thought you were gonna see based on the MRI. And with more experience, you learn to kind of peel back that layer of blood vessels and see the underlying pattern of wrinkles in the brain and use that as a landmark for where you are.
Paragraph 1044
[Lex] The wrinkles are a landmark? So like- - Yeah. So I was describing hand knob earlier. That's a pattern of the wrinkles in the brain. It's sort of this sort of Greek letter, omega-shaped area of the brain.
Paragraph 1045
So you could recognize the hand knob area. Like if I show you a thousand brains and give you like one minute with each, you'd be like, "Yep, that's that?"
Paragraph 1046
Sure.
Paragraph 1047
And so there is some uniqueness to that area of the brain, like in terms of the geometry, the topology of the thing.
Paragraph 1048
Yeah.
Paragraph 1049
Where is it about in the- - So you have this strip of brain running down the top called the primary motor area. And I'm sure you've seen this picture of the homunculus laid over the surface of the brain, the weird little guy with huge lips and giant hands. That guy sort of lays with his legs up at the top of the brain and face, arm, areas farther down and then some kind of mouth, lip, tongue areas farther down. And so the hand is right in there.
Paragraph 1050
And then the areas that control speech, at least on the left side of the brain in most people are just below that. And so any muscle that you voluntarily move in your body, the vast majority of that references that strip or those intentions come from that strip of brain. And the wrinkle for hand knob is right in the middle of that.
Paragraph 1051
And vision is back here.
Paragraph 1052
Back, yep.
Paragraph 1053
Also, close to the surface?
Paragraph 1054
Vision's a little deeper. And so this gets to your question about how deep can you get to do vision. We can't just do the surface of the brain. We have to be able to go in, not as deep as we have to go for DBS, but maybe a centimeter deeper than we're used to for hand insertions. And so that's work in progress. That's a new set of challenges to overcome.
Paragraph 1055
By the way, you mentioned the Utah array. And I just saw a picture of that, and that thing looks terrifying.
Paragraph 1056
[Matthew] Yeah, bed of nails.
Paragraph 1057
It's because it's rigid. And then if you look at the threads, they're flexible. What can you say that's interesting to you about the flexible, that kind of approach of the flexible threads to deliver the electrodes next to the neurons?
Paragraph 1058
Yeah, I mean, the goal there comes from experience. I mean, we stand on the shoulders of people that made Utah arrays and used Utah arrays for decades before we ever even came along.
Paragraph 1059
Neuralink arose, partly this approach to technology arose out of a need recognized after Utah arrays would fail routinely because the rigid electrodes, those spikes that are literally hammered using an air hammer into the brain, those spikes generate a bad immune response that encapsulates the electrode spikes in scar tissue essentially.
Paragraph 1060
And so one of the projects that was being worked on in the Andersen lab at Caltech when I got there, was to see if you could use chemotherapy to prevent the formation of scar. Things are pretty bad when you're jamming a bed of nails into the brain and then treating that with chemotherapy to try to prevent scar tissue. It's like, maybe we've gotten off track here, guys. Maybe there's a fundamental redesign necessary.
Paragraph 1061
And so Neuralink's approach of using highly flexible, tiny electrodes avoids a lot of the bleeding, avoids a lot of the immune response that ends up happening when rigid electrodes are pounded into the brain. And so what we see is our electrode longevity and functionality and the health of the brain tissue immediately surrounding the electrode is excellent. I mean, it goes on for years now in our animal models.
Paragraph 1062
What do most people not understand about the biology of the brain? We mention the vasculature. That's really interesting.
Paragraph 1063
I think the most interesting maybe underappreciated fact is that it really does control almost everything. I mean, I don't know, for out of the blue example, imagine you want a lever on fertility, you wanna be able to turn fertility on and off. I mean, there are legitimate targets in the brain itself to modulate fertility, say blood pressure. You wanna modulate blood pressure. There are legitimate targets in the brain for doing that.
Paragraph 1064
Things that aren't immediately obvious as brain problems are potentially solvable in the brain. And so I think it's an under-explored area for primary treatments of all the things that bother people.
Paragraph 1065
That's a really fascinating way to look at it. Like there's a lot of conditions we might think have nothing to do with the brain, but they might just be symptoms of something that actually started in the brain. The actual source of the problem. The primary source is something in the brain.
Paragraph 1066
Yeah, not always. I mean, kidney disease is real. But there are levers you can pull in the brain that affect all of these systems.
Paragraph 1067
There's knobs.
Paragraph 1068
Yeah.
Paragraph 1069
On-off switches and knobs in the brain, from which this all originates. Would you have a Neuralink chip implanted in your brain?
Paragraph 1070
Yeah. I think use case right now is use a mouse, right? I can already do that. And so there's no value proposition. On safety grounds alone, sure. I'll do it tomorrow.
Paragraph 1071
You say the use case of the mouse, is it after like researching all this and part of it is just watching Noland have so much fun? If you can get that bits per second look really high with a mouse, like being able to interact, 'cause if you think about it, the way, on the smartphone, the way you swipe, that was transformational how you interact with a thing. It's subtle.
Paragraph 1072
You don't realize it, but to able to touch a phone and to scroll with your finger, that's like, that changed everything. People were sure you need a keyboard to type. There's a lot of HCI aspects to that that changed how we interact with computers. So there could be a certain rate of speed with the mouse that would change everything.
Paragraph 1073
Yeah.
Paragraph 1074
It's like you might be able to just click around a screen extremely fast. And that, I can see myself getting a Neuralink for much more rapid interaction with the digital devices.
Paragraph 1075
Yeah, I think recording speech intentions from the brain might change things as well. The value proposition for the average person, a keyboard is a pretty clunky human interface, requires a lot of training. It's highly variable in the maximum performance that the average person can achieve. I think taking that out of the equation and just having a natural word to computer interface might change things for a lot of people.
Paragraph 1076
It'd be hilarious if that is the reason people do it. Even if you have speech to text, that's extremely accurate, it currently isn't, but say it gotten super accurate, it'd be hilarious if people went for Neuralink just so you avoid the embarrassing aspect of speaking, like looking like a douche bag speaking to your phone in public, which is a real, like that's a real constraint.
Paragraph 1077
Yeah. I mean, with a bone conducting case, that can be an invisible headphone, say, and the ability to think words into software and have it respond to you, that starts to sound sort of like embedded super intelligence. If you can silently ask for the Wikipedia article on any subject and have it read to you, without any observable change happening in the outside world, for one thing, standardized testing is obsolete. (laughs) - Yeah.
Paragraph 1078
If it's done well in the UX side, it could change. I don't know if it transforms society, but it really can create a kind of shift in the way we interact with digital devices in the way that a smartphone did. Just having to look into the safety of everything involved, I would totally try it so it doesn't have to go to some like incredible thing where you have, it connects your vision or to some other, like it connects all over your brain.
Paragraph 1079
That could be like just connecting to the hand knob. You might have a lot of interesting interaction, human-computer interaction possibilities. That's really interesting.
Paragraph 1080
Yeah, and the technology on the academic side is progressing at light speed here. I think there was a really amazing paper out of UC Davis, Sergey Stavisky's lab that basically made a initial solve of speech decode. It was something like 125,000 words that they we're getting with very high accuracy, which is- - So you're just thinking the word?
Paragraph 1081
Yeah.
Paragraph 1082
Thinking the word and you're able to get it?
Paragraph 1083
Yeah.
Paragraph 1084
Oh boy. Like you have to have the intention of speaking it.
Paragraph 1085
Right.
Paragraph 1086
So like do the inner voice. Man, it's so amazing to me that you can do the intention, the signal mapping. All you have to do is just imagine yourself doing it. And if you get the feedback that it actually worked, you can get really good at that. Like your brain will first of all adjust and you develop it like any other skill. Like touch typing, you develop in that same kind of way.
Paragraph 1087
To me, it's just really fascinating to be able to even to play with that. Honestly, like I would get a Neuralink just to be able to play with that. Just to play with the capacity, the capability of my mind to learn this skill. It's like learning the skill of typing and learning the skill of moving a mouse. It's another skill of moving the mouse, not with my physical body, but with my mind.
Paragraph 1088
I can't wait to see what people do with it. I feel like we're cavemen right now. We're like banging rocks with a stick and thinking that we're making music. At some point, when these are more widespread, there's gonna be the equivalent of a piano that someone can make art with their brain in a way that we didn't even anticipate. I'm looking forward to it.
Paragraph 1089
Give it to like a teenager. Like anytime I think I'm good at something, I'll always go to like, I don't know. Even with the bit per second of playing a video game, you realize you give it to a teenager, you've given your link to a teenager, just the large number of them, the kind of stuff, they get good at stuff. They're gonna get like hundreds of bits per second. Even just with the current technology.
Paragraph 1090
Probably. Probably.
Paragraph 1091
'Cause it's also addicting, the number go up aspect of it of like improving and training, 'cause it is almost like a skill. And plus, there's the software on the other end that adapts to you. And especially if the adapting procedure algorithm becomes better and better and better, you're like learning together.
Paragraph 1092
Yeah, we're scratching the surface on that right now. There's so much more to do.
Paragraph 1093
So on the complete other side of it, you have an RFID chip implanted in you.
Paragraph 1094
Yeah.
Paragraph 1095
So I hear, nice.
Paragraph 1096
Little subtle thing.
Paragraph 1097
It's a passive device that you use for unlocking like a safe with top secrets, or what do you use it for? What's the story behind it?
Paragraph 1098
I'm not the first one. There's this whole community of weirdo biohackers that have done this stuff, and I think one of the early use cases was storing private crypto wallet keys and whatever. I dabbled in that a bit and had some fun with it.
Paragraph 1099
You have some bitcoin implanted in your body somewhere. You can't tell where, yeah.
Paragraph 1100
Yeah, actually, yeah. (Lex laughing) It was the modern day equivalent of finding change in the sofa cushions after I put some orphan crypto on there that I thought was worthless and forgot about it for a few years. Went back and found that some community of people loved it and had propped up the value of it. And so it had gone up 50 fold.
Paragraph 1101
Wow.
Paragraph 1102
So there was a lot of change in those cushions. (Lex laughing) - That's hilarious.
Paragraph 1103
But the primary use case was mostly as a tech demonstrator. It has my business card on it. You can scan that in by touching it to your phone. It opens the front door to my house, whatever simple stuff.
Paragraph 1104
It's a cool step. It's a cool leap to implant something in your body. I mean, perhaps, it's a similar leap to a Neuralink because for a lot of people, that kind of notion of putting something inside your body, something electronic inside a biological system is a big leap.
Paragraph 1105
Yeah, we have a kind of a mysticism around the barrier of our skin. We're completely fine with knee replacements, hip replacements, dental implants. But there's a mysticism still around the inviable barrier that the skull represents. And I think that needs to be treated like any other pragmatic barrier. The question isn't, how incredible is it to open the skull? The question is, what benefit can we provide?
Paragraph 1106
So from all the surgeries you done, from everything you understand the brain, how much does neuroplasticity come into play? How adaptable is the brain, for example, just even in the case of healing from surgery or adapting to the post-surgery situation.
Paragraph 1107
The answer that is sad for me and other people of my demographic is that plasticity decreases with age. Healing decreases with age. I have too much gray hair to be optimistic about that. There are theoretical ways to increase plasticity using electrical stimulation. Nothing that is totally proven out as a robust enough mechanism to offer widely to people.
Paragraph 1108
But yeah, I think there's cause for optimism that we might find something useful in terms of, say, an implanted electrode that improves learning.
Paragraph 1109
Certainly, there's been some really amazing work recently from Nicholas Schiff, Jonathan Baker, and others who have a cohort of patients with moderate traumatic brain injury who have had electrodes placed in the deep nucleus in the brain called the centromedian nucleus or just near central media nucleus. And when they apply small amounts of electricity to that part of the brain, it's almost like electronic caffeine.
Paragraph 1110
They're able to improve people's attention and focus. They're able to improve how well people can perform a task. I think in one case, someone who was unable to work after the device was turned on, they were able to get a job.
Paragraph 1111
And that's sort of one of the holy grails for me with Neuralink and other technologies like this is from a purely utilitarian standpoint, can we make people able to take care of themselves and their families economically again? Can we make it so someone who's fully dependent and even maybe requires a lot of caregiver resources, can we put them in a position to be fully independent, taking care of themselves, giving back to their communities?
Paragraph 1112
I think that's a very compelling proposition, and what motivates a lot of what I do and what a lot of the people at Neuralink are working for.
Paragraph 1113
It's just a cool possibility that if you put a Neuralink in there, that the brain adapts, like the other part of the brain adapts too.
Paragraph 1114
Yeah.
Paragraph 1115
And integrates it. The capacity of the brain to do that is really interesting. Probably unknown to the degree to which you can do that, but you're now connecting an external thing to it, especially once it's doing stimulation, like the biological brain and the electronic brain outside of it working together. Like the possibilities there are really interesting. It's still unknown but interesting. It feels like the brain is really good at adapting to whatever.
Paragraph 1116
Yeah.
Paragraph 1117
But of course, it is a system that by itself is already, like everything serves a purpose and so you don't wanna mess with it too much.
Paragraph 1118
Yeah, it's like, eliminating a species from an ecology. You don't know what the delicate interconnections and dependencies are. The brain is certainly a delicate, complex beast. And we don't know every potential downstream consequence of a single change that we make.
Paragraph 1119
Do you see yourself doing, so you mentioned P1, surgeries of P2, P3, P4, P5? Just more and more and more humans.
Paragraph 1120
I think it's a certain kind of brittleness or a failure on the company's side if we need me to do all the surgeries. I think something that I would very much like to work towards is a process that is so simple and so robust on the surgery side that literally anyone could do it. We wanna get away from requiring intense expertise or intense experience to have this successfully done and make it as simple and translatable as possible.
Paragraph 1121
I mean, I would love it if every neurosurgeon on the planet had no problem doing this. I think we're probably far from a regulatory environment that would allow people that aren't neurosurgeons to do this, but not impossible.
Paragraph 1122
All right, I'll sign up for that. Did you ever anthropomorphize the robot R1? Like do you give it a name? Do you see it as like a friend, as like working together with you?
Paragraph 1123
I mean, to a certain degree it's- - Or anatomy who's gonna get the gap.
Paragraph 1124
To a certain degree, yeah, it's complex relationship.
Paragraph 1125
All the good relationships are.
Paragraph 1126
It's funny when, in the middle of the surgery, there's a part of it where I stand basically shoulder to shoulder with the robot. And so if you're in the room reading the body language, that's my brother in arms there. We're working together on the same problem. Yeah, I'm not threatened by it.
Paragraph 1127
Keep telling yourself that. (laughs) How have all the surgeries that you've done over the years, the people you've helped and the stakes, the high stakes that you've mentioned, how has that changed your understanding of life and death?
Paragraph 1128
Yeah. It gives you a very visceral sense, and this makes sound trite, but it gives you a very visceral sense that death is inevitable. On one hand, you are, as a neurosurgeon, you're deeply involved in these like just hard to fathom tragedies: young parents dying, leaving a four-year-old behind say. And on the other hand, it takes the sting out of it a bit because you see how just mind numbingly universal death is.
Paragraph 1129
There's zero chance that I'm going to avoid it. I know techno optimists right now and longevity buffs right now would disagree on that 0. 0% estimate. But I don't see any chance that our generation is going to avoid it. Entropy is a powerful force, and we are very ornate, delicate, brittle DNA machines that aren't up to the cosmic ray bombardment that we're subjected to. So on the one hand, every human that has ever lived died or will die.
Paragraph 1130
On the other hand, it's just one of the hardest things to imagine inflicting on anyone that you love is having them gone. I'm sure you've had friends that aren't living anymore and it's hard to even think about them. And so I wish I had arrived at the point of nirvana where death doesn't have a sting. I'm not worried about it, but I can at least say that I'm comfortable with the certainty of it.
Paragraph 1131
If not, having found out how to take the tragedy out of it when I think about my kids either not having me or me not having them or my wife.
Paragraph 1132
Maybe I have come to accepting intellectual certainty of it, but it may be the pain that comes with losing the people you love, I don't think I've come to understand the existential aspect of it. Like that this is gonna end. And I don't mean like in some trite way. I mean like, it certainly feels like it's not going to end. Like you live life like it's not going to end.
Paragraph 1133
[Matthew] Right.
Paragraph 1134
And the fact that this light that's shining this consciousness is going to no longer be, in one moment, maybe today, it fills me when I really am able to load all that in with Ernest Becker's terror. Like it's a real fear. I think people aren't always honest with how terrifying it is. I think the more you are able to really think through it, the more terrifying it is. It's not such a simple thing. Oh well, it's the way life is.
Paragraph 1135
If you really can load that in, it's hard. But I think that's why the stoics did it, because it like helps you get your shit together and be like, well, the moment, every single moment you're alive is just beautiful. And it's terrifying that it's gonna end, like almost like you're shivering in the cold a child helpless, this kind of feeling.
Paragraph 1136
And then it makes you, when you have warmth, when you have the safety, when you have the love to really appreciate it. I feel like sometimes, in your position, when you mentioned armor, just to see death, it might make you not be able to see that, the finiteness of life, because if you kept looking at that, it might break you.
Paragraph 1137
So it's good to know that you're kind of still struggling with that, that there's the neurosurgeon and then there's a human. And the human is still able to struggle with that and feel the fear of that and the pain of that.
Paragraph 1138
Yeah, it definitely makes you ask the question of how long, how many of these can you see? And not say, "I can't do this anymore." But I mean, you said it well. I think it gives you an opportunity to just appreciate that you're alive today. And I've got three kids and an amazing wife and I'm really happy. Things are good. I get to help on a project that I think matters. I think it moves us forward. I'm a very lucky person.
Paragraph 1139
It's the early steps of a potentially gigantic leap for humanity. It's a really interesting one. And it's cool 'cause like you, you read about all this stuff in history where it's like the early days. I've been reading, before going to the Amazon, I would read about explorers that would go and explore even the Amazon jungle for the first time. Those are the early steps.
Paragraph 1140
Or early steps into space, early steps in any discipline, in physics and mathematics. And it's cool 'cause this is like, on the grand scale, these are the early steps into delving deep into the human brain. So not just observing the brain, but be able to interact with the human brain. It's gonna help a lot of people, but it also might help us understand what the hell's going on in there.
Paragraph 1141
Yeah, I think ultimately, we wanna give people more levers that they can pull, right? Like you wanna give people options. If you can give someone a dial that they can turn on how happy they are, I think that makes people really uncomfortable. But now, talk about major depressive disorder. Talk about people that are committing suicide at an alarming rate in this country. And try to justify that queasiness in that light of you can give people a knob to take away suicidal ideation, suicidal intention. I would give them that knob. I don't know how you justify not doing that.
Paragraph 1142
Yeah, you can think about like all the suffering that's going on in the world. Like every single human being that's suffering right now, it'll be a glowing red dot. The more suffering, the more it's glowing. And you just see the map of human suffering, and any technology that allows you to dim that light of suffering on a grand scale is pretty exciting, because there's a lot of people suffering and most of them suffer quietly. We look away too often, and we should remember those that are suffering, 'cause once again, most of them are suffering quietly.
Paragraph 1143
Well, and on a grander scale, the fabric of society, people have a lot of complaints about how our social fabric is working or not working, how our politics is working or not working. Those things are made of neurochemistry too, in aggregate, right?
Paragraph 1144
Like our politics is composed of individuals with human brains and, the way it works or doesn't work is potentially tunable in the sense that, I don't know, say remove our addictive behaviors or tune our addictive behaviors for social media or our addiction to outrage, our addiction to sharing the most angry political tweet we can find. I don't think that leads to a functional society.
Paragraph 1145
And if you had options for people to moderate that maladaptive behavior, there could be huge benefits to society. Maybe we could all work together a little more harmoniously toward useful ends.
Paragraph 1146
There's a sweet spot, like you mentioned, you don't wanna completely remove all the dark sides of human nature 'cause those kind of are somehow necessary to make the whole thing work. But there's a sweet spot.
Paragraph 1147
Yeah, I agree. We gotta suffer a little, just not so much that you lose hope.
Paragraph 1148
Yeah. We knew all the surgeries you've done. Have you seen consciousness in there ever? Was there like a glowing light?
Paragraph 1149
I have this sense that I never found it. Never removed it, like a dementor in Harry Potter. I have this sense that consciousness is a lot less magical than our instincts wanna claim it is. It seems to me like a useful analog for thinking about what consciousness is in the brain, is that we have a really good intuitive understanding of what it means to, say, touch your skin and know what's being touched.
Paragraph 1150
I think consciousness is just that level of sensory mapping applied to the thought processes in the brain itself. So what I'm saying is consciousness is the sensation of some part of your brain being active. So you feel it working. You feel the part of your brain that thinks of red things or winged creatures or the taste of coffee. You feel those parts of your brain being active the way that I'm feeling my palm being touched, right?
Paragraph 1151
And that sensory system that feels the brain working is consciousness.
Paragraph 1152
That's so brilliant. It's the same way, it's the sensation of touch when you're touching a thing. Consciousness is the sensation of you feeling your brain working, your brain thinking, your brain perceiving.
Paragraph 1153
Which isn't like a warping of space and time or some quantum field effect, right? It's nothing magical. People always wanna ascribe to consciousness something truly different. And there's this awesome long history of people looking at whatever the latest discovery in physics is to explain consciousness, because it's the most magical, the most out there thing that you can think of. And people always wanna do that with consciousness. I don't think that's necessary. It's just a very useful and gratifying way of feeling your brain work.
Paragraph 1154
And as we said, it's one heck of a brain.
Paragraph 1155
[Matthew] Yeah.
Paragraph 1156
Everything we see around us, everything we love, everything that's beautiful, it came from brains like these.
Paragraph 1157
It's all electrical activity happening inside your skull.
Paragraph 1158
And I, for one, am grateful that there's people like you that are exploring all the ways that it works and all the ways it can be made better.
Paragraph 1159
Thanks, Lex.
Paragraph 1160
Thank you so much for talking today.
Paragraph 1161
It's been a joy. Thanks for listening to this conversation with Matthew MacDougall. And now, dear friends, here's Bliss Chapman, Brain Interface Software lead at Neuralink. You told me that you've met hundreds of people with spinal cord injuries or with ALS and that your motivation for helping at Neuralink is grounded in wanting to help them. Can you describe this motivation?
Paragraph 1162
Yeah. First, just a thank you to all the people I've gotten a chance to speak with, for sharing their stories with me. I don't think there's any world really in which I can share their stories as powerful way as they can.
Paragraph 1163
But just I think to summarize at a very high level what I hear over and over again is that people with ALS or severe spinal cord injury in a place where they basically can't move physically anymore, really at the end of the day are looking for independence. And that can mean different things for different people.
Paragraph 1164
For some folks, it can mean the ability just to be able to communicate again independently without needing to wear something on their face, without needing a caretaker to be able to put something in their mouth.
Paragraph 1165
For some folks, it can mean independence to be able to work again, to be able to navigate a computer digitally, efficiently enough to be able to get a job, to be able to support themself, to be able to move out and ultimately be able to support themself after their family maybe isn't there anymore to take care of them.
Paragraph 1166
And for some folks, it's as simple as just being able to respond to their kid in time before they run away or get interested in something else. And these are deeply personal and sort of very human problems. And what strikes me again and again when talking with these folks is that this is actually an engineering problem. This is a problem that with the right resources, with the right team, we can make a lot of progress on.
Paragraph 1167
And at the end of the day, I think that's a deeply inspiring message and something that makes me excited to get up every day.
Paragraph 1168
So it's both an engineering problem in terms of a BCI, for example, that can give them capabilities where they can interact with the world. But also on the other side, it's an engineering problem for the rest of the world to make it more accessible for people living with quadriplegia.
Paragraph 1169
Yeah, and actually, I'll take a broad view sort of lens on this for a second. I think I'm very in favor of anyone working in this problem space. So beyond BCI, I'm happy and excited and willing to support any way I can folks working on eye tracking systems, working on speech to text systems, working on head trackers or mouse sticks or quad sticks. And I've met many engineers and folks in the community that do exactly those things.
Paragraph 1170
And I think for the people we're trying to help, it doesn't matter what the complexity of the solution is as long as the problem is solved. And I wanna emphasize that there can be many solutions out there that can help with these problems. And BCI is one of a collection of such solutions. So BCI, in particular, I think offers several advantages here.
Paragraph 1171
And I think the folks that recognize this immediately are usually the people who have spinal cord injury or some form of paralysis. Usually, you don't have to explain to them why this might be something that could be helpful. It's usually pretty self-evident.
Paragraph 1172
But for the rest of us folks that don't live with severe spinal cord injury or who don't know somebody with ALS, it's not often obvious why you would want a brain implant to be able to connect and navigate a computer. And it's surprisingly nuanced, and to the degree that I've learned a huge amount just working with Noland in the first Neuralink clinical trial and understanding from him in his words why this device is impactful for him.
Paragraph 1173
And it's a nuanced topic. It can be the case that even if you can achieve the same thing, for example, with a mouse stick when navigating a computer, he doesn't have access to that mouse stick every single minute of the day. He only has access when someone's available to put it in front of him.
Paragraph 1174
And so a BCI can really offer a level of independence and autonomy that if it wasn't literally physically part of your body, it'd be hard to achieve in any other way.
Paragraph 1175
So there's a lot of fascinating aspects to what it takes to get Noland to be able to control a cursor on the screen with his mind. You texted me something that I just love. You said, "I was part of the team that interviewed and selected P1. I was in the operating room during the first human surgery monitoring live signals coming out of the brain. I work with the user basically every day to develop new UX paradigm's decoding strategies.
Paragraph 1176
And I was part of the team that figured out how to recover useful BCI to new world record levels when the signal quality degraded." We'll talk about I think every aspect of that, but just zooming out, what was it like to be part of that team and part of that historic, I would say, historic first?
Paragraph 1177
Yeah, I think for me, this is something I've been excited about for close to 10 years now. And so to be able to be even just some small part of making it a reality is extremely exciting. A couple maybe special moments during that whole process that I'll never really truly forget, one of them is during the actual surgery, at that point in time, I know Noland quite well. I know his family.
Paragraph 1178
And so I think the initial reaction when Noland is rolled into the operating room is just a "oh shit" kind of reaction. But at that point, muscle memory kicks in and you sort of go into, you let your body just do all the talking. And I have the lucky job in that particular procedure to just be in charge of monitoring the implant.
Paragraph 1179
So my job is to sit there, to look at the signals coming off the implant, to look at the live brain data streaming off the device as threads are being inserted into the brain and just to basically observe and make sure that nothing is going wrong or that there's no red flags or fault conditions that we need to go and investigate or pause the surgery to debug.
Paragraph 1180
And because I had that sort of spectator view of the surgery, I had a slightly removed perspective than I think most folks in the room. I got to sit there and think to myself, "Wow, that brain is moving a lot." When you look into the side look craniectomy, that we stick the threads in, one thing that most people don't realize is the brain moves. The brain moves a lot when you breathe, when your heart beats, and you can see it visibly.
Paragraph 1181
So that's something that I think was a surprise to me and very, very exciting to be able to see someone's brain who you physically know and have talked with at length actually pulsing and moving inside their skull.
Paragraph 1182
And they use that brain to talk to you previously, and now it's right there moving.
Paragraph 1183
Yep.
Paragraph 1184
Actually, I didn't realize that in terms of the thread sending, so the Neuralink implant is active during surgery, and one thread at a time, you're able to start seeing the signal?
Paragraph 1185
Yeah.
Paragraph 1186
So that's part of the way you test that the thing is working?
Paragraph 1187
Yeah, so actually in the operating room, right after we sort of finished all the thread insertions, I started collecting what's called broadband data. So broadband is basically the most raw form of signal you can collect from a Neuralink electrode. It's essentially a measurement of the local field potential or the, yeah, the voltage essentially measured by that electrode.
Paragraph 1188
And we have a certain mode in our application that allows us to visualize where detected spikes are. So it visualizes sort of where, in the broadband symbol, and it's very, very raw form of the data a neuron is actually spiking.
Paragraph 1189
And so one of these moments that I'll never forget as part of this whole clinical trial is seeing live in the operating room, while he's still under anesthesia, beautiful spikes being shown in the application, just streaming live to a device I'm holding in my hand.
Paragraph 1190
So this is no signal processing the raw data and then the signals processings on top of it, you're seeing the spikes detected?
Paragraph 1191
[Bliss] Right, yeah.
Paragraph 1192
And that's a UX too because- - Yes.
Paragraph 1193
That looks beautiful as well.
Paragraph 1194
During that procedure, there was actually a lot of cameramen in the room. So they also were curious and wanted to see. There's several neurosurgeons in the room who are all just excited to see robots taking their job and they're all crowded around a small little iPhone watching this live brain data stream out of his brain.
Paragraph 1195
What was that like seeing the robot do some of the surgery? So the computer vision aspect where it detects all the spots that avoid the blood vessels and then obviously with the human supervision, then actually doing the really high precision connection of the threads to the brain.
Paragraph 1196
That's a good question. My answer's gonna be pretty lame here, but it was boring. I've seen it so many times. Yeah, that's exactly how you want surgery to be. You want it to be boring, because I've seen it so many times. I've seen the robot do the surgery literally hundreds of times, and so it was just one more time.
Paragraph 1197
Yeah, all the practice surgeries and the proxies and this is just another day.
Paragraph 1198
Yep.
Paragraph 1199
So what about when Noland woke up? Do you remember a moment where he was able to move the cursor, not move the cursor, but get signal from the brain such that it was able to show that there's a connection?
Paragraph 1200
Yeah, yeah. So we are quite excited to move as quickly as we can, and Noland was really, really excited to get started. He wanted to get started actually the day of surgery, but we waited till the next morning very patiently. So a long night. And the next morning in the ICU, where he was recovering, he wanted to get started and actually start to understand what kind of signal we can measure from his brain.
Paragraph 1201
And maybe for folks who are not familiar with the Neuralink system, we implant the Neuralink system or the Neuralink implant in the motor cortex. So the motor cortex is responsible for representing things like motor intent, sort of if you imagine closing and opening your hand, that kind of signal representation would be present in the motor cortex.
Paragraph 1202
If you imagine moving your arm back and forth or wiggling a pinky, this sort of signal can be present in the motor cortex. So one of the ways we start to sort of map out, what kind of signal do we actually have access to in any particular individual's brain is through this task called body mapping. And body mapping is where you essentially present a visual to the user and you say, "Hey, imagine doing this."
Paragraph 1203
And that visual is a 3D hand opening and closing, or index finger modulating up and down. And you ask the user to imagine that, and obviously, you can't see them do this, 'cause they're paralyzed so you can't see them actually move their arm, but while they do this task, you can record neural activity, and you can basically offline model and check, can I predict or can I detect the modulation corresponding with those different actions?
Paragraph 1204
And so we did that task and we realized, hey, there's actually some modulation associated with some of his hand motion, which was the first indication that, okay, we can potentially use that modulation to do useful things in the world. For example, control a computer cursor. And he started playing with it, the first time we showed him it, and we actually just took the same live view of his brain activity and put it in front of him.
Paragraph 1205
And we said, "Hey, you tell us what's going on. We're not you. You're able to imagine different things, and we know that it's modulating some of these neurons so you figure out for us what that is actually representing." And so he played with it for a bit. He was like, "I don't quite get it yet." He played for a bit longer. And he said, "Oh, when I move this finger, I see this particular neuron start to fire more."
Paragraph 1206
And I said, okay, "Prove it, do it again." And so he said, "Okay, three, two, one, boom." And the minute he moved, you can see like instantaneously this neuron is firing - single neuron. I can tell you the exact channel number if you're interested. It's stuck in my brain now forever.
Paragraph 1207
But that single channel firing was a beautiful indication that it was behaved really modulated neural activity that could then be used for downstream tasks like decoding a computer cursor.
Paragraph 1208
And when you say single channel, is that associated with a single electrode?
Paragraph 1209
Yeah, channel electrode are interchangeable.
Paragraph 1210
And there's 1,024 of those?
Paragraph 1211
1,024, yeah.
Paragraph 1212
That's incredible that that works.
Paragraph 1213
When I was learning about all this and like loading it in, it was just blowing my mind that the intention, you can visualize yourself moving the finger, that can turn into a signal, and the fact that you can then skip that step and visualize the cursor moving or have the intention of the cursor moving and that leading to a signal that can then be used to move the cursor, there is so many exciting things there to learn about the brain, about the way the brain works, the very fact of their existing signal that can be used is really powerful.
Paragraph 1214
But it feels like that's just like the beginning of figuring out how that signal could be used really, really effectively. I should also just, there's so many fascinating details here, but you mentioned the body mapping step. At least in the version I saw that Noland was showing off, there's like a super nice interface, like a graphical interface.
Paragraph 1215
But like it just felt like I was like in the future 'cause it like, you know, I guess it visualizes you moving the hand. And there's very like a sexy, polished interface. Hello. I don't know if there's a voice component, but it just felt like when you wake up in a really nice video game and this is a tutorial at the beginning of that video game. "This is what you're supposed to do." It's cool.
Paragraph 1216
No, I mean, the future should feel like the future.
Paragraph 1217
But it's not easy to pull that off. I mean, it needs to be simple but not too simple.
Paragraph 1218
Yeah, and I think the UX design component here is underrated for BCI development in general. There's a whole interaction effect between the ways in which you visualize an instruction to the user and the kinds of signal you can get back. And that quality of sort of your behavioral alignment to the neural signal is a function of how good you are at expressing to the user what you want them to do.
Paragraph 1219
And so yeah, we spend a lot of time thinking about the UX, of how we build our applications, of how the decoder actually functions, the control surfaces it provides to the user. All these little details matter a lot.
Paragraph 1220
So maybe it'd be nice to get into a little bit more detail of what the signal looks like and what the decoding looks like. So there's a N1 implant that has, like we mentioned, 1,024 electrodes and that's collecting raw data, raw signal. What does that signal look like, and what are the different steps along the way before it's transmitted, and what is transmitted, all that kind of stuff?
Paragraph 1221
Yeah, yep. This is gonna be a fun one. Let's go. So maybe before diving into what we do, it's worth understanding what we're trying to measure, because that dictates a lot of the requirements for the system that we build. And what we're trying to measure is really individual neurons producing action potentials. And action potential is, you can think of it like a little electrical impulse that you can detect if you're close enough.
Paragraph 1222
And by being close enough, I mean, like within let's say 100 microns of that cell. And 100 microns is a very, very tiny distance. And so the number of neurons that you're gonna pick up with any given electrode is just a small radius around that electrode. And the other thing worth understanding about the underlying biology here is that when neurons produce an action potential, the width of that action potential is about one millisecond.
Paragraph 1223
So from the start of the spike to the end of the spike, that whole width of that sort of characteristic feature of a neuron firing is one millisecond wide. And if you want to detect that an individual spike is occurring or not, you need to sample that signal or sample the local full potential nearby that neuron much more frequently than once a millisecond.
Paragraph 1224
You need to sample many, many times per millisecond to be able to detect that this is actually the characteristic waveform of a neuron producing an action potential. And so we sample across all 1,024 electrodes about 20,000 times a second. 20,000 times a second means we've already given one millisecond window. We have about 20 samples that tell us what that exact shape of that action potential looks like.
Paragraph 1225
And once we've sort of sampled at super high rate the underlying electrical field nearby these cells, we can process that signal into just where do we detect a spike or where do we not, sort of a binary signal one or zero. Do we detect a spike in this one millisecond or not? And we do that because the actual information carrying sort of subspace of neural activity is just when are spikes occurring.
Paragraph 1226
Essentially, everything that we care about for decoding can be captured or represented in the frequency characteristics of spike trains, meaning how often are spikes firing in any given window of time.
Paragraph 1227
And so that allows us to do sort of a crazy amount of compression from this very rich, high density signal to something that's much, much more sparse and compressible that can be sent out over a wireless radio, like a Bluetooth communication, for example.
Paragraph 1228
Quick tangents here. You mentioned electrode neuron. There's a local neighborhood of neurons nearby. How difficult is it to like isolate from where the spike came from?
Paragraph 1229
Yeah, so there's a whole field of sort of academic neuroscience work on exactly this problem, of basically given a single electrode or given a set of electrodes measuring a set of neurons, how can you sort of sort, spike sort which spikes are coming from what neuron? And this is a problem that's pursued in academic work because you care about it for understanding what's going on in the underlying sort of neuroscience of the brain.
Paragraph 1230
If you care about understanding how the brain's representing information, how that's evolving through time, then that's a very, very important question to understand. For sort of the engineering side of things, at least at the current scale, if the number of neurons per electrode is relatively small, you can get away with basically ignoring that problem completely.
Paragraph 1231
You can think of it like sort of a random projection of neurons to electrodes, and there may be in some cases more than one neuron per electrode. But if that number is small enough, those signals can be thought of as sort of a union of the two. And for many applications, that's a totally reasonable trade off to make and can simplify the problem a lot.
Paragraph 1232
And as you sort of scale out channel count, the relevance of distinguishing individual neurons becomes less important, because you have more overall signal and you can start to rely on sort of correlations or covariance structure in the data to help understand when that channel's firing, what does that actually represent? 'Cause you know that when that channel's firing in concert with these other 50 channels, that means move left.
Paragraph 1233
But when that same channel's firing with concert with these other 10 channels, that means move right.
Paragraph 1234
Okay, so you have to do this kind of spike detection on board, and you have to do that super efficiently, so fast and not use too much power, 'cause you don't wanna be generating too much heat. So it has to be a super simple signal processing step.
Paragraph 1235
[Bliss] Yeah.
Paragraph 1236
Is there some wisdom you can share about what it takes to overcome that challenge?
Paragraph 1237
Yeah, so we've tried many different versions of basically turning this raw signal into sort of a feature that you might wanna send off the device. And I'll say that I don't think we're at the final step of this process. This is a long journey. We have something that works clearly today, but there can be many approaches that we find in the future that are much better than what we do right now.
Paragraph 1238
So some versions of what we do right now, and there's a lot of academic heritage to these ideas, so I don't wanna claim that these are original Neuralink ideas or anything like that. But one of these ideas is basically to build a sort of like a convolutional filter almost, if you will, that slides across the signal and looks for a certain template to be matched.
Paragraph 1239
And that template consists of sort of how deep the spike modulates, how much it recovers, and what the duration and window of time is that the whole process takes. And if you can see in the signal that that template is matched within certain bounds, then you can say, "Okay, that's a spike."
Paragraph 1240
One reason that approach is super convenient is that you can actually implement that extremely efficiently in hardware, which means that you can run it in low power across 1,024 channels all at once. Another approach that we've recently started exploring, and this can be combined with the spike detection approach, something called spike band power.
Paragraph 1241
And the benefits of that approach are that you may be able to pick up some signal from neurons that are maybe too far away to be detected as a spike, because the farther away you are from an electrode, the weaker that actual spike waveform will look like on that electrode.
Paragraph 1242
So you might be able to pick up population level activity of things that are maybe slightly outside the normal recording radius, what neuroscientists sometimes refer to as the hash of activity, the other stuff that's going on, and you can look at sort of across many channels how that sort of background noise is behaving and you might be able to get more juice out of the signal that way. But it comes at a cost.
Paragraph 1243
That signal is now a floating point representation, which means it's more expensive to send out over a power. It means you have to find different ways to compress it that are different than what you can apply to binary signals. So there's a lot of different challenges associated with these different modalities.
Paragraph 1244
So also, in terms of communication, you're limited by the amount of data you can send.
Paragraph 1245
[Bliss] Yeah.
Paragraph 1246
And also, because you're currently using the Bluetooth protocol, you have to batch stuff together. But you have to also do this keeping the latency crazy low. Like crazy low. Anything to say about the latency?
Paragraph 1247
Yeah, this is a passion project of mine, so I wanna build the best mouse in the world. I don't wanna build like the, you know, the Chevrolet Spark or whatever of electric cars. I wanna build like the Tesla Roadster version of a mouse. And I really do think it's quite possible that within 5 to 10 years, that most eSports competitions are dominated by people with paralysis. This is like a very real possibility for number of reasons.
Paragraph 1248
One is that they'll have access to the best technology to play video games effectively. The second is they have the time to do so. So those two factors together are particularly potent for eSport competitors.
Paragraph 1249
Unless people without paralysis are also allowed to implant.
Paragraph 1250
(laughs) Right.
Paragraph 1251
Which is, it is another way to interact with a digital device. And there's something to that, if it's a fundamentally different experience, more efficient experience. Even if it's not like some kinda full on high bandwidth communication, if it's just the ability to move the mouse 10x faster, like the bits per second, if I can achieve a bits per second, that 10x, what I can do with the mouse, that's a really interesting possibility of what that can do, especially as you get really good at it with training.
Paragraph 1252
It's definitely the case that you have a higher ceiling performance, because you don't have to buffer your intention through your arm, through your muscle. You get just, by nature of having a brain implant at all, like 75 millisecond lead time on any action that you're actually trying to take.
Paragraph 1253
And there's some nuance to this, like there's evidence that the motor cortex, you can sort of plan out sequences of action so you may not get that whole benefit all the time. But for sort of like reaction time style games where you just wanna, somebody's over here, snipe 'em, that kind of thing. You actually do have just an inherent advantage 'cause you don't need to go through muscle. So the question is, just how much faster can you make it?
Paragraph 1254
And we're already than you what you would do if you're going through muscle from a latency point of view, and we're in the early stage of that. I think we can push it sort of our end-to-end latency right now from brain spike to cursor movement, it's about 22 milliseconds. If you think about the best mice in the world, the best gaming mice, that's about five milliseconds-ish of latency, depending on how you measure.
Paragraph 1255
Depending how fast your screen refreshes, there's a lot of characteristics that matter there. But yeah, and the rough time for like a neuron in the brain to actually impact your command of your hand is about 75 millisecond. So if you look at those numbers, you can see that we're already like competitive and slightly faster than what you'd get by actually moving your hand.
Paragraph 1256
And this is something that, if you ask Noland about it, when he moved the cursor for the first time, we asked him about this. This was something I was super curious about, like what does it feel like when you're modulating, a click intention or when you're trying to just move the cursor to the right.
Paragraph 1257
He said it moves before he is like actually intending it to, which is kind of a surreal thing and something that I would love to experience myself one day. What is that like to have that thing just be so immediate, so fluid that it feels like it's happening before you're actually intending it to move.
Paragraph 1258
Yeah, I suppose we've gotten used to that latency, that natural latency that happens. So is the currently the bottleneck, the communication, so like the Bluetooth communication, what's the actual bottleneck? I mean, there's always gonna be a bottleneck. What's the current bottleneck?
Paragraph 1259
Yeah, a couple things. So kind of hilariously, Bluetooth low energy protocol has some restrictions on how fast you can communicate. So the protocol itself establishes a standard of the most frequent sort of updates you can send are on the order of 7. 5 milliseconds.
Paragraph 1260
And as we push latency down to the level of sort of individual spikes impacting control, that level of resolution, that kind of protocol is gonna become a limiting factor at some scale. Another sort of important nuance to this is that it's not just the Neuralink itself that's part of this equation. If you start pushing latency sort of below the level of how fast screens refresh, then you have another problem.
Paragraph 1261
Like you need your whole system to be able to be as reactive as the sort of limits of what the technology can offer. Like you need the screen like 120 hertz just doesn't work anymore if you're trying to have something respond at something that's at the level of one millisecond.
Paragraph 1262
That's a really cool challenge. I also like that for a T-shirt, the best mouse in the world. Tell me on the receiving end, so the decoding step, now we figured out what the spikes are, we got them all together, now we're sending that over to the app. What's the decoding step look like?
Paragraph 1263
Yeah, so maybe first, what is decoding? I think there's probably a lot of folks listening that just have no clue what it means to decode brain activity.
Paragraph 1264
Actually, even if we zoom out beyond that, what is the app? So there's an implant that's wirelessly communicating with any digital device that has an app installed. So maybe can you tell me a high level what the app is, what the software is outside of the brain?
Paragraph 1265
Yeah, so maybe working backwards from the goal, the goal is to help someone with paralysis, in this case Noland, be able to navigate his computer independently. And we think the best way to do that is to offer them the same tools that we have to navigate our software because we don't wanna have to rebuild an entire software ecosystem for the brain. At least not yet. Maybe someday you can imagine there's UXs that are built natively for BCI.
Paragraph 1266
But in terms of what's useful for people today, I think most people would prefer to be able to just control mouse and keyboard inputs to all the applications that they wanna use for their daily jobs, for communicating with their friends, et cetera. And so the job of the application is really to translate this wireless stream of brain data coming off the implant into control of the computer.
Paragraph 1267
And we do that by essentially building a mapping from brain activity to sort of the HID inputs to the actual hardware. So HID is just the protocol for communicating like input device events. So for example, move mouse to this position or press this key down. And so that mapping is fundamentally what the app is responsible for.
Paragraph 1268
But there's a lot of nuance of how that mapping works that we spend a lot of time to try to get right and we're still in the early stages of a long journey to figure out how to do that optimally. So one part of that process is decoding. So decoding is this process of taking the statistical patterns of brain data that's being channeled across this Bluetooth connection to the application and turning it into, for example, a mouse movement.
Paragraph 1269
And that decoding step, you can think of it in a couple different parts. So similar to any machine learning problem, there's a training step and there's an inference step. The training step in our case is a very intricate behavioral process where the user has to imagine doing different actions. So for example, they'll be presented a screen with a cursor on it and they'll be asked to push that cursor to the right.
Paragraph 1270
Then imagine pushing that cursor to the left, push it up, push it down, and we can basically build up a pattern or using any sort of modern ML method, a mapping of given this brain data and that imagined behavior map one to the other. And then at test time, you take that same pattern matching system.
Paragraph 1271
In our case, it's a deep neural network, and you run it and you take the live stream of brain data coming off their implant, you decode it by pattern matching to what you saw at calibration time, and you use that for a control of the computer.
Paragraph 1272
Now, a couple like sort of rabbit holes that are I think are quite interesting, one of them has to do with how you build that best template matching system because there's a variety of behavioral challenges and also debugging challenges when you're working with someone who's paralyzed. Because again, fundamentally, you don't observe what they're trying to do. You can't see them attempt to move their hand.
Paragraph 1273
And so you have to figure out a way to instruct the user to do something and validate that they're doing it correctly such that then you can downstream, build with confidence the mapping between the neural spikes and the intended action. And by doing the action correctly, what I really mean is at this level of resolution of what neurons are doing.
Paragraph 1274
So if in ideal world, you could get a signal of behavioral intent that is ground truth accurate at the scale of sort of one millisecond resolution, then with high confidence, I could build a mapping from my neuro spikes to that behavioral intention. But the challenge is, again, that you don't observe what they're actually doing.
Paragraph 1275
And so there's a lot of nuance to how you build user experiences that give you more than just sort of a course on average correct representation of what the user's intending to do. If you want to build the world's best mouse, you really want it to be as responsive as possible.
Paragraph 1276
You want it to be able to do exactly what the user's intending at every sort of step along the way, not just on average be correct when you're trying to move it from left to right. And building a behavioral sort of calibration game or sort of software experience that gives you that level of resolution is what we spend a lot of time working.
Paragraph 1277
So the calibration process, the interface has to encourage precision, meaning like whatever it does, it should be super intuitive that the next thing the human is going to likely do is exactly that intention that you need and only that intention. And you don't have any feedback except that may be speaking to you afterwards what they actually did. You can't, "Oh yeah."
Paragraph 1278
Right.
Paragraph 1279
So that's fundamentally, that is a really exciting UX challenge, 'cause that's all on the UX. It's not just about being friendly or nice or usable.
Paragraph 1280
Yeah.
Paragraph 1281
It's like- - User experience is how it works.
Paragraph 1282
It's how it works.
Paragraph 1283
Yeah.
Paragraph 1284
For the calibration, and calibration, at least at this stage of Neuralink, is like fundamental to the operation of the thing and not just calibration but continued calibration essentially.
Paragraph 1285
[Bliss] Yeah.
Paragraph 1286
Wow, yeah.
Paragraph 1287
You said something that I think is worth exploring there a little bit.
Paragraph 1288
You said it's primarily a UX challenge, and I think a large component of it is, but there is also a very interesting machine learning challenge here, which is given some data set, including some on average correct behavior of asking the user to move up or move down, move right, move left, and given a data set of neural spikes, is there a way to infer in some kind of semi-supervised or entirely unsupervised way what that high resolution version of their intention is?
Paragraph 1289
And if you think about it, like there probably is because there are enough data points in the dataset, enough constraints on your model that there should be a way with the right sort of formulation to let the model figure out itself. For example, at this millisecond, this is exactly how hard they're pushing upwards. And at this millisecond, this is how hard they're trying to push upwards.
Paragraph 1290
It's really important to have very clean labels, yes. So like the problem becomes much harder from the machine learning perspective if the labels are noisy.
Paragraph 1291
[Bliss] That's correct.
Paragraph 1292
And then to get the clean labels, that's a UX challenge.
Paragraph 1293
Correct, although clean labels, I think maybe it's worth exploring what that exactly means. I think any given labeling strategy will have some number of assumptions it makes about what the user's attempting to do. Those assumptions can be formulated in a loss function, or they can be formulated in terms of heuristics that you might use to just try to estimate or guesstimate what the user's trying to do.
Paragraph 1294
And what really matters is how accurate are those assumptions. For example, you might say, "Hey, user, push upwards and follow the speed of this cursor," and your heuristic might be that they're trying to do it exactly what that cursor's trying to do. Another competing heuristic might be they're actually trying to go slightly faster at the beginning of the movement and slightly slower at the end.
Paragraph 1295
And those competing heuristics may or may not be accurate reflections of what the user's trying to do. Another version of the task might be, "Hey, user, imagine moving this cursor a fixed offset. So rather than follow the cursor, just try to move it exactly 200 pixels to the right." So here's the cursor, here's the target. Okay, cursor disappears. Try to move that now invisible cursor 200 pixels to the right.
Paragraph 1296
And the assumption in that case would be that the user can actually modulate correctly that position offset, but that position offset assumption might be a weaker assumption, and therefore, potentially you can make it more accurate than these heuristics that are trying to guesstimate at each millisecond what the user's trying to do.
Paragraph 1297
So you can imagine different tasks that make different assumptions about the nature of the user intention and those assumptions being correct is what I would think of as a clean label.
Paragraph 1298
For that step, what are we supposed to be visualizing? There's a cursor and you wanna move that cursor to the right or the left or up and down or maybe move them by a certain offset. So that's one way, is that the best way to do calibration? So for example, an alternative crazy way that probably is playing a role here is a game like Webgrid, where you're just getting a very large amount of data, the person playing a game, where if they're in a state of flow, maybe you can get clean signal as a side effect.
Paragraph 1299
[Bliss] Yep.
Paragraph 1300
Is that not an effective way for initial calibration?
Paragraph 1301
Yeah, great question. There's a lot to unpack there. So the first thing I would draw a distinction between a sort of open loop, first closed loop. So open loop, what I mean by that is the user is sort of going from zero to one. They have no model at all, and they're trying to get to the place where they have some level of control at all.
Paragraph 1302
In that setup, you really need to have some task that gives the user a hint of what you want them to do such that you can build this mapping again from brain data to output. Then once they have a model, you could imagine them using that model and actually adapting to it and figuring out the right way to use it themself and then retraining on that data to give you sort of a boost in performance.
Paragraph 1303
There's a lot of challenges associated with both of these techniques and we can sort of rabbit hole into both of 'em, if you're interested. But the sort of challenge with the open loop task is that the user themself doesn't get proprioceptive feedback about what they're doing. They don't necessarily perceive themself or feel the mouse under their hand when they're using an open, when they're trying to do an open loop calibration.
Paragraph 1304
They're being asked to perform something. Like imagine if you sort of had your whole right arm numbed and you stuck it in a box and you couldn't see it. So you had no visual feedback and you had no proprioceptive feedback about what the position or activity of your arm was. And now you're asked, okay, given this thing on the screen that's moving from left to right, match that speed.
Paragraph 1305
And you basically can try your best to invoke whatever that imagined action is in your brain that's moving the cursor from left to right. But in any situation, you're gonna be inaccurate and maybe inconsistent in how you do that task. And so that's sort of the fundamental challenge of open loop.
Paragraph 1306
The challenge with closed loop is that, once the user's given a model, and they're able to start moving the mouse on their own, they're going to very naturally adapt to that model. And that co-adaptation between the model learning, what they're doing, and the user learning how to use the model may not find you the best sort of global minima.
Paragraph 1307
And maybe that your first model was noisy in some ways or maybe just had some like quirk, like if there's some like part of the data distribution that didn't cover super well, and the user now figures out because they're a brilliant user like Noland. They figured out the right sequence of imagined motions or the right angle they have to hold their hand at to get it to work.
Paragraph 1308
And they'll get it to work great, but then the next day, they come back to their device and maybe they don't remember exactly all the tricks that they used the previous day. And so there's a complicated sort of feedback cycle here that can emerge and can make it a very, very difficult debugging process.
Paragraph 1309
Okay, there's a lot of really fascinating things there. Yeah, actually, just to stay on the closed loop, I've seen situations, this actually happened watching psychology grad students. They use piece of software when they don't know how to program themselves. They use piece of software that somebody else wrote, and it has a bunch of bugs. And they figure out like, and they've been using it for years. They figured out ways to work around it.
Paragraph 1310
Oh, that just happens. Like nobody like considers maybe we should fix this. They just adapt. And that's a really interesting notion, that we were really good at adapting, but you need to still, that might not be the optimal. Okay, so how do you solve that problem? Do you have to restart from scratch every once in a while kind of thing?
Paragraph 1311
Yeah, it's a good question. First and foremost, I would say this is not a solved problem. And for anyone who's listening in academia who works on BCIs, I would also say this is not a problem that's solved by simply scaling channel account. Maybe that can help and you can get sort of richer covariance structures that you can use to exploit when trying to come up with good labeling strategies.
Paragraph 1312
But if you're interested in problems, that aren't gonna be solved inherently by scaling channel account, this is one of them. Yeah, so how do you solve it? It's not a solved problem. That's the first thing I wanna make sure gets across. The second thing is, any solution that involves closed loop is going to become a very difficult debugging problem.
Paragraph 1313
And one of my sort of general heuristics for choosing what prompts to tackle is that you wanna choose the one that's gonna be the easiest to debug, 'cause if you can do that, even if the ceiling is lower, you're gonna be able to move faster because you have a tighter iteration loop debugging the problem. And in the open loop setting, there's not a feedback cycle debug with the user in the loop.
Paragraph 1314
And so there's some reason to think that that should be an easier debugging problem. The other thing that's worth understanding is that even in a closed loop setting, there's no special software magic of how to infer what the user is truly attempting to do. In the closed loop setting, although they're moving the cursor on the screen, they may be attempting something different than what your model is outputting.
Paragraph 1315
So what the model is outputting is not a signal that you can use to retrain if you want to be able to improve the model further. You still have this very complicated guesstimation or unsupervised problem of figuring out what is the true user intention underlying that signal. And so the open loop problem has the nice property of being easy to debug.
Paragraph 1316
And the second nice property of, it has all the same information and content as the closed loop scenario. Another thing I wanna mention and call out is that this problem doesn't need to be solved in order to give useful control to people. Even today with the solutions we have now and that academia has built up over decades, the level of control that can be given to a user today is quite useful.
Paragraph 1317
It doesn't need to be solved to get to that level of control. But again, I wanna build the world's best mouse. I wanna make it so good that it's not even a question that you want it. And to build the world's best mouse, the superhuman version, you really need to nail that problem.
Paragraph 1318
And a couple maybe details of previous studies that we've done internally that I think are very interesting to understand when thinking about how to solve this problem, the first is that even when you have ground truth data of what the user's trying to do, and you can get this with an able-bodied monkey, a monkey that has a Neuralink device implanted and moving a mouse to control a computer, even with that ground truth dataset, it turns out that the optimal thing to predict to produce high performance BCI is not just the direct control of the mouse.
Paragraph 1319
You can imagine building dataset of what's going on in the brain and what is the mouse exactly doing on the table. And it turns out that if you build the mapping from neuro spikes to predict exactly what the mouse is doing, that model will perform worse than a model that is trained to predict sort of higher level assumptions about what the user might be trying to do.
Paragraph 1320
For example, assuming that the monkey is trying to go in a straight line to the target, it turns out that making those assumptions is actually more effective in producing a model than actually predicting the underlying hand movement.
Paragraph 1321
So the intention, not like the physical movement or whatever.
Paragraph 1322
Yeah.
Paragraph 1323
There's obviously a really strong correlation between the two, but the intention is a more powerful thing to be chasing.
Paragraph 1324
[Bliss] Right.
Paragraph 1325
Well, that's also super interesting. I mean, the intention itself is fascinating, because yes, with the BCI here, in this case, with a digital telepathy, you're acting on the intention, not the action, which is why there's an experience of like feeling like it's happening before you meant for it to happen. That is so cool. And that is why you could achieve like superhuman performance problem in terms of the control of the mouse.
Paragraph 1326
So for open loop, just to clarify, so whenever the person is tasked to like move the mouse to the right, you said there's not feedback so they don't get to get that satisfaction of like actually getting it to move, right?
Paragraph 1327
You could imagine giving the user feedback on a screen, but it's difficult, because at this point, you don't know what they're attempting to do. So what can you show them that would basically give them a signal of I'm doing this correctly or not correctly. So let's take this very specific example. Like maybe your calibration task looks like you're trying to move the cursor a certain position offset.
Paragraph 1328
So your instructions to the user are, "Hey, the cursor's here. Now, when the cursor disappears, imagine moving it 200 pixels from where it was to the right to be over this target." In that kind of scenario, you could imagine coming up with some sort of consistency metric that you could display to the user of, "Okay, I know what the spike train looks like on average when you do this action to the right.
Paragraph 1329
Maybe I can produce some sort of probabilistic estimate of how likely is that to be the action you took given the latest trial or trajectory that you imagined." And that could give the user some sort of feedback of how consistent are they across different trials.
Paragraph 1330
You could also imagine that if the user is prompted with that kind of consistency metric, that maybe they just become more behaviorally engaged to begin with because the task is kind of boring when you don't have any feedback at all. And so there may be benefits to the user experience of showing something on the screen, even if it's not accurate, just because it keeps the user motivated to try to increase that number or push it upwards.
Paragraph 1331
So there's a psychology element here.
Paragraph 1332
Yeah, absolutely.
Paragraph 1333
And again, all of that is UX challenge. How much signal drift is there, hour to hour, day to day, week to week, month to month? How often do you have to recalibrate because of the signal drift?
Paragraph 1334
Yeah, so this is a problem we've worked on, both with NHP, non-human primates, before our clinical trial and then also with Noland during the clinical trial. Maybe the first thing that's worth stating is what the goal is here. So the goal is really to enable the user to have a plug and play experience where I guess they don't have to plug anything in, but a play experience where they can use the device whenever they want to, however they want to.
Paragraph 1335
And that's really what we're aiming for. And so there can be a set of solutions that get to that state without considering this non-stationary problem. So maybe the first solution here that's important is that they can recalibrate whenever they want. This is something that Noland has the ability to do today.
Paragraph 1336
So he can recalibrate the system at 2:00 AM in the middle of the night, without his caretaker or parents or friends around to help push a button for him. The other important part of the solution is that when you have a good model calibrated, that you can continue using that without needing to recalibrate it. So how often he has to do this recalibration today depends really on his appetite for performance.
Paragraph 1337
We observe sort of a degradation through time of how well any individual model works, but this can be mitigated behaviorally by the user adapting their control strategy. It can also be mitigated through a combination of sort of software features that we provide to the user. For example, we let the user adjust exactly how fast the cursor is moving.
Paragraph 1338
We call that the gain, for example, the gain of how fast the cursor reacts to any given input intention. They can also adjust the smoothing, how smooth the output of that cursor intention actually is. They can also adjust the friction, which is how easy is it to stop and hold still. And all these software tools allow the user a great deal of flexibility and troubleshooting mechanisms to be able to solve this problem for themselves.
Paragraph 1339
By the way, all of this is done by looking to the right side of the screen, selecting the mixer, and the mixer you have- - It's like DJ mode. DJ mode for your VCI.
Paragraph 1340
So I mean, it's a really well done interface. It's really, really well done. And so yeah, there's that bias that there's a cursor drift that Noland talked about in a stream, although he said that you guys were just playing around with it with him and then constantly improving. So that could have been just a snapshot of that particular moment, a particular day. But he said that there was this cursor drift and this bias that could be removed by him, I guess looking to the right side of the screen, the left side of the screen to kind of adjust the bias.
Paragraph 1341
Yeah, yeah.
Paragraph 1342
That's one interface action I guess to adjust the bias.
Paragraph 1343
Yeah, so this is actually an idea that comes out of academia. There are some prior work with sort of BrainGate clinical trial participants where they pioneered this idea of bias correction.
Paragraph 1344
The way we've done it I think is, yeah, it's very prototized, very beautiful user experience where the user can essentially flash the cursor over to the side of the screen and it opens up a window where they can actually sort of adjust or tune exactly the bias of the cursor. So bias maybe, for people who aren't familiar, is just sort of what is the default motion of the cursor if you're imagining nothing.
Paragraph 1345
And it turns out that that's one of the first sort of qualia of the cursor control experience that's impacted by neuro non-stationarity.
Paragraph 1346
Quality off the cursor experience.
Paragraph 1347
I don't know how else to describe it. I'm not the guy moving- - It's very poetic, I love it. The quality of the cursor experience. Yeah, I mean, it sounds poetic but it is deeply true. There is an experience, when it works well, it is a joyful, a really pleasant experience. And when it doesn't work well, it's a very frustrating experience. That's actually the art of UX.
Paragraph 1348
It's like you have the possibility to frustrate people or the possibility to give them joy, - And at the end of the day, it really is truly the case that UX is how the thing works. And so it's not just like what's showing on the screen, it's also what control surfaces does a decode provide the user? Like we want them to feel like they're in the F1 card, not like some like minivan, right? And that really truly is how we think about it.
Paragraph 1349
Noland himself is an F1 fan, so we refer to ourself as a pit crew. He really is truly the F1 driver, and there's different control surfaces that different kinds of cars and airplanes provide the user. And we take a lot of inspiration from that when designing how the cursor should behave..
Paragraph 1350
And what maybe one nuance of this is, even details like when you move a mouse on a MacBook track pad, the sort of response curve of how that input that you give the track pad translates to cursor movement is different than how it works with a mouse. When you move on the track pad, there's a different response function, a different curve to how much a movement translates to input to the computer than when you do it physically with a mouse.
Paragraph 1351
And that's because somebody sat down a long time ago when they're designed the initial input systems to any computer and they thought through exactly how it feels to use these different systems. And now we're designing sort of the next generation of this input system to a computer, which is entirely done via the brain, and there's no proprioceptive feedback.
Paragraph 1352
Again, you don't feel the mouse in your hand, you don't feel the keys under your fingertips, and you want a control surface that still makes it easy and intuitive for the user to understand the state of the system and how to achieve what they wanna achieve. And ultimately, the end goal is that that UX is completely, it fades into the background.
Paragraph 1353
It becomes something that's so natural and intuitive that it's subconscious to the user, and they just should feel like they have basically direct control over the cursor. It just does what they want it to do. They're not thinking about the implementation of how to make it do what they want it to do. It's just doing what they want it to do.
Paragraph 1354
Is there some kind of things along the lines of like Fitts' law where you should move the mouse in a certain kind of way that maximizes your chance to hit the target? I don't even know what I'm asking, but I'm hoping the intention of my question will land on a profound answer. No, is there some kind of understanding of the laws of UX when it comes to the context of somebody using their brain to control it? Like that's different than actual with a mouse?
Paragraph 1355
I think we're in the early stages of discovering those laws, so I wouldn't claim to have solved that problem yet. But there's definitely some things we've learned that make it easier for the user to get stuff done. And it's pretty straightforward when you verbalize it, but it takes a while to actually get to that point when you're in the process of debugging the stuff in the trenches.
Paragraph 1356
One of those things is that any machine learning system you build has some number of errors, and it matters how those errors translate to the downstream user experience. For example, if you're developing a search algorithm in your photos, if you search for your friend Joe and it pulls up a photo of your friend Josephine, maybe that's not a big deal because the cost of an error is not that high.
Paragraph 1357
In a different scenario where you're trying to detect insurance fraud or something like this and you're directly sending someone to court because of some machine learning model output, then the errors make a lot more sense to be careful about. You wanna be very thoughtful about how those errors translate to downstream effects. The same is true in BCI.
Paragraph 1358
So for example, if you're building a model that's decoding a velocity output from the brain versus an output where you're trying to modulate the left click, for example. These have sort of different trade-offs of how precise you need to be before it becomes useful to the end user. For velocity, it's okay to be on average correct, because the output of the model is integrated through time.
Paragraph 1359
So if the user's trying to click at position A, and they're currently in position B, they're trying to navigate over time to get between those two points. And as long as the output of the model is on average correct, they can sort of steer it through time with the user control loop in the mix. They can get to the point they wanna get to. The same is not true of a click.
Paragraph 1360
For a click, you're performing it almost instantly at the scale of neurons firing. And so you want to be very sure that that click is correct because a false click can be very destructive to the user. They might accidentally close the tab that they're trying to do something and lose all their progress. They might accidentally like hit some Send button on some text that it's only like half-composed and reads funny after.
Paragraph 1361
So there's different sort of cost functions associated with errors in this space. And part of the UX design is understanding how to build a solution that is when it's wrong, still useful to the end user.
Paragraph 1362
It's so fascinating that assigning cost to every action when an error occurs. So every action, if an error occurs, has a certain cost, and incorporating that into how you interpret the intention, mapping it to the action is really important. I didn't quite until you said it realize there's a cost to like sending the text early. It's like very expensive cost.
Paragraph 1363
Yeah. It's super annoying if you accidentally, like if you're a cursor, imagine if your cursor misclick every once in a while, that's like super obnoxious. And the worst part of it is, usually, when the user's trying to click, they're also holding still because they're over the target they wanna hit and they're getting ready to click, which means that in the data sets that we build, on average, it's the case that sort of low speeds or desire to hold still. It's correlated with when the user's attempting to click.
Paragraph 1364
Wow, that is really fascinating.
Paragraph 1365
It's also not the case. People think that, "Oh, a click is a binary signal. This must be super easy to decode." Well, yes it is, but the bar is so much higher for it to become a useful thing for the user, and there's ways to solve this. I mean, you can sort of take the compound approach of, well, let's just give the like, let's take five seconds to click. Let's take a huge window of time so it can be very confident about the answer.
Paragraph 1366
But again, world's best mouse. The world's best mouse doesn't take a second to click or 500 milliseconds to click. It takes five milliseconds to click or less. And so if you're aiming for that kind of high bar, then you really wanna solve the underlying problem.
Paragraph 1367
So maybe this is a good place to ask about how to measure performance, this whole bits per second. Can you like explain what you mean by that? Maybe a good place to start is to talk about Webgrid as a game, as a good illustration of the measurement of performance.
Paragraph 1368
Yeah, maybe I'll take one zoom out step there, which is just explaining why we care to measure this at all. So again, our goal is to provide the user the ability to control the computer as well as I can and hopefully better. And that means that they can do it at the same speed as what I can do.
Paragraph 1369
It means that they have access to all the same functionality that I have, including all those little details like command tab, command space, all this stuff. They need to be able to do it with their brain and with the same level of reliability as what I can do with my muscles. And that's a high bar. And so we intend to measure and quantify every aspect of that to understand how we're progressing towards that goal.
Paragraph 1370
There's many ways to measure BPS, by the way. This isn't the only way, but we present the user a creative targets, and basically, we compute a score which is dependent on how fast and accurate they can select, and then how small are the targets. And the more targets that are on the screen, the smaller they are, the more information you present per click.
Paragraph 1371
And so if you think about it from information theory point of view, you can communicate across different information theoretic channels. And one such channel is a typing interface you could imagine that's built out of a grid, just like a software keyboard on the screen. And bits per second is a measure that's computed by taking the log of the number of targets on the screen.
Paragraph 1372
You can subtract one if you care to model a keyboard because you have to subtract one for the Delete key on the keyboard, but log of the number of targets on the screen times the number of correct selections minus incorrect, divided by some time window. For example, 60 seconds. And that's sort of the standard way to measure a cursor control task in academia. And all credit in the world goes to this great professor, Dr.
Paragraph 1373
Shenoy of Stanford who came up with that task. And he's also one of my inspirations for being in the field. So all the credit in the world to him for coming up with a standardized metric to facilitate this kind of bragging rights that we have now, to say that Noland is the best in the world at this task with his BCI.
Paragraph 1374
It's very important for progress that you have standardized metrics that people can compare across different techniques and approaches. How well does this do? So yeah, big kudos to him and to all the team at Stanford. Yeah, so for Noland, and for me playing this task, there's also different modes that you can configure this task.
Paragraph 1375
So the Webgrid task can be presented as just sort of a left click on the screen, or you could have targets that you just dwell over, or you could have targets that you left, right click on. You could have targets that are left, right click, middle click, scrolling, clicking, and dragging. You could do all sorts of things within this general framework. But the simplest, purest form is just blue targets show up on the screen. Blue means left click.
Paragraph 1376
That's the simplest form of the game. And the sort of prior records here in academic work and at Neuralink internally with sort of NPS have all been matched or beaten by Noland with his Neuralink device. So sort of prior to Neuralink, the sort of world record for a human using device is somewhere between 4. 2 to 4. 6 BPS, depending on exactly what paper you read and how you interpret it. Noland's current record is 8. 5 BPS.
Paragraph 1377
And again, this sort of median Neuralink performance is 10 BPS. So you can think of it roughly as he's 85% the level of control of a median Neuralinker using their cursor to select blue targets on the screen. And yeah, I think there's a very interesting journey ahead to get us to that same level of 10 BPS performance.
Paragraph 1378
It's not the case that sort of the tricks that got us from four to six BPS, and then six to eight BPS are gonna be the ones that get us from eight to 10. And in my view, the core challenge here is really the labeling problem. It's how do you understand at a very, very fine resolution what the user's attempting to do. And yeah, I highly encourage folks in academia to work on this problem.
Paragraph 1379
What's the journey with Noland on that quest of increasing the BPS on Webgrid? In March, you said that he selected 89,285 targets in Webgrid.
Paragraph 1380
Yep.
Paragraph 1381
So he loves this game. He's really serious about improving his performance in this game. So what is that journey of trying to figure out how to improve that performance? How much can that be done on the decoding side? How much can that be done on the calibration side? How much can that be done on the Noland side of like figuring out how to convey his intention more cleanly?
Paragraph 1382
Yeah, no, this is a great question. So in my view, one of the primary reasons why Noland's performance is so good is because of Noland. Noland is extremely focused and very energetic. He'll play Webgrid sometimes for like four hours in the middle of the night. Like from 2:00 AM to 6:00 AM, he'll be playing Webgrid, just because he wants to push it to the limits of what he can do. And this is not us like asking him to do that. I wanna be clear.
Paragraph 1383
Like we're not saying, "Hey, you should play Webgrid tonight." We just gave him the game as part of our research, and he is able to play it independently and practice whenever he wants, and he really pushes hard to push it, the technology's the absolute limit. And he views that as like his job really to make us be the bottleneck. And boy, has he done that well.
Paragraph 1384
And so the first thing to acknowledge is that he's extremely motivated to make this work. I've also had the privilege to meet other clinical trial participants from BrainGate and other trials, and they very much shared the same attitude of like they view this as their life's work to advance the technology as much as they can. And if that means selecting targets on the screen for four hours from 2:00 AM to 6:00 AM, then so be it.
Paragraph 1385
And there's something extremely admirable about that that's worth calling out. Okay, so then how do you sort of get from where he started, which is no cursor control to a BPS? So I mean, when he started, there's a huge amount of learning to do on his side and our side to figure out what's the most intuitive control for him.
Paragraph 1386
And the most intuitive control for him is sort of, you have to find the set intersection of what do we have this signal to decode. So we don't pick up every single neuron in the motor cortex, which means we don't have representation for every part of the body. So there may be some signals that we have better sort of decode performance on than others.
Paragraph 1387
For example, on his left hand, we have a lot of difficulty distinguishing his left ring finger from his left middle finger. But on his right hand, we have a good control and good modulation detected from the neurons that we're able to record for his pinky, his thumb, and his index finger. So you can imagine how these different sub spaces of modulated activity intersect with what's the most intuitive for him. And this has evolved over time.
Paragraph 1388
So once we gave him the ability to calibrate models on his own, he was able to go and explore various different ways to imagine controlling the cursor. For example, he could imagine controlling the cursor by wiggling his wrist side to side, or by moving his entire arm. I think at one point, he did his feet. He tried like a whole bunch of stuff to explore the space of what is the most natural way for him to control the cursor.
Paragraph 1389
That at the same time, it's easy for us to decode rules.
Paragraph 1390
Just to clarify, it's through the body mapping procedure that you're able to figure out which finger he can move?
Paragraph 1391
Yes, yeah, that's one way to do it. Maybe one nuance of when he's doing it, he can imagine many more things than we represent in that visual on the screen. So we show him sort of abstractly, "Here's a cursor. You figure out what works the best for you." And we obviously have hints about what will work best from that body mapping procedure of we know that this particular action, we can represent well. But it's really up to him to go and explore and figure out what works the best.
Paragraph 1392
But at which point does he no longer visualize the movement of his body and is just visualizing the movement of the cursor?
Paragraph 1393
Yeah.
Paragraph 1394
How quickly does he go from, how quickly does he get there?
Paragraph 1395
So this happened on a Tuesday. I remember this day very clearly, because at some point during the day, it looked like he wasn't doing super well. It looked like the model wasn't performing super well and he was like getting distracted. But he actually, it wasn't the case. Like what actually happened was he was trying something new where he was just controlling the cursor. So he wasn't imagining moving his hand anymore.
Paragraph 1396
He was just imagining, I don't know what it is, some like abstract intention to move the cursor on the screen. And I cannot tell you what the difference between those two things are. I really truly cannot. He's tried to explain it to me before. I cannot give a first person account of what that's like.
Paragraph 1397
But the expletives that he uttered in that moment were enough to suggest that it was a very qualitatively different experience for him to just have direct neural control over a cursor.
Paragraph 1398
I wonder if there's a way through UX to encourage a human being to discover that, because he discovered it, like you said to me, that he's a pioneer. So he discovered that on his own through all of this, the process of trying to move the cursor with different kinds of intentions. But that is clearly a really powerful thing to arrive at, which is to let go of trying to control the fingers and the hand and control the actual digital device with your mind.
Paragraph 1399
That's right, UX is how it works. And the ideal UX is one that the user doesn't have to think about what they need to do in order to get it done. It just does it.
Paragraph 1400
That is so fascinating. But I wonder on the biological side how long it takes for the brain to adapt.
Paragraph 1401
Yeah.
Paragraph 1402
So is it just simply learning like high level software, or is there like a neuroplasticity component where like the brain is adjusting slowly?
Paragraph 1403
Yeah, the truth is, I don't know. I'm very excited to see with sort of the second participant that we implant what the journey is like for them, because we'll have learned a lot more. Potentially, we can help them understand and explore that direction more quickly. This is something I didn't know. This wasn't me prompting Noland to go try this. He was just exploring how to use his device and figure it out himself.
Paragraph 1404
But now that we know that that's a possibility, that maybe there's a way to, for example, hint the user, "Don't try super hard during calibration. Just do something that feels natural, or just directly control the cursor. Don't imagine explicit action." And from there, we should be able to hopefully understand how this is for somebody who has not experienced that before. Maybe that's the default mode of operation for them.
Paragraph 1405
You don't have to go through this intermediate phase of explicit motions.
Paragraph 1406
Or maybe if that naturally happens for people, you can just occasionally encourage them to allow themselves to move the cursor. Actually sometimes, just like with a four minute mile, just the knowledge that that's possible.
Paragraph 1407
Pushes you to do it.
Paragraph 1408
Yeah, enables you to do it, and then it becomes trivial. And then it also makes you wonder, it's the cool thing about humans. Once there's a lot more human participants, they will discover things that are possible.
Paragraph 1409
Yes, and share their experiences.
Paragraph 1410
Yeah, and share.
Paragraph 1411
With each other.
Paragraph 1412
And that because of them sharing it, they'll be able to do it. All of a sudden, that's unlocked for everybody, because just the knowledge sometimes is the thing that enables it to do it.
Paragraph 1413
Yeah, I mean, and just to comment on that too, we've probably tried like a thousand different ways to do various aspects of decoding, and now we know like what the right subspace is to continue exploring further. Again, thanks to Noland and the many hours he's put into this.
Paragraph 1414
And so even just that help, like help constraints sort of the beam search of different approaches that we could explore really helps accelerate for the next person the set of things that we'll get to try on day one, how fast we hope to get them to useful control, how fast we can enable 'em to use it independently, and to get value out of the system.
Paragraph 1415
So yeah, massive hats off to Noland and all the participants that came before him to make this technology a reality.
Paragraph 1416
So how often are the updates to the decoder? 'Cause Noland mentioned like, okay, there's a new update that we're working on, and that in the stream, he said he plays the snake game because it's like super hard. It's a good way for him to test like how good the update is. And he says like sometimes the update is a step backwards. It's a constant like iteration. Like what does the update entail? Is it mostly on the decoder side?
Paragraph 1417
Yeah, a couple comments. So one is it's probably worth drawing distinction between sort of research sessions where we're actively trying different things to understand like what the best approach is versus sort of independent use where we wanted to have an ability to just go use the device, how anybody would wanna use their MacBook.
Paragraph 1418
And so what he's referring to is, I think usually in the context of a research session, where we're trying many, many different approaches to even unsupervised approaches like we talked about earlier to try to come up with better ways to estimate his true intention and more accurately decode it. And in those scenarios, I mean we try, in any given session, he'll sometimes work for like eight hours a day.
Paragraph 1419
And so that can be hundreds of different models that we would try in that day. Like a lot of different things. Now, it's also worth noting that we update the application he uses quite frequently. I think sometimes, up to like four or five times a day. We'll update his application with different features or bug fixes or feedback that he's given us. He's a very articulate person who is part of the solution. He's not a complaining person.
Paragraph 1420
He says, "Hey, here's this thing that I've discovered is not optimal in my flow. Here's some ideas how to fix it. Let me know what your thoughts are. Let's figure out how to solve it." And it often happens that those things are addressed within a couple hours of him giving us his feedback. That's the kind of iteration cycle we'll have.
Paragraph 1421
And so sometimes, at the beginning of the session, he'll give us feedback, and at the end of the session, he's giving us feedback on the next iteration of that process or that set up.
Paragraph 1422
That's fascinating, because one of the things you mentioned, that there was 271 pages of notes taken from the BCI sessions, and this was just in March. So one of the amazing things about human beings that they can provide, especially ones who are smart and excited and all like positive and good vibes like Nolan, that they can provide feedback, continuous feedback.
Paragraph 1423
Yeah, it also requires, just to brag on the team a little bit, I work with a lot of exceptional people, and it requires the team being absolutely laser focused on the user and what will be the best for them. And it requires like a level of commitment of, "Okay, this is what the user feedback was. I have all these meetings. We're gonna skip that today and we're gonna do this." That level of focus commitment is, I would say, underappreciated in the world. And also, you obviously have to have the talent to be able to execute on these things effectively. And yeah, we have that in loads.
Paragraph 1424
Yeah, and this is such a interesting space of UX design because there's so many unknowns here. And I can tell UX is difficult because of how many people do it poorly. It's just not a trivial thing.
Paragraph 1425
Yeah, it's also, you know, UX is not something that you can always solve by just constant iterating on different things. Like sometimes, you really need to step back and think globally, am I even in like the right sort of minima to be chasing down for a solution? Like there's a lot of problems in which sort of fast iteration cycle is the predictor of how successful you will be.
Paragraph 1426
As a good example, like in RL simulation, for example, the more frequently you get reward, the faster you can progress. It's just an easier learning prompt, the more frequently you get feedback. But UX is not that way. I mean, users are actually quite often wrong about what the right solution is, and it requires a deep understanding of the technical system and what's possible, combined with what the problem is you're trying to solve.
Paragraph 1427
Not just how the user expressed it, but what the true underlying problem is to actually get to the right place.
Paragraph 1428
Yeah, that's the old like stories of Steve Jobs like rolling in there, like, yeah, the user is a useful signal, but it's not a perfect signal. And sometimes, you have to remove the floppy disk drive or whatever the, I forgot all the crazy stories of Steve Jobs like making wild design decisions. But there, some of it is aesthetic, that some of it is about the love you put into the design, which is very much a Steve Jobs-Jony Ive type of thing.
Paragraph 1429
But when you have a human being using their brain to interact with it, it also is deeply about function. It's not just aesthetic. And that you have to empathize with a human being before you, while not always listening to them directly. Like you have to deeply empathize. It's fascinating. It's really, really fascinating. And at the same time, iterate. But not iterate in small ways. Sometimes, a complete like rebuilding the design.
Paragraph 1430
He said that, Noland said in the early days, the UX sucked. But you improved quickly. What was that journey like?
Paragraph 1431
Yeah, I mean, I'll give one concrete example. So he really wanted to be able to read manga. This is something that he, I mean, it sounds like a simple thing, but it's actually a really big deal for him. And he couldn't do it with his mouth stick. It wasn't accessible. You can't scroll with a mouse stick on his iPad on the website that he wanted to be able to use to read the newest manga.
Paragraph 1432
Might be a good quick pause to say the mouth stick is the thing he's using, holding a stick in his mouth to scroll on a tablet.
Paragraph 1433
Right, yeah, it's basically, you can imagine it's a stylus that you hold between your teeth. Yeah, it's basically a very long stylus.
Paragraph 1434
And it's exhausting, it hurts, and it's inefficient.
Paragraph 1435
Yeah, and maybe it's also worth calling out, there are other alternative assisted technologies, but the particular situation Noland's in, and this is not uncommon, and I think it's also not well understood by folks, is that he's relatively spastic so he'll have muscle spasms from time to time.
Paragraph 1436
And so any assistive technology that requires him to be positioned directly in front of a camera, for example, an eye tracker, or anything that requires him to put something in his mouth just is a no go, 'cause he'll either be shifted out of frame when he has a spasm, or if he has something in his mouth, it'll stab him in the face if he spasms too hard.
Paragraph 1437
So these kind of considerations are important when thinking about what advantages a BCI has in someone's life. If it fits ergonomically into your life in a way that you can use it independently when your caretaker's not there, wherever you want to, either in the bed or in the chair, depending on your comfort level and your desire to have pressure sores, all these factors matter a lot in how good the solution is in that user's life.
Paragraph 1438
So one of these very fun examples is scroll. So again, manga is something he wanted to be able to read, and there's many ways to do scroll with a BCI. You can imagine like different gestures, for example. The user could do that, would move the page. But scroll is a very fascinating control surface, because it's a huge thing on the screen in front of you.
Paragraph 1439
So any sort of jitter in the model output, any sort of error in the model output causes like an earthquake on the screen. Like you really don't wanna have your manga page you're trying to read be shifted up and down a few pixels just because your scroll decoder is not completely accurate.
Paragraph 1440
And so this was an example where we had to figure out how to formulate the problem in a way that the errors of the system, whenever they do occur, and we'll do our best to minimize them, whenever those errors do occur, that it doesn't interrupt the qualia again of the experience that the user is having. It doesn't interrupt their flow of reading their book. And so what we ended up building is this really brilliant feature.
Paragraph 1441
This is teammate named Ruse, who worked on this really brilliant work called quick scroll. And quick scroll basically looks at the screen, and it identifies where on the screen are scroll bars. And it does this by deeply integrated with MacOs to understand where are the scroll bars actively present on the screen using the sort of accessibility tool that's available to MacOs apps.
Paragraph 1442
And we identified where those scroll bars are and we provided a BCI scroll bar. And the BCI scroll bar looks similar to a normal scroll bar but it behaves very differently in that once you sort of move over to it, your cursor sort of morphs onto it. It sort of attaches or latches onto it. And then once you push up or down in the same way that you'd use a push to control the normal cursor, it actually moves the screen for you.
Paragraph 1443
So it's basically like remapping the velocity to a scroll action. And the reason that feels so natural and intuitive is that, when you move over to attach to it, it feels like magnetic. So you're like sort of stuck onto it. And then it's one continuous action. You don't have to like switch your imagined movement. You sort of snap onto it and then you're good to go. You just immediately can start pulling the page down or pushing it up.
Paragraph 1444
And even once you get that right, there's so many little nuances of how the scroll behavior works to make it natural and intuitive. So one example is momentum. Like when you scroll a page with your fingers on the screen, you actually have some like flow. Like it doesn't just stop right when you lift your finger up. The same is true with BCI scroll. So we had to spend some time to figure out what are the right nuances.
Paragraph 1445
When you don't feel the screen under your fingertip anymore, what is the right sort of dynamic or what's the right amount of page give, if you will, when you push it to make it flow the right amount for the user to have a natural experience reading their book.
Paragraph 1446
And there's a million, I mean, I could tell you like there's so many little minutia of how exactly that scroll works that we spent probably like a month getting right to make that feel extremely natural and easy for the user to navigate.
Paragraph 1447
I mean, even to scroll on a smartphone with your finger feels extremely natural and pleasant. And it probably takes a extremely long time to get that right. And actually, the same kind of visionary UX design that we're talking about, don't always listen to the users but also listen to them, and also have like visionary big, like throw everything out, think from first principles but also not.
Paragraph 1448
Yeah, yeah, by the way, it just makes me think that scroll bars on the desktop probably have stagnated and never taken that like, 'cause the snap, same as like snap the grid, snap the scroll bar action you're talking about is something that could potentially be extremely useful in the desktop setting, even just for users to just improve the experience, 'cause the current scroll bar experience in the desktop is horrible.
Paragraph 1449
Yeah, agreed.
Paragraph 1450
It's hard to find, hard to control. There's not a momentum. And the intention should be clear. When I start moving towards a scroll bar, there should be a snapping to the scroll bar action. But of course, maybe I'm okay paying that cost, but there's hundreds of millions of people paying that cost nonstop. But anyway, but in this case, this is necessary because there's an extra cost paid by Noland for the jitteriness. So you have to switch between the scrolling and the reading. There has to be a phase shift between the two. Like when you're scrolling, you're scrolling.
Paragraph 1451
Right, right, so that is one drawback of the current approach. Maybe one other just sort of case study here, so again, UX is how it works, and we think about that holistically from like the, even the feature detection level of what we detect in the brain to how we design the decoder, what we choose to decode, to then how it works once it's being used by the user.
Paragraph 1452
So another good example in the sort of how it works once they're actually using the decoder, the output that's displayed on the screen is not just what the decoder says. It's also a function of what's going on on the screen.
Paragraph 1453
So we can understand, for example, that when you're trying to close a tab, that very small, stupid little X that's extremely tiny, which is hard to get precisely hit if you're dealing with sort of a noisy output of the decoder, we can understand that that is a small little X you might be trying to hit and actually make it a bigger target for you.
Paragraph 1454
Similar to how when you're typing on your phone, if you are used to like the iOS keyboard, for example, it actually adapts the target size of individual keys based on an underlying language model. So it'll actually understand if I'm typing, "Hey, I'm going to see L." It'll make the E key bigger, because in those Lex is the person I'm gonna go see.
Paragraph 1455
And so that kind of predictiveness can make the experience much more smooth even without improvements to the underlying decoder or feature detection part of the stack. So we do that with a feature called magnetic targets. We actually indexed the screen and we understand, okay, these are the places that are very small targets that might be difficult to hit.
Paragraph 1456
Here's the kind of cursor dynamics around that location that might be indicative of the user trying to select it. Let's make it easier. Let's blow up the size of it in a way that makes it easier for the user to sort of snap onto that target. So all these little details, they matter a lot in helping the user be independent in their day-to-day living.
Paragraph 1457
So how much of the work on the decoder is generalizable to P2, P3, P4, P5, PM? How do you improve the decoder in a way that's generalizable?
Paragraph 1458
Yeah, great question. So the underlying signal we're trying to decode is gonna look very different in P2 than in P1. For example, channel number 345 is gonna mean something different in user one than it will in user two, just because that electrode that corresponds with channel 345 is gonna be next to a different neuron in user one versus user two.
Paragraph 1459
But the approach is the methods, the user experience of how do you get the right sort of behavioral pattern from the user to associate with that neural signal, we hope that will translate over multiple generations of users. And beyond that, it's very, very possible. In fact, quite likely that we've overfit to sort of Noland's user experience desires and preferences.
Paragraph 1460
And so what I hope to see is that when we get second, third, fourth participant, that we find sort of what the right wide minimums are that cover all the cases, that make it more intuitive for everyone. And hopefully, there's a crosspollination of things where, "Oh, we didn't think about that with this user, because they can speak. But with this user who just can fundamentally not speak at all, this user experience is not optimal."
Paragraph 1461
And that will actually, those improvements that we make there should hopefully translate then to even people who can't speak but don't feel comfortable doing so because we're in a public setting, like their doctor's office.
Paragraph 1462
So the actual mechanism of open loop labeling and then closed loop labeling will be the same, and hopefully, can generalize across the different users as they're doing the calibration step. And the calibration step is pretty cool. I mean, that in itself, the interesting thing about Webgrid, which is like closed loop, it's like fun. I love it when there's like, they used to be kind of idea of human computation, which is using actions a human would want to do anyway to get a lot of signal from.
Paragraph 1463
Yeah.
Paragraph 1464
And like Webgrid is that, like a nice video game that also serves as great calibration.
Paragraph 1465
It's so funny. I've heard this reaction so many times. Before sort of the first user was implanted, we had an internal perception that the first user would not find this fun. And so we thought really quite a bit actually about like, should we build other games that like are more interesting for the user so we can get this kind of data and help facilitate research for long duration and stuff like this? Turns out that like people love this game.
Paragraph 1466
Yeah.
Paragraph 1467
I always loved it, but I didn't know that that was a shared perception.
Paragraph 1468
Yeah, and just in case it's not clear, Webgrid is, there's a grid of, let's say 35 by 35 cells, and one of them lights up blue and you have to move your mouse over that and click on it. And if you miss it and it's red- - [Bliss] I played this game for so many hours. So many hours.
Paragraph 1469
And what's your record, you said?
Paragraph 1470
I think I have the highest at Neuralink. Right now, my record's 17 BPS.
Paragraph 1471
17 BPS.
Paragraph 1472
Which is about, if you imagine that 35 by 35 grid, you're hitting about 100 trials per minute in. So 100 correct selections in that one minute window. So you're averaging about, between 500, 600 milliseconds per selection.
Paragraph 1473
So one of the reasons that I think I struggle with that game is I'm such a keyboard person, so everything is done with via keyboard. If I can avoid touching the mouse, it's great. So how can you explain your high performance?
Paragraph 1474
I have like a whole ritual I go through when I play Webgrid. So it's actually like a diet plan associated with this. Like it's a whole thing, so great.
Paragraph 1475
The first thing- - You have to fast for five days, have to go up to the mountains.
Paragraph 1476
Actually, it kind of, I mean, the fasting thing is important. So this is like, you know- - Focuses the mind, yeah?
Paragraph 1477
Yeah, it's true. So what I do is actually, I don't eat for a little bit beforehand. And then I'll actually eat like a ton of peanut butter right before, and I get like- - This is a real thing?
Paragraph 1478
This is a real thing, yeah. And then it has to be really late at night. This is again a night owl thing I think we share, but it has to be like midnight, 2:00 AM kind of time window. And I have a very specific like physical position I'll sit in, which is, I used to be, I was homeschooled growing up and so I did most of my work like on the floor, just like in my bedroom or whatever. And so I have a very specific situation on the floor.
Paragraph 1479
On the floor, I sit and play, and then you have to make sure like there's not a lot of weight on your elbow when you're playing so that you can move quickly. And then I turn the gain of the cursor, so the speed of the cursor way, way up. So it's like small motions that actually move the cursor.
Paragraph 1480
Are you moving with your wrist or you're never moving on- - I move my fingers. So my wrist is almost completely still. I'm just moving my fingers.
Paragraph 1481
Yeah. You know those, just on a small tangent, which I've been meaning to go down this rabbit hole of people that set the world record in Tetris. Those folks, they're playing, there's a way to, did you see this?
Paragraph 1482
I see like the three, like all the fingers are moving.
Paragraph 1483
Yeah, you could find a way to do it where like it's using a loophole, like a bug that you can do some incredibly fast stuff. So it's along that line but not quite. But you do realize there'll be like a few programmers right now listening to this cool fast and eat peanut butter and be like- - Yeah, please, please trade my record. I mean, the reason I did this, literally, was just because I wanted the bar to be high. The team, like I wanted the number that we aim for should not be like the median performance. It should be able to beat all of us at least. Like that should be the minimum bar.
Paragraph 1484
What do you think is possible, like 20 scrapes?
Paragraph 1485
Yeah, I don't know what the limits. I mean, the limits you can calculate just in terms of like screen refresh rate and like cursor immediately jumping to the next target. But I mean, I'm sure there's limits before that with just sort of reaction time and visual perception and things like this. I would guess it's in the below 40 but above 20, somewhere in there is probably the right that I'd never be thinking about.
Paragraph 1486
It also matters like how difficult the task is. You could imagine like some people might be able to do like 10,000 targets on the screen and maybe they can do better that way. There's some like task optimizations you could do to try to boost your performance as well.
Paragraph 1487
What do you think it takes for Noland to be able to do above 8.5? To keep increasing that number? You said like every increase in the number might require different- - [Bliss] Yeah.
Paragraph 1488
Different improvements in the system.
Paragraph 1489
Yeah, I think the nature of this work is, I mean, the first answer that's important to say is, I don't know. This is edge of the research. So again, nobody's gotten to that number before. So what's next is gonna be a heuristic guess from my part. What we've seen historically is that different parts of the stack follow next to different time points.
Paragraph 1490
So when I first joined Neuralink, like three years ago or so, one of the major problems was just the latency of the Bluetooth connection. It was just like the radial device wasn't super good. It was an early revision of the implant, and it just like, no matter how good your decoder was, if your thing is updating every 30 milliseconds or 50 milliseconds, it's just gonna be choppy.
Paragraph 1491
And no matter how good you are, that's gonna be frustrating and lead to challenges. So at that point, it was very clear that the main challenge is just get the data off the device in a very reliable way, such that you can enable the next challenge to be tackled.
Paragraph 1492
And then at some point, it was, actually, the modeling challenge of how do you just build a good mapping, like the supervised learning problem of you have a bunch of data and you have a label you're trying to predict, just what is the right like neuro decoder architecture and hyper parameters to optimize that? And that was a problem for a bit. And once you solve that, it became a different bottleneck.
Paragraph 1493
I think the next bottleneck after that was actually just sort of software stability and reliability. If you have widely varying sort of inference latency in your system, or your app just lags out every once in a while, it decreases your ability to maintain and get in a state of flow, and it basically just disrupts your control experience.
Paragraph 1494
And so there's a variety of different software bugs and improvements we made that basically increased the performance of the system, made it much more reliable, much more stable, and led to a state where we could reliably collect data to build better models with. So that was a bottleneck for a while. It's just sort of like the software stack itself.
Paragraph 1495
If I were to guess right now, there's sort of two major directions you could think about for improving BPS further. The first major direction is labeling. So labeling is again this fundamental challenge of given a window of time where the user is expressing some behavioral intent. What are they really trying to do at the granularity of every millisecond? And that, again, is a task design problem. It's a UX problem.
Paragraph 1496
It's a machine learning problem. It's a software problem. Sort of touches all those different domains. The second thing you can think about to improve BPS further is either completely changing the thing you're decoding or just extending the number of things that you're decoding. So this is serving the direction of functionality, okay? So you can imagine giving more clicks. For example, a left click, a right click, a middle click.
Paragraph 1497
Different actions like click and drag, for example. And that can improve the effective bit rate of your communication prosthesis. If you're trying to allow the user to express themselves through any given communication channel, you can measure that with bits per second. But what action measures at the end of the day is how effective are they at navigating their computer.
Paragraph 1498
And so from the perspective of the downstream tasks that you care about, functionality and extending functionality is something we're very interested in, because not only can it improve the sort of number of BPS, but it can also improve the downstream sort of independence that the user has and the skill and efficiency with which they can operate their computer.
Paragraph 1499
Would the number of threads increasing also potentially help?
Paragraph 1500
Yes, short answer is yes. It's a bit nuanced how that curve or how that manifests in the numbers. So what you'll see is that if you sort of plot a curve of number of channels that you're using for decode versus either the offline metric of how good you are at decoding, or the online metric of sort of, in practice, how good is the user using this device, you see roughly a log curve.
Paragraph 1501
So as you move further out in number of channels, you get a corresponding sort of logarithmic improvement in control quality and offline validation metrics. The important nuance here is that each channel corresponds with a specific represented intention in the brain. So for example, if you have a channel 254, it might correspond with moving to the right. Channel 256 might mean move to the left.
Paragraph 1502
If you want to expand the number of functions you want to control, you really want to have a broader set of channels that covers a broader set of imagined movements. You can think of it kinda like Mr. Potato man actually. Like if you had a bunch of different imagined movements you could do, how would you map those imagined movements to input to a computer? You can imagine handwriting to output characters on the screen.
Paragraph 1503
You can imagine just typing with your fingers and have that output text on the screen. You can imagine different finger modulations for different clicks. You can imagine wiggling your big nose or opening some menu or wiggling your big toe to have like command tab occur or something like this. So it's really, the amount of different actions you can take in the world depends on how many channels you have on the information content that they carry.
Paragraph 1504
Right, so that's more about the number of actions. So actually, as you increase the number of threads, that's more about increasing the number of actions you're able to perform.
Paragraph 1505
One other nuance there that is worth mentioning, so again, our goal is really to enable a user with process to control their computer as fast as I can. So that's BPS. With all the same functionality I have, which is what we just talked about, but then also as reliably as I can. And that last point is very related to channel account discussion.
Paragraph 1506
So as you scale out number of channels, the relative importance of any particular feature of your model input to the output control of the user diminishes, which means that if the sort of neural non-stationary effect is per channel, or if the noise is independent, such that more channels means, on average, less output effect, then your reliability of your system will improve.
Paragraph 1507
So one sort of core thesis that at least I have is that scaling channel account should improve the reliability system without any work on the decoder itself.
Paragraph 1508
Can you linger on the reliability here? So first of all, when you see non-stationarity of the signal, which aspect are you referring to?
Paragraph 1509
Yeah, so maybe let's talk briefly what the actual underlying signal looks like. So again, I spoke very briefly at the beginning about how when you imagine moving to the right or imagine moving to the left, neurons might fire more or less. And their frequency content of that signal, at least in the motor cortex, it's very correlated with the output intention, the behavioral task that the user is doing.
Paragraph 1510
You can imagine actually, this is not obvious, that rate coding, which is the name of that phenomenon is like the only way the brain could represent information. You can imagine many different ways in which the brain could encode intention. And there's actually evidence like in bats, for example, that there's temporal codes. So timing codes of like exactly when particular neurons fire is the mechanism of information representation.
Paragraph 1511
But at least in the motor cortex, there's a substantial evidence that it's rate coding, or at least one, like first order of fact is that it's rate coding. So then if the brain is representing information by changing the sort of frequency of a neuron firing, what really matters is sort of the delta between sort of the baseline state of the neuron and what it looks like when it's modulated.
Paragraph 1512
And what we've observed and what has also been observed in academic work is that that baseline rate, sort of the, if you're to target the scale, if you imagine that analogy for like measuring flour or something when you're baking, that baseline state of how much the pot weighs is actually different day to day.
Paragraph 1513
And so if what you're trying to measure is how much rice is in the pot, you're gonna get a different measurement different days, because you're measuring with different pots. So that baseline rate shifting is really the thing that, at least from a first order description of the problem, is what's causing this downstream bias.
Paragraph 1514
There can be other effects, non-linear effects on top of that, but at least, at a very first order description of the problem, that's what we observe day to day is that the baseline firing rate of any particular neuron are observed on a particular channel is changing.
Paragraph 1515
So can you just adjust to the baseline to make it relative to the baseline nonstop?
Paragraph 1516
Yeah, this is a great question. So with monkeys, we have found various ways to do this. One example way to do this is you ask them to do some behavioral task, like play the game with a joystick, you measure what's going on in the brain, you compute some mean of what's going on across all the input features, and you subtract it in the input when you're doing your BCI session. Works super well.
Paragraph 1517
For whatever reason, that doesn't work super well with Noland. I actually don't know the full reason why, but I can imagine several explanations. One such explanation could be that the context effect difference between some open loop task and some closed loop task is much more significant with Noland than it is with a monkey.
Paragraph 1518
Maybe in this open loop task, he's watching the Lex Fridman podcast while he's doing the task, or he's whistling and listening to music and talking with his friend and ask his mom what's for dinner while he's doing this task.
Paragraph 1519
And so the exact sort of difference in context between those two states may be much larger, and thus lead to a bigger sort of generalization gap between the features that you're normalizing at sort of open loop time and what you're trying to use at close loop time.
Paragraph 1520
That's interesting. Just on that point, it's kind of incredible to watch Noland be able to do, to multitask, to do multiple tasks at the same time, to be able to move the mouse courser effectively while talking and while being nervous, because he's talking in front of- - Kicking my ass and chest too, yeah.
Paragraph 1521
Kicking your ass. And talk trash while doing it. So all at the same time. And yes, if you are trying to normalize to the baseline, that might throw everything off. Boy is that interesting.
Paragraph 1522
Maybe one comment on that too. For folks that aren't familiar with assisted technology, I think there's a common belief that, well, why can't you just use an eye tracker or something like this for helping somebody move a mouse on the screen? And it's a really a fair question, and one that I actually did was not confident before Noland, that this was gonna be a profoundly transformative technology for people like him.
Paragraph 1523
And I'm very confident now that it will be, but the reasons are subtle. It really has to do with ergonomically how it fits into their life. Even if you can just offer the same level of control as what they would have with an eye tracker or with a mouse stick, but you don't need to have that thing in your face. You don't need to be positioned a certain way. You don't need your caretaker to be around to set it up for you.
Paragraph 1524
You can activate it when you want, how you want, wherever you want. That level of independence is so game changing for people. It means that they can text a friend at night privately without their mom needing to be in the loop. It means that they can like open up and browse the internet at 2:00 AM when nobody's around to set their iPad up for them. This is like a profoundly game changing thing for folks in that situation.
Paragraph 1525
And this is even before we start talking about folks that may not be able to communicate at all or ask for help when they want to. This can be potentially the only link that they have to the outside world. And yeah, that one doesn't I think need explanation of why that's so impactful.
Paragraph 1526
You mentioned neural decoder. How much machine learning is in the decoder? How much magic, how much science, how much art, how difficult is it to come up with a decoder that figures out what these sequence of spikes mean?
Paragraph 1527
Yeah, good question. There's a couple different ways to answer this. So maybe I'll zoom out briefly first, and then I'll go down one of the rabbit holes. So the zoomed out view is that building the decoder is really the process of building the dataset, plus compiling it into the weights. And each of those steps is important.
Paragraph 1528
The direction I think of further improvement is primarily going to be in the dataset side of how do you construct the optimal labels for the model. But there's an entirely separate challenge of then how do you compile the best model. And so I'll go briefly down the second one, down the second rabbit hole. One of the main challenges with designing the optimal model for BCI is that offline metrics don't necessarily correspond to online metrics.
Paragraph 1529
It's fundamentally a control problem. The user is trying to control something on the screen, and the exact sort of user experience of how you output the intention impacts their ability to control. So for example, if you just look at validation loss, as predicted by your model, there can be multiple ways to achieve the same validation loss. Not all of them are equally controllable by the end user.
Paragraph 1530
And it might be as simple as saying, "Oh, you could just add auxiliary loss terms that like help you capture the thing that actually matters." But this is a very complex nuanced process. So how you turn the labels into the model is more of a nuanced process than just like a standard supervised learning problem.
Paragraph 1531
One very fascinating anecdote here, we've tried many different sort of neural network architectures that translate brain data to velocity outputs, for example. And one example that's stuck in my brain from a couple years ago now is, at one point, we were using just fully connected networks to decode the brain activity.
Paragraph 1532
We tried A/B test where we were measuring the relative performance in online control sessions of sort of 1D convolution over the input signal. So if you imagine per channel, you have a sliding window that's producing some Commvault feature for each of those input sequences for every single channel simultaneously.
Paragraph 1533
You can actually get better validation metrics, meaning you're fitting the data better, and it's generalizing better on offline data if you use this convolutional architecture. You're reducing parameters. It's sort of a standard procedure when you're dealing with time series data. Now it turns out that when using that model online, the controllability was worse, was far worse, even though the offline metrics were better.
Paragraph 1534
And there can be many ways to interpret that, but what that taught me at least was that, hey, it's at least the case right now that if you were to just throw a bunch of computer at this problem, and you were trying to sort of hyper parameter optimize, or let some GPT model hard code or come up with or invent many different solutions, if you were just optimizing for loss, it would not be sufficient, which means that there's still some inherent modeling gap here.
Paragraph 1535
There's still some artistry left to be uncovered here of how to get your model to scale with more compute. And that may be fundamentally labeling problem, but there may be other components to this as well.
Paragraph 1536
Is it data constrained at this time? Which is what it sounds like. How do you get a lot of good labels?
Paragraph 1537
Yeah, I think it's data quality constrained, not necessarily data quantity constrained.
Paragraph 1538
But even like even just a quantity. I mean, 'cause it has to be trained on the interactions. I guess there's not that many interactions.
Paragraph 1539
Yeah, so it depends what version of this you're talking about. So if you're talking about like, let's say the simplest example of just 2D velocity, then I think, yeah, data quality is the main thing.
Paragraph 1540
If you're talking about how to build a sort of multifunction output that lets you do all the inputs, the computer that you and I can do, then it's actually a much more sophisticated nuanced modeling challenge, because now you need to think about not just when the user's left clicking, but when you're building the left click model, you also need to be thinking about how to make sure it doesn't fire when they're trying to right click or when they're trying to move the mouse.
Paragraph 1541
So one example of an interesting bug from like sort of week one of a BCI with Nolan was, when he moved the mouse, the click signal sort of dropped off a cliff, and when he stopped, the click signal went up. So again, there's a contamination between the two inputs. Another good example was, at one point, he was trying to do sort of a left click and drag. And the minute he started moving, the left click signal dropped off a cliff.
Paragraph 1542
So again, 'cause there's some contamination between the two signals, you need to come up with some way to either in the dataset or in the model, build robustness against this kind of, you think of it like overfitting, but really, it's just that the model has not seen this kind of variability before. So you need to find some way to help the model with that.
Paragraph 1543
This is super cool because it feels like all of this is very solvable, but it's hard.
Paragraph 1544
Yes, it is fundamentally an engineering challenge. This is important to emphasize, and it's also important to emphasize that it may not need fundamentally new techniques, which means that people who work on, let's say, unsupervised speech classification using CTC loss, for example, with internal to Siri, they could potentially have very applicable skills to this.
Paragraph 1545
So what things are you excited about in the future development of the software stack on Neuralink? So everything we've been talking about, the decoding, the UX- - I think there's some I'm excited about, like something I'm excited about from the technology side, and some I'm excited about for understanding how this technology is going to be best situated for entering the world. So I'll work backwards.
Paragraph 1546
On the technology entering the world side of things, I'm really excited to understand how this device works for folks that cannot speak at all. They have no ability to sort of bootstrap themselves into useful control by voice command, for example, and are extremely limited in their current capabilities.
Paragraph 1547
I think that will be an incredibly useful signal for us to understand, I mean, really what is an existential threat for all startups, which is product market fit. Does this device have the capacity and potential to transform people's lives in the current state? And if not, what are the gaps? And if there are gaps, how do we solve them most efficiently?
Paragraph 1548
So that's what I'm very excited about for the next sort of year or so of clinical trial operations. The technology side, I'm quite excited about basically everything we're doing. I think it's gonna be awesome. The most prominent one, I would say, is scaling channel count. So right now, we have a thousand channel device. The next version, we'll have between 3 and 6,000 channels. And I would expect that curve to continue in the future.
Paragraph 1549
And it's unclear what set of problems will just disappear completely at that scale, and what set of problems will remain and require for their focus. And so I'm excited about the clarity of gradient that that gives us in terms of the user experiences we choose to focus our time and resources on. And also in terms of the, yeah, even things as simple as not stationary.
Paragraph 1550
Like does that problem just completely go away at that scale, or do we need to come up with new creative UXs still even at that point?
Paragraph 1551
And also, when we get to that time point, when we start expanding out dramatically the set of functions that you can output from one brain, how to deal with all the nuances of both the user experience of not being able to feel the different keys under your fingertips, but still needing to be able to modulate all of them in synchrony to achieve the thing you want.
Paragraph 1552
And again, you don't have that appropriate set to feedback loop, so how can you make that intuitive for a user to control a high dimensional control surface without feeling the thing physically? I think that's gonna be a super interesting problem. I'm also quite excited to understand, do these scaling laws continue? Like as you scale channel count, how much further out do you go before that saturation point is truly hit?
Paragraph 1553
And it's not obvious today. I think we only know what's in the sort of interpolation space. We only know what's between zero and 1,024. We don't know what's beyond that. And then there's a whole sort of like range of interesting sort of neuroscience and brain questions, which is when you stick more stuff in the brain, in more places, you get to learn much more quickly about what those brain regions represent.
Paragraph 1554
And so I'm excited about that fundamental neuroscience learning, which is also important for figuring out how, and to most efficiently, insert electrodes in the future. So yeah, I think all those dimensions, I'm really, really excited about. And that doesn't even get close to touching the sort of software stack that we work on every single day and what we're working on right now.
Paragraph 1555
Yeah, it seems virtually impossible to me that a thousand electrodes is where it saturates. It feels like this would be one of those silly notions in the future, where obviously, you should have millions of electrodes, and this is where like the true breakthroughs happen.
Paragraph 1556
Yeah.
Paragraph 1557
You tweeted- - Oh.
Paragraph 1558
"Some thoughts are most precisely described in poetry." Why do you think that is?
Paragraph 1559
I think it's because the information bottleneck of language is pretty steep. And yet you're able to reconstruct in the other person's brain more effectively without being literal. If you can express a sentiment such that in their brain, they can reconstruct the actual true underlying meaning and beauty of the thing that you're trying to get across. The sort of the generator function in their brain's more powerful than what language can express. And so the mechanism of poetry is really just to feed or seed that generator function.
Paragraph 1560
So being literal sometimes is a suboptimal compression for the thing you're trying to convey.
Paragraph 1561
And it's actually in the process of the user going through that generation that they understand what you mean. That's the beautiful part. It's also like when you look at a beautiful painting, like it's not the pixels of the painting that are beautiful, it's the thought process that occurs when you see that, the experience of that. That actually is a thing that matters.
Paragraph 1562
Yeah, it's resonating with some deep- - [Bliss] Yeah.
Paragraph 1563
Thing within you that the artist also experienced and was able to convey that through the pixels. And that's actually gonna be relevant for full on telepathy. It's like if you just read the poetry, literally, that doesn't say much of anything interesting. It requires a human to interpret it.
Paragraph 1564
So it's the combination of the human mind and all the experiences that human being has within the context of the collective intelligence of the human species that makes that poem make sense. And they load that in. And so in that same way, the signal that carries from human to human meaning may seem trivial, but may actually carry a lot of power because of the complexity of the human mind on the receiving end. Yeah, that's interesting.
Paragraph 1565
Poetry still doesn't, who is it? I think Yoshi Bako first said something about all the people that think we've achieved AGI explain why humans like music.
Paragraph 1566
[Bliss] Oh yeah.
Paragraph 1567
And until the AGI likes music, you haven't achieved AGI or something like this.
Paragraph 1568
Do you not think that's like some next token entropy surprise kind of thing going on there?
Paragraph 1569
I don't know.
Paragraph 1570
I don't know either. I listen to a lot of classical music and also read a lot of poetry. And yeah, I do wonder if like there is some element of the next token surprise factor going on there.
Paragraph 1571
Yeah, maybe.
Paragraph 1572
Because I mean, like a lot of the tricks in both poetry and music are like basically, you have some repeated structure. And then you do like a twist. It's like, okay, verse or like clause one, two, three is one thing, and then clause four is like, okay, now we're onto the next theme. And they kind of play with exactly when the surprise happens and the expectation of the user.
Paragraph 1573
And that's even true like, through history as musicians evolve music, they take like some nuanced structure that people are familiar with and they just tweak it a little bit. Like they tweak it and add a surprising element. This is especially true in classical music heritage. But that's what I'm wondering, like is it all just entropy- - So breaking structure or breaking symmetry is something that humans seem to like. Maybe as simple as that.
Paragraph 1574
Yeah, and I mean, great artists copy, and they also, knowing which rules to break is the important part. And that fundamentally, it must be about the listener of the piece. Like which rule is the right one to break, it's about the user or the audience member perceiving that as interesting.
Paragraph 1575
What do you think is the meaning of human existence?
Paragraph 1576
There's a TV show I really like called "The West Wing." And in "The West Wing," there's a character. He's the president of the United States who's having a discussion about the Bible with one of their colleagues. And the colleague says something about, "The Bible says X, Y, and Z." And the President says, "Yeah, but it also says A, B, C." And the person says, "Do you believe the Bible to be literally true?"
Paragraph 1577
And the President says, "Yes, but I also think that neither of us are smart enough to understand it." I think the analogy here for the meaning of life is that largely, we don't know the right question to ask. And so I think I'm very aligned with sort of "The Hitchhiker's Guide to the Galaxy" version of this question, which is basically, if we can ask the right questions, it's much more likely we find the meaning of human existence.
Paragraph 1578
And so in the short term, as a heuristic in the sort of search policy space, we should try to increase the diversity of people asking such questions, or generally of consciousness and conscious beings asking such questions. So again, I think I'll take the 'I don't know card' here, but say I do think there are meaningful things we can do that improve the likelihood of answering that question.
Paragraph 1579
It's interesting how much value you assign to the task of asking the right questions. That's the main thing is not the answers, it's the questions.
Paragraph 1580
This point by the way is driven home in a very painful way when you try to communicate with someone who cannot speak, because a lot of the time, the last thing to go is they have the ability to somehow wiggle a lip or move something that allows them to say yes or no. And in that situation, it's very obvious that what matters is, are you asking them the right question to be able to say yes or no to?
Paragraph 1581
Wow, that's powerful. Well, Bliss, thank you for everything you do, and thank you for being you, and thank you for talking today.
Paragraph 1582
Thank you.
Paragraph 1583
Thanks for listening to this conversation with Bliss Chapman. And now, dear friends, here's Noland Arbaugh, the first human being to have a Neuralink device implanted in his brain. You had a diving accident in 2016 that left you paralyzed with no feeling from the shoulders down. How did that accident change your life?
Paragraph 1584
That's sort of a freak thing that happened. Imagine you're running into the ocean, although this is a lake, but you're running into the ocean and you get to about waist high, and then you kind of like dive in, take the rest of the plunge under the wave or something. That's what I did. And then I just never came back up. Not sure what happened. I did it running into the water with a couple of guys.
Paragraph 1585
And so my idea of what happened is really just that I took like a stray fist, elbow, knee, foot, something to the side of my head. The left side of my head was sore for about a month afterwards. So must must've taken a pretty big knock. And then they both came up, and I didn't. And so I was face down in the water for a while. I was conscious. And then eventually just realized I couldn't hold my breath any longer.
Paragraph 1586
And I keep saying, "Took a big drink." People, I don't know if they like that I say that. It seems like I'm making light of it all, but it's just kind of how I am. And I don't know, like I'm a very relaxed sort of stress-free person. I rolled with the punches for a lot of this. I kind of took it in stride. It's like, "All right, well what can I do next? How can I improve my life even a little bit on a day-to-day basis?"
Paragraph 1587
At first, just trying to find some way to heal as much of my body as possible, to try to get healed, to try to get off a ventilator, learn as much as I could so I could somehow survive once I left the hospital. And then thank God I had like my family around me. If I didn't have my parents, my siblings, then I would've never made it this far. They've done so much for me, more than like I can ever thank them for, honestly.
Paragraph 1588
And a lot of people don't have that. A lot of people in my situation, their families either aren't capable of providing for them, or honestly, just don't want to. And so they get placed somewhere in some sort of home. So thankfully I had my family. I have a great group of friends, a great group of buddies from college who have all rallied around me, and we're all still incredibly close.
Paragraph 1589
People always say, "If you're lucky, you'll end up with one or two friends from high school that you keep throughout your life." I have about 10 or 12 from high school that have all stuck around, and we still get together, all of us twice a year. We call it the spring series and the fall series. This last one we all did, we dressed up like X-Men. So I did a Professor Xavier, and it was freaking awesome. It was so good.
Paragraph 1590
So yeah, I have such a great support system around me. And so being a quadriplegic isn't that bad. I get waited on all the time. People bring me food and drinks, and I get to sit around and watch as much TV and movies and anime as I want. I get to read as much as I want. I mean, it's great.
Paragraph 1591
It's beautiful to see that you see the silver lining in all of this. Just going back, do you remember the moment when you first realized you were paralyzed from the neck down?
Paragraph 1592
Yeah, yep. I was face down in the water right when whatever something hit my head. I tried to get up and I realized I couldn't move and it just sort of clicked. I'm like, "All right, I'm paralyzed. Can't move. What do I do? If I can't get up, I can't flip over, can't do anything, then I'm gonna drown eventually." And I knew I couldn't hold my breath forever, so I just held my breath and thought about it for maybe 10, 15 seconds.
Paragraph 1593
I've heard from other people that like look on liquors, I guess the two girls that pulled me out of the water were two of my best friends. They're lifeguards. And one of them said that it looked like my body was sort of shaking in the water, like I was trying to flip over and stuff. But I knew, I knew immediately. And I just kind of, I realized that that's what my situation was from here on out.
Paragraph 1594
Maybe if I got to the hospital, they'd be able to do something. When I was in the hospital, like right before surgery, I was trying to calm one of my friends down. I had like brought her with me from college to camp, and she was just bawling over me, and I was like, "Hey, it's gonna be fine. Like don't worry." I was cracking some jokes to try to lighten the mood. The nurse had called my mom, and I was like, "Don't tell my mom.
Paragraph 1595
She's just gonna be stressed out. Call her after I'm out of surgery," 'cause at least she'll have some answers then, like whether I live or not, really. And I didn't want her to be stressed through the whole thing. But I knew. And then when I first woke up after surgery, I was super drugged up. They had me on fentanyl like three ways, which was awesome.
Paragraph 1596
I don't recommend it, but I saw some crazy stuff on that fentanyl, and it was still the best I've ever felt on drugs. Medication, sorry, on medication. And I remember the first time I saw my mom in the hospital. I was just bawling. I had like ventilator in, like I couldn't talk or anything, and I just started crying, because it was more like seeing her.
Paragraph 1597
I mean, the whole situation obviously was pretty rough, but it was just like seeing her face for the first time was pretty hard. But yeah, I never had like a moment of, "Man, I'm paralyzed. This sucks. I don't wanna like be around anymore." It was always just, "I hate that I have to do this, but like sitting here and wallowing isn't gonna help."
Paragraph 1598
So immediate acceptance.
Paragraph 1599
[Noland] Yeah, yeah.
Paragraph 1600
Has there been low points along the way?
Paragraph 1601
Yeah, yeah, sure. I mean, there are days when I don't really feel like doing anything. Not so much anymore. Like not for the last couple years, I don't really feel that way. I've more so just wanted to try to do anything possible to make my life better at this point. But at the beginning, there were some ups and downs. There were some really hard things to adjust to.
Paragraph 1602
First off, just like the first couple months, the amount of pain I was in was really, really hard. I mean, I remember screaming at the top of my lungs in the hospital, because I thought my legs were on fire. And obviously, I can't feel anything, but it's all nerve pain. And so that was a really hard night. I asked them to give me as much pain meds as possible. They're like, "You've had as much as you can have, so just kind of deal with it.
Paragraph 1603
Go to a happy place sort of thing." So that was a pretty low point. And then every now and again, it's hard, like realizing things that I wanted to do in my life that I won't be able to do anymore. I always wanted to be a husband and father, and I just don't think that I could do it now as a quadriplegic. Maybe it's possible, but I'm not sure I would ever put someone I love through that, like having to take care of me and stuff.
Paragraph 1604
Not being able to go out and play sports. I was a huge athlete growing up, so that was pretty hard. Little things too, when I realized I can't do them anymore. Like there's something really special about being able to hold a book and smell a book. Like the feel, the texture, the smell, like as you turn the pages, like I just love it. I can't do it anymore. And it's little things like that. The two-year mark was pretty rough.
Paragraph 1605
Two years is when they say you will get back basically as much as you're ever gonna get back, as far as movement and sensation goes. And so for the first two years, that was the only thing on my mind was like try as much as I can to move my fingers, my hands, my feet, everything possible to try to get sensation and movement back. And then when the two-year mark hit, so June 30th, 2018, I was really sad that that's kind of where I was.
Paragraph 1606
And then just randomly here and there, but I was never like depressed for long periods of time. Just it never seemed worthwhile to me.
Paragraph 1607
What gave you strength?
Paragraph 1608
My faith. My faith in God was a big one. My understanding that it was all for a purpose. And even if that purpose wasn't anything involving Neuralink, even if that purpose was, there's a story in the Bible about Job, and I think it's a really, really popular story, about how Job has all of these terrible things happen to him, and he praises God throughout the whole situation.
Paragraph 1609
I thought, and I think a lot of people think for most of their lives that they are Job, that they're the ones going through something terrible, and they just need to praise God through the whole thing and everything will work out. At some point after my accident, I realized that I might not be Job, that I might be one of his children that gets killed or kidnapped or taken from him.
Paragraph 1610
And so it's about terrible things that happen to those around you who you love. So maybe, in this case, my mom would be Job, and she has to get through something extraordinarily hard and I just need to try and make it as best as possible for her, because she's the one that's really going through this massive trial. And that gave me a lot of strength. And obviously, my family.
Paragraph 1611
My family and my friends, they give me all the strength that I need on a day-to-day basis. So it makes things a lot easier having that great support system around me.
Paragraph 1612
From everything I've seen of you online, your streams and the way you are today, I really admire, let's say, your unwavering positive outlook on life. Has that always been this way?
Paragraph 1613
Yeah, yeah. I mean, I've just always thought I could do anything I ever wanted to do. There was never anything too big. Like whatever I set my mind to, I felt like I could do it. I didn't wanna do a lot. I wanted to like travel around and be sort of like a gypsy and like go work odd jobs. I had this dream of traveling around Europe and being like, I don't know, a shepherd in like Wales or Ireland.
Paragraph 1614
And then going and being a fisherman in Italy, doing all these things for like a year. Like it's such like cliche things, but I just thought it would be so much fun to go and travel and do different things. And so I've always just seen the best in people around me too. And I've always tried to be good to people. And growing up with my mom too, she's like the most positive energetic person in the world. And we're all just people people.
Paragraph 1615
I just get along great with people. I really enjoy meeting new people, and so I just wanted to do everything. This is just kind of just how I've been.
Paragraph 1616
It's just great to see that cynicism didn't take over, given everything you've been through.
Paragraph 1617
Yeah.
Paragraph 1618
Was that like a deliberate choice you made that you're not gonna let this keep you down?
Paragraph 1619
Yeah, a bit. Also, like it's just kind of how I am. Like I said, I roll with the punches with everything. I always used to tell people, like I don't stress about things much. And whenever I'd see people getting stressed, I would just say, "It's not hard. Just don't stress about it." And that's all you need to do. And they're like, "That's not how that works." I'm like, "It works for me. Like just don't stress, and everything will be fine.
Paragraph 1620
Like everything will work out." Obviously, not everything always goes well, and it's not like it all works out for the best all the time, but I just don't think stress has had any place in my life since I was a kid.
Paragraph 1621
What was the experience like of you being selected to be the first human being to have a Neuralink device implanted in your brain? Were you scared, excited?
Paragraph 1622
No, no, it was cool. (Lex laughing) Like I was never afraid of it. I had to think through a lot. Should I do this? Like be the first person? I could wait until number two or three and get a better version of the Neuralink. Like the first one might not work. Maybe it's actually gonna kind of suck it. It's gonna be the worst version ever in a person. So why would I do the first one? Like I've already kind of been selected.
Paragraph 1623
I could just tell them, like, "Okay, find someone else, and then I'll do number two or three." Like I'm sure they would let me. They're looking for a few people anyways. But ultimately, I was like, I don't know, there's something about being the first one to do something. It's pretty cool.
Paragraph 1624
I always thought that if I had the chance that I would like to do something for the first time, this seemed like a pretty good opportunity, and I was never scared. I think my like faith had a huge part in that. I always felt like God was preparing me for something. I almost wish it wasn't this, because I had many conversations with God about not wanting to do any of this as a quadriplegic. I told him, "I'll go out and talk to people.
Paragraph 1625
I'll go out and travel the world and talk to stadiums, thousands of people, give my testimony, I'll do all of it, but like heal me first. Don't make me do all of this in a chair. That sucks." And I guess he won that argument. I didn't really have much of a choice. I always felt like there was something going on.
Paragraph 1626
And to see how, I guess, easily I made it through the interview process and how quickly everything happened, how the stars sort of aligned with all of this, it just told me like, as the surgery was getting closer, it just told me that it was all meant to happen. It was all meant to be. And so I shouldn't be afraid of anything that's to come. And so I wasn't.
Paragraph 1627
I kept telling myself like, "You say that now, but as soon as the surgery comes, you're probably gonna be freaking out. Like you're about to have brain surgery." And brain surgery is a big deal for a lot of people, but it's a even bigger deal for me. Like it's all I have left.
Paragraph 1628
The amount of times I've been like, "Thank you God that you didn't take my brain and my personality and my ability to think, my like love of learning, like my character, everything, like thank you so much. Like as long as you left me that, then I think I can get by." And I was about to let people go like root around, and they're like, "Hey, we're gonna go put some stuff in your brain. Hopefully, it works out."
Paragraph 1629
And so it was something that gave me pause. But like I said, how smoothly everything went, I never expected for a second that anything would go wrong. Plus, the more people I met on the Barrow's side and on the Neuralink side, they're just the most impressive people in the world. Like I can't speak enough to how much I trust these people with my life and how impressed I am with all of them.
Paragraph 1630
And to see the excitement on their faces, to like walk into a room and roll into a room and see all of these people looking at me, like we're so excited. Like we've been working so hard on this, and it's finally happening. It's super infectious, and it just makes me wanna do it even more and to help them achieve their dreams. Like I don't know, it's so rewarding. And I'm so happy for all of them, honestly.
Paragraph 1631
What was the day of surgery like? When did you wake up? What'd you feel?
Paragraph 1632
Yeah.
Paragraph 1633
Minute by minute.
Paragraph 1634
Yeah.
Paragraph 1635
Were you freaking out?
Paragraph 1636
No, no. I thought I was going to, but as surgery approached the night before, the morning of, I was just excited. Let's make this happen. I think I said something like that to Elon on the phone beforehand. We were like FaceTiming, and I was like, "Let's rock and roll." And he's like, "Let's do it." I don't know, I wasn't scared. So we woke up. I think we had to be at the hospital at like 5:30 AM. I think surgery was at like 7:00 AM.
Paragraph 1637
So we woke up pretty early. I'm not sure much of us slept that night. Got to the hospital 5:30, went through like all the pre-op stuff. Everyone was super nice. Elon was supposed to be there in the morning, but something went wrong with his plane, so we ended up FaceTiming. That was cool. Had one of the greatest one-liners of my life. After that phone call, hung up with him.
Paragraph 1638
There were like 20 people around me, and I was like, I just hope he wasn't too starstruck talking to me.
Paragraph 1639
Nice.
Paragraph 1640
Yeah, it was good.
Paragraph 1641
Well done.
Paragraph 1642
Yeah, yeah.
Paragraph 1643
Did you write that ahead of time, or it just came to you?
Paragraph 1644
No, it just came to me. I was like, "This seems right." When in surgery, I asked if I could pray right beforehand. So I like prayed over the room. I asked God if you would like be with my mom in case anything happened to me. And just to like calm her nerves out there. Woke up and played a bit of a prank on my mom. I don't know if you've heard about it.
Paragraph 1645
Yeah, I read about it.
Paragraph 1646
Yeah, she was not happy.
Paragraph 1647
Can you take me through the prank - Yeah, this is something- - Do you regret doing that now?
Paragraph 1648
No, no, not one bit. It was something I had talked about ahead of time with my buddy, Bain. I was like, "I would really like to play a prank on my mom." Very specifically, my mom. She's very gullible. I think she had knee surgery once even. And after she came out of knee surgery, she was super groggy. She's like, "I can't feel my legs." And my dad looked at her, he was like, "You don't have any legs. Like they had to amputate both your legs."
Paragraph 1649
We just do very mean things to her all the time. I'm so surprised that she still loves us. But right after surgery, I was really worried that I was going to be too like groggy, like not all there. I had anesthesia once before, and it messed me up. Like I could not function for a while afterwards. And I like said a lot of things that I was like, I was really worried that I was gonna start, I don't know, like dropping some bombs.
Paragraph 1650
And I wouldn't even know, I wouldn't remember. So I was like, "Please God, don't let that happen. And please let me be there enough to do this to my mom." And so she walked in after surgery. It was like the first time they had been able to see me after surgery. And she just looked at me, she said, "Hi, like, how are you? How are you doing? How do you feel?"
Paragraph 1651
And I looked at her in this very, I think the anesthesia helped, very like groggy, sort of confused look on my face. It's like, "Who are you?" And she just started looking around the room, like at the surgeons, at the doctors, like, "What did you do to my son? You need to fix this right now." Tears started streaming. I saw how much she was freaking out. I was like, "I can't let this go on." And so I was like, "Mom, I'm fine. Like it's all right."
Paragraph 1652
And still, she was not happy about it. She still says she's gonna get me back someday. But I mean, I don't know. I don't know what that's gonna look like.
Paragraph 1653
It's a lifelong battle.
Paragraph 1654
Yeah, it was good.
Paragraph 1655
In some sense, it was a demonstration that you still got- - That's all I wanted it to be - Sense of humor.
Paragraph 1656
That's all I wanted it to be. And I knew that doing something super mean to her like that would show her- - - - Yeah.
Paragraph 1657
To show that you're still there, that you love her.
Paragraph 1658
[Noland] Yeah, exactly, exactly.
Paragraph 1659
It's a dark way to do it, but I love it. What was the first time you were able to feel that you can use the Neuralink device to effect the world around you?
Paragraph 1660
Yeah, the first little taste I got of it was actually not too long after surgery. Some of the Neuralink team had brought in like a little iPad, a little tablet screen, and they had put up eight different channels that were recording some of my neuron spikes. They put it in front of me. Like this is like real time your brain firing. That's super cool.
Paragraph 1661
My first thought was, "I mean, if they're firing now, let's see if I can affect them in some way." So I started trying to like wiggle my fingers and I just started like scanning through the channels, and one of the things I was doing was like moving my index finger up and down. And I just saw this yellow spike on like top row, like third box over or something. I saw this yellow spike every time I did it. And I was like, "Oh, that's cool."
Paragraph 1662
And everyone around me was just like, "Well, what are you seeing?" I was like, "Look, look at this one. Look at like this top row, third box over this yellow spike. Like that's me right there, there, there." And everyone was freaking out. They started like clapping. I was like, "That's super unnecessary."
Paragraph 1663
That's awesome.
Paragraph 1664
This is what's supposed to happen, right?
Paragraph 1665
So you're imagining yourself moving each individual finger one at a time, and then seeing like that you can notice something, and then when you did the index finger, you're like, "Oh."
Paragraph 1666
Yeah, I was wiggling kind of all of my fingers to see if anything would happen. There was a lot of other things going on, but that big yellow spike was the one that stood out to me. Like I'm sure that if I would've stared at it long enough, I could have mapped out maybe 100 different things. But the big yellow spike was the one that I noticed.
Paragraph 1667
Maybe you could speak to what it's like to sort of wiggle your fingers, to imagine that the mental, the cognitive effort required to sort of wiggle your index finger, for example. How easy is that to do?
Paragraph 1668
Pretty easy for me. It's something that, at the very beginning, after my accident, they told me to try and move my body as much as possible. Even if you can't, just keep trying, because that's going to create new like neural pathways or pathways in my spinal cord to like reconnect these things to hopefully regain some movement someday.
Paragraph 1669
That's fascinating.
Paragraph 1670
Yeah, I know. It's bizarre, but I- - So that's part of the recovery process is to keep trying to move your body?
Paragraph 1671
Yep, as much as you can.
Paragraph 1672
And that's, and the nervous system does its thing. It starts reconnecting.
Paragraph 1673
Yeah. It'll start reconnecting for some people. Some people, it never works. Some people, they'll do it. Like for me, I got some bicep control back, and that's about it. I can, if I try enough, I can wiggle some of my fingers. Not like on command. It's more like if I try to move, say my right pinky and I just keep trying to move it, after a few seconds, it'll wiggle. So I know there's stuff there.
Paragraph 1674
Like I know, and that happens with a few - -different of my fingers and stuff. But yeah, that's what they tell you to do. One of the people at the time when I was in the hospital came in and told me, for one guy who had recovered most of his control, what he thought about every day was actually walking, like the act of walking just over and over again. So I tried that for years. I tried just imagining walking, which it's hard.
Paragraph 1675
It's hard to imagine like all of the steps that go into, well, taking a step, like all of the things that have to move, like all of the activations that have to happen along your leg in order for one step to occur.
Paragraph 1676
But you're not just imagining. You're like doing it, right?
Paragraph 1677
I'm trying, yeah. So it's like, it's imagining over again what I had to do to take a step, because it's not something any of us think about. You wanna walk and you take a step. You don't think about all of the different things that are going on in your body. So I had to recreate that in my head as much as I could. And then I practice it over and over and over.
Paragraph 1678
So it's not like a third person perspective. It's a first person perspective. You're like, it's not like you're imagining yourself walking. You're like literally doing this, everything, all the same stuff as if you're walking.
Paragraph 1679
Yeah, which was hard. It was hard at the beginning.
Paragraph 1680
Like frustrating hard, or like actually cognitively hard? Like which way?
Paragraph 1681
It was both. There's a scene in one of the "Kill Bill" movies actually, oddly enough, where she is like paralyzed, I don't know, from like a drug that was in her system. And then she like finds some way to get into the back of a truck or something, and she stares at her toe and she says, "Move." Like move your big toe. And after a few seconds on screen, she does it. And she did that with every one of her like body parts until she can move again.
Paragraph 1682
I did that for years, just stared at my body and said, "Move your index finger. Move your big toe." Sometimes, vocalizing it like out loud, sometimes just thinking it. I tried every different way to do this to try to get some movement back.
Paragraph 1683
And it's hard because it actually is like taxing, like physically taxing on my body, which is something I would've never expected, 'cause it's not like I'm moving, but it feels like there's a buildup of, I don't know, the only way I can describe it is there are like signals that aren't getting through from my brain down, 'cause there's that gap in my spinal cord. So brain down, and then from my hand back up to the brain.
Paragraph 1684
And so it feels like those signals get stuck in whatever body part that I'm trying to move. And they just build up and build up and build up until they burst. And then once they burst, I get like this really weird sensation of everything sort of like dissipating back out to level, and then I do it again. It's also just like a fatigue thing, like a muscle fatigue, but without actually moving your muscles. It's very, very bizarre.
Paragraph 1685
And then if you try to stare at a body part or think about a body part and move for two, three, four, sometimes eight hours, it's very taxing on your mind. It's takes a lot of focus. It was a lot easier at the beginning because I wasn't able to like control a TV in my room or anything. I wasn't able to control any of my environment. So for the first few years, a lot of what I was doing was staring at walls.
Paragraph 1686
And so obviously, I did a lot of thinking, and I tried to move a lot just over and over and over again.
Paragraph 1687
So you never gave up sort of hope there?
Paragraph 1688
No.
Paragraph 1689
Just training hard essentially?
Paragraph 1690
Yep, and I still do it. I do it like subconsciously. And I think that that helped a lot with things with Neuralink, honestly. It's something that I talked about the other day at the all hands that I did at Neuralink's Austin facility.
Paragraph 1691
Welcome to Austin, by the way.
Paragraph 1692
Yeah, hey, thanks man. I went to school- - Nice hat.
Paragraph 1693
Hey, thanks, thanks man. The gigafactory was super cool. I went to school at Texas A&M, so I've been around before.
Paragraph 1694
So you should be saying welcome to me.
Paragraph 1695
Yeah.
Paragraph 1696
Welcome to Texas, Lex. Yeah, I get you.
Paragraph 1697
But yeah, I was talking about how a lot of what they've had me do, especially at the beginning, well, I still do it now is body mapping. So like there will be a visualization of a hand or an arm on the screen and I have to do that motion, and that's how they sort of train the algorithm to like understand what I'm trying to do. And so it made things very seamless for me, I think.
Paragraph 1698
That's really, really cool. So yeah, it's amazing to know 'cause I've learned a lot about the body mapping procedure with the interface and everything like that. It's cool to know that you've been essentially like training to be like world class at that task.
Paragraph 1699
Yeah, yeah. I don't know if other quadriplegics, like other paralyzed people give up. I hope they don't. I hope they keep trying, because I've heard other paralyzed people say, like don't ever stop. They tell you two years, but you just never know. The human body is capable of amazing things. So I've heard other people say, "Don't give up."
Paragraph 1700
Like I think one girl had spoken to me through some family members and said that she had been paralyzed for 18 years, and she'd been trying to like wiggle her index finger for all that time, and she finally got it back like 18 years later. So like I know that it's possible, and I'll never give up doing it. I do it when I'm lying down. Like watching TV, I'll find myself doing it, kind of just almost like on its own.
Paragraph 1701
It's just something I've gotten so used to doing that I don't know, I don't think I'll ever stop.
Paragraph 1702
That's really awesome to hear, 'cause I think it's one of those things that can really pay off, in the long term. 'Cause like that is training. You're not visibly seeing the results of that training at the moment, but like, there's that like Olympic level nervous system getting ready for something.
Paragraph 1703
Which honestly was like something that I think Neuralink gave me that I can't thank them enough for it. Like I can't show my appreciation for it enough was being able to visually see that what I'm doing is actually having some effect. It's a huge part of the reason why, like I know now that I'm gonna keep doing it forever, because before Neuralink, I was doing it every day, and I was just assuming that things were happening.
Paragraph 1704
Like it's not like I knew. I wasn't getting back any mobility or sensation or anything. So I could have been running up against a brick wall for all I knew. And with Neuralink, I get to see like all the signals happen in real time, and I get to see that what I'm doing can actually be mapped. When we started doing like click calibrations and stuff, when I go to click my index finger for a left click, that it actually recognizes that.
Paragraph 1705
Like it changed how I think about what's possible with like retraining my body to move. And so yeah, I'll never give up now.
Paragraph 1706
And also, just the signal that there's still a powerhouse of a brain there that's like- - Exactly.
Paragraph 1707
And as the technology develops, that brain is, I mean, that's the most important thing about the human body is the brain. And it can do a lot of the control. So what did it feel like when you first, could wiggle the index finger and saw the environment respond? Like that little- - Yeah.
Paragraph 1708
Wherever, just being way too dramatic according to you.
Paragraph 1709
Yeah, it was very cool. I mean, it was cool, but I keep telling this to people, it made sense to me. Like it made sense that like there are signals still happening in my brain, and that as long as you had something near it that could measure those, that could record those, then you should be able to visualize it in some way. Like see it happen. And so that was not very surprising to me. I was just like, "Oh, cool." We found one.
Paragraph 1710
Like we found something that works. It was cool to see that their technology worked, and that everything that they had worked so hard for was like going to pay off. But I hadn't like moved a cursor or anything at that point. I had like interacted with a computer or anything at that point. So it just made sense, it was cool. I didn't really know much about BCI at that point either, so I didn't know like what sort of step this was actually making.
Paragraph 1711
Like I didn't know if this was like a huge deal, or if this was just like, okay, it's cool that we got this far, but we're actually hoping for something like much better down the road. It's like, okay. I just thought that they knew that it turned on. So I was like, cool. Like this is cool.
Paragraph 1712
Well, did you like read up on the specs of the hardware you're getting installed? Like the number of threads, this kind of stuff?
Paragraph 1713
Yeah, I do all of that, but it's all Greek to me. I was like, okay, threads, 64 threads, 16 electrodes, 1,024 channels. Okay. Like that math checks out.
Paragraph 1714
Sounds right.
Paragraph 1715
Yeah.
Paragraph 1716
When was the first time you were able to move a mouse cursor?
Paragraph 1717
I know it must have been within the first maybe week, a week or two weeks that I was able to like first move the cursor. And again, like it kind of made sense to me. It didn't seem like that big of a deal. It was like, okay, well, hmm, how do I explain this? When everyone around you starts clapping for something that you've done, it's easy to say, "Okay, like I did something cool. Like that was impressive in some way."
Paragraph 1718
What exactly that meant, what it was hadn't really like set in for me. So again, I knew that me trying to move a body part and then that being mapped in some sort of like machine learning algorithm to be able to identify like my brain signals and then take that and give me cursor control, that all kind of made sense to me. I don't know like all the ins and outs of it, but I was like, there are still signals in my brain firing.
Paragraph 1719
They just can't get through because there's like a gap in my spinal cord. And so they can't get all the way down and back up, but they're still there. So when I moved the cursor for the first time, I was like, "That's cool, but I expected that that should happen." Like it made sense to me. When I moved the cursor for the first time with just my mind, without like physically trying to move, so I guess I can get into that just a little bit.
Paragraph 1720
Like the difference between attempted movement and imagined movement.
Paragraph 1721
Yeah, that's a fascinating difference.
Paragraph 1722
Yeah.
Paragraph 1723
From one to the other.
Paragraph 1724
Yeah, yeah, yeah. So like attempted movement is me physically trying to attempt to move, say, my hand. I try to attempt to move my hand to the right, to the left, forward and back. And that's all attempted. Attempt to like lift my finger up and down. Attempt to kick or something. I'm physically trying to do all of those things, even if you can't see it. This would be like me attempting to like shrug my shoulders or something.
Paragraph 1725
That's all attempted movement. That's what I was doing for the first couple of weeks when they were going to give me cursor control. When I was doing body mapping, it was attempt to do this, attempt to do that. When Nir was telling me to like imagine doing it, it like kind of made sense to me, but it's not something that people practice. Like if you started school as a child and they said, "Okay, write your name with this pencil."
Paragraph 1726
And so you do that. "Okay, now imagine writing your name with that pencil." Kids would think like, I guess like that kind of makes sense. And they would do it. But that's not something we're taught. It's all like how to do things physically. We think about like thought experiments and things, but that's not like a physical action of doing things. It's more like what you would do in certain situations.
Paragraph 1727
So imagined movement, it never really connected with me. Like I guess you could maybe describe it as like a professional athlete, like swinging a baseball bat or swinging like a golf club. Like imagine what you're supposed to do. But then you go right to that and physically do it, then you get a bat in your hand and then you do what you've been imagining. And so I don't have that like connection.
Paragraph 1728
So telling me to imagine something versus attempting it, there wasn't a lot that I could do there mentally. I just kind of had to accept what was going on and try. But the attempted moving thing, it all made sense to me. Like if I try to move, then there's a signal being sent in my brain. And as long as they can pick that up, then they should be able to map it to what I'm trying to do.
Paragraph 1729
And so when I first moved the cursor like that, it was just like, "Yes, this should happen." Like I'm not surprised by that.
Paragraph 1730
But can you clarify, is there supposed to be a difference between imagined movement and attempted movement?
Paragraph 1731
Yeah, just that in imagined movement, you're not attempting to move at all.
Paragraph 1732
You're like visualizing doing. And then theoretically, is that supposed to be a different part of the brain that lights up in those two different situations?
Paragraph 1733
[Bliss] Yeah, not necessarily. I think all these signals can still be represented in motor cortex, but the difference I think has to do with the naturalness of imagining something versus attempting it- - Got it.
Paragraph 1734
[Bliss] And sort of the fatigue of that over time.
Paragraph 1735
And by the way, on the mic is Bliss. So this is just different ways to prompt you to kind of get to the thing that you arrived at.
Paragraph 1736
[Noland] Yeah, yeah.
Paragraph 1737
Attempted movement does sound like the right thing - try.
Paragraph 1738
Yeah, I mean, it makes sense to me.
Paragraph 1739
'Cause imagine for me, I would start visualizing, like in my mind, visualizing. Attempted, I would actually start trying to like, there's, I mean, I did like combat sports my whole life, like wrestling. When I'm imagining a move, see, I'm like moving my muscle.
Paragraph 1740
Exactly.
Paragraph 1741
Like there is a bit of an activation almost, versus like visualizing yourself like a picture doing it.
Paragraph 1742
Yeah, it's something that I feel like naturally, anyone would do. If you try to tell someone to imagine doing something, they might close their eyes and then start physically doing it. But it's just- - Just didn't click.
Paragraph 1743
Yeah. It's hard. It was very hard at the beginning.
Paragraph 1744
But attempted worked.
Paragraph 1745
Attempted worked. It worked just like it should. Worked like a charm.
Paragraph 1746
[Bliss] I remember there was like one Tuesday we were messing around, and I think, I forget what swear word you used, but there's a swear word that came out of your mouth when you figured out you could just do the direct cursor control.
Paragraph 1747
Yeah, that's it. It blew my mind. Like no pun intended, blew my mind when I first moved the cursor just with my thoughts and not attempting to move. It's something that I've found like over the couple of weeks, like building up to that, that as I get better cursor controls, like the model gets better, then it gets easier for me to like, like I don't have to attempt as much to move it.
Paragraph 1748
And part of that is something that I'd even talked with them about when I was watching the signals of my brain one day, I was watching, when I like attempted to move to the right and I watched the screen, it's like I saw the spikes. It's like I was seeing the spike, the signals being sent before I was actually attempting to move.
Paragraph 1749
I imagined just because when you go to say move your hand or any body part, that signal gets sent before you're actually moving has to make it all the way down and back up before you actually do any sort of movement. So there's a delay there. And I noticed that there was something going on in my brain before I was actually attempting to move, that my brain was like anticipating what I wanted to do.
Paragraph 1750
And that all started sort of, I don't know, like percolating in my brain. It was just sort of there, like always in the back. Like that's so weird that it could do that. It kind of makes sense, but I wonder what that means as far as like using the Neuralink.
Paragraph 1751
And then as I was playing around with the attempted movement and playing around with the cursor, and I saw that like, as the cursor control got better, that it was anticipating my movements and what I wanted it to do. Like cursor movements, what I wanted to do a bit better and a bit better. And then one day, I just randomly, as I was playing Webgrid, I like looked at a target before I had started like attempting to move.
Paragraph 1752
I was just trying to like get over, like train my eyes to start looking ahead. Like, okay, this is the target I'm on, but if I look over here to this target, I know I can like maybe be a bit quicker getting there. And I looked over and the cursor just shot over it. It was wild. I had to take a step back. Like I was like, "This should not be happening." All day, I was just smiling, I was so giddy. I was like, "Guys, do you know that this works?
Paragraph 1753
Like I can just think it and it happens," which like they'd all been saying this entire time. like I can't believe like you're doing all this with your mind. I'm like, "Yeah, but is it really with my mind?" Like I'm attempting to move, and it's just picking that up so it doesn't feel like it's with my mind. When I moved it for the first time like that, it was, oh man.
Paragraph 1754
It made me think that this technology, that what I'm doing is actually way, way more impressive than I ever thought. It was way cooler than I ever thought. And it just opened up a whole new world of possibilities of like what could possibly happen with this technology and what I might be able to be capable of with it.
Paragraph 1755
Because you had felt for the first time, like this was digital telepathy. Like you're controlling a digital device with your mind.
Paragraph 1756
[Noland] Yep.
Paragraph 1757
I mean, that's a real moment of discovery. That's really cool. Like you've discovered something. I've seen like scientists talk about like a big aha moment. Like Nobel Prize winning, they'll have this like holy crap.
Paragraph 1758
Yeah.
Paragraph 1759
Like whoa.
Paragraph 1760
That's what it felt. I didn't feel like, like I felt like I had discovered something but for me. Maybe not necessarily for like the world at large or like this field at large. It just felt like an aha moment for me. Like, "Oh this works." Like obviously, it works. And so that's what I do like all the time now. I kind of intermix the attempted movement and imagined movement.
Paragraph 1761
I do it all like together, because I've found that there is some interplay with it that maximizes efficiency with the cursor. So it's not all like one or the other. It's not all just, I only use attempted or I only use like imagined movements. It's more I use them in parallel, and I can do one or the other. I can just completely think about whatever I'm doing. But I don't know. I like to play around with it.
Paragraph 1762
I also like to just experiment with these things. Like every now and again, I'll get this idea in my head like, "Hmm, I wonder if this works." And I'll just start doing it, and then afterwards, I'll tell them, "By the way, I wasn't doing that like you guys wanted me to. I thought of something and I wanted to try it and so I did. It seems like it works, so maybe we should like explore that a little bit."
Paragraph 1763
So I think that discovery is not just for you, at least from my perspective, that's the discovery for everyone else who ever uses a Neuralink that this is possible. Like I don't think this an obvious thing that this is even possible. It's like, I was saying to Bliss earlier, it's like the four-minute mile. People thought it was impossible to run a mile in four minutes, and once the first person did it, then everyone just started doing it.
Paragraph 1764
So like just to show that it's possible, that paves the way to like anyone can not do it. That's the thing that's actually possible. You don't need to do the attempted movement. You can just go direct. That's crazy.
Paragraph 1765
It is crazy, it's crazy.
Paragraph 1766
For people who don't know, can you explain how the Link app works? You have an amazing stream on the topic. Your first stream, I think, on X describing the app. Can you just describe how it works?
Paragraph 1767
Yeah, so it's just an app that Neuralink created to help me interact with the computer. So on the Link app, there are a few different settings and different modes and things I can do on it. So there's like the body mapping, which we kind of touched on. There's a calibration. Calibration is how I actually get cursor control. So calibrating what's going on in my brain to translate that into cursor control. So it will pop out models.
Paragraph 1768
What they use I think is like time. So it would be five minutes, and calibration will give me so good of a model. And then if I'm in it for 10 minutes and 15 minutes, the models will progressively get better. And so the longer I'm in it generally, the better the models will get.
Paragraph 1769
That's really cool, 'cause you often refer to the models. The model's the thing that's constructed once you go through the calibration step. And then you also talked about, sometimes you'll play like a really difficult game, like Snake, just to see how good the model is.
Paragraph 1770
Yeah, yeah, so Snake is kind of like my litmus test for models. If I can control Snake decently well, then I know I have a pretty good model. So yeah, the Link app has all of those. It has Webgrid in it now. It's also how I like connect to the computer just in general. So they've given me a lot of like voice controls with it at this point, so I can say like connect or implant disconnect.
Paragraph 1771
And as long as I have that charger handy, then I can connect to it. So the charger is also how I connect to the Link app, to connect to the computer. I have to have the implant charger over my head when I wanna connect to have it wake up, 'cause the implant's in hibernation mode, like always when I'm not using it. I think there's a setting to like wake it up every so long.
Paragraph 1772
So we could set it to half an hour or five hours or something if I just want it to wake up periodically. So yeah, I'll like connect to the Link app, and then go through all sorts of things. Calibration for the day, maybe body mapping. I made them give me like a little homework tab, because I am very forgetful and I forget to do things a lot. So I have like a lot of data collection things that they want me to do.
Paragraph 1773
Is the body mapping part of the data collection, or is that also part of the- - Yeah, it is. It's something that they want me to do daily, which I've been slacking on, 'cause I've been doing so much media and traveling so much. So I've been- - You've been super famous.
Paragraph 1774
Yeah, I've been a terrible first candidate for how much I've been slacking on my homework. But yeah, it's just something that they want me to do every day to track how well the Neuralink is performing over time and to have something to give. I imagine to give to the FDA to create all sorts of fancy charts and stuff and show like, "Hey, this is what the Neuralink, this is how it's performing day one versus day 90 versus day 180 and things like that.
Paragraph 1775
What's the calibration step like? Is it like move left, move right?
Paragraph 1776
It's a bubble game. So there will be like yellow bubbles that pop up on the screen. At first, it is open loop. So open loop, this is something that I still don't fully understand, the open loop and closed loop thing.
Paragraph 1777
And me and Bliss talked for a long time about the difference between the two on the technical side.
Paragraph 1778
Okay.
Paragraph 1779
So it'd be great to hear your side of the story.
Paragraph 1780
Open loop is basically, I have no control over the cursor. The cursor will be moving on its own across the screen, and I am following by intention the cursor to different bubbles. And then the algorithm is training off of what like the signals it's getting are as I'm doing this. There are a couple different ways that they've done it. They call it center out target. So there will be a bubble in the middle and then eight bubbles around that.
Paragraph 1781
And the cursor will go from the middle to one side. So say middle to left, back to middle, to up, to middle, like up, right. And they'll do that all the way around the circle. And I will follow that cursor the whole time, and then it will train off of my intentions, what it is expecting my intentions to be throughout the whole process.
Paragraph 1782
Can you actually speak to, when you say follow- - Yes.
Paragraph 1783
You don't mean with your eyes. You mean with your intentions.
Paragraph 1784
Yeah, so generally for calibration, I'm doing attempted movements, 'cause I think it works better. I think the better models, as I progress through calibration, make it easier to use imagined movements.
Paragraph 1785
Wait, wait, wait, wait. So calibrated on attempted movement will create a model that makes it really effective for you to then use the force?
Paragraph 1786
Yes, I've tried doing calibration with imagined movement, and it just doesn't work as well for some reason. So that was the center out targets. There's also one where a random target will pop up on the screen and it's the same. I just like move, I follow along with wherever the cursor is to that target, all across the screen.
Paragraph 1787
I've tried those with imagined movement, and for some reason, the models just don't, they don't give as high level as quality when we get into closed loop. I haven't played around with it a ton, so maybe like the different ways that we're doing calibration now might make it a bit better. But what I've found is there will be a point in calibration where I can use imagined movement. Before that point, it doesn't really work.
Paragraph 1788
So if I do calibration for 45 minutes, the first 15 minutes, I can't use imagined movement. It just like doesn't work for some reason. And after a certain point, I can just sort of feel it. I can tell it moves different. That's the best way I can describe it. Like it's almost as if it is anticipating what I am going to do again before I go to do it.
Paragraph 1789
And so using attempted movement for 15 minutes, at some point, I can kind of tell when I like move my eyes to the next target that the cursor is starting to like pick up. Like it's starting to understand, it's learning like what I'm going to do.
Paragraph 1790
So first of all, it's really cool that, I mean, you are a true pioneer in all of this. You're like exploring how to do every aspect of this most effectively, and there's just, I imagine so many lessons learned from this. So thank you for being a pioneer in all these kinds of different like super technical ways.
Paragraph 1791
And it's also cool to hear that there's like a different like feeling to the experience when it's calibrated in different ways, 'cause I mean, I imagine your brain is doing something different, and that's why there's a different feeling to it. And then trying to find the words and the measurements to those feelings would be also interesting.
Paragraph 1792
But at the end of the day, you can also measure that your actual performance on whether it's Snake or Webgrid, you could see like what actually works well. And you're saying, for the open loop calibration, the attempted movement works best for now.
Paragraph 1793
Yep, yep.
Paragraph 1794
So the open loop, you don't get the feedback that you did something.
Paragraph 1795
Yeah- - Is that frustrating?
Paragraph 1796
No, no, it makes sense to me. Like we've done it with a cursor and without a cursor in open loop. So sometimes, it's just, say, for like the center out, you'll start calibration with a bubble lighting up, and I push towards that bubble. And then when that bubble, when it's pushed towards that bubble for say, three seconds a bubble will pop and then I come back to the middle. So I'm doing it all just by my intentions. Like that's what it's learning anyways. So it makes sense that as long as I follow what they want me to do, like follow the yellow brick road, that it'll all work out.
Paragraph 1797
You're full of great references. Is the bubble game fun?
Paragraph 1798
Yeah, they always feel so bad making me do calibration. Like, "Oh, we're about to do a 40-minute calibration." I'm like, "All right, do you guys wanna do two of them?" Like I'm always asking to, like whatever they need, I'm more than happy to do. And it's not bad. Like I get to lie there or sit in my chair and like do these things with some great people. I get to have great conversations. I can give them feedback. I can talk about all sorts of things. I could throw something on on my TV in the background and kinda like split my attention between them. Like it's not bad at all. I don't mind it.
Paragraph 1799
Is there a score that you get? Like can you do better on the bubble game?
Paragraph 1800
No, I would love that. I would love- - Yeah. Writing down suggestions from Noland.
Paragraph 1801
That's- - Make it more fun. Gamified.
Paragraph 1802
Yeah, that's one thing that I really, really enjoy about Webgrid is 'cause I'm so competitive. Like the higher the BPS, the higher the score, I know the better I'm doing. I think I've asked at one point one of the guys, like if he could give me some sort of numerical feedback for calibration, like I would like to know what they're looking at.
Paragraph 1803
Like, "Oh, we see like this number while you're doing calibration, and that means, at least on our end, that we think calibration is going well." And I would love that, because I would like to know if what I'm doing is going well or not. But then they've also told me like, "Yeah, not necessarily like one-to-one." It doesn't actually mean that calibration is going well in some ways.
Paragraph 1804
So it's not like 100%, and they don't wanna like skew what I'm experiencing or want me to change things based on that. If that number isn't always accurate to like how the model will turn out or how like the end result, that's at least what I got from it. One thing I do that I have asked them and something that I really enjoy striving for is towards the end of calibration, there is like a time between targets.
Paragraph 1805
And so I like to keep, like at the end, that number as low as possible. So at the beginning, it can be four or five, six seconds between me popping bubbles. But towards the end, I like to keep it below like 1. 5. Or if I could, get it to like one second between like bubbles, because in my mind, that translates really nicely to something like Webgrid where I know if I can hit a target one every second that I'm doing real, real well.
Paragraph 1806
There you go, that's a way to get a score on the calibrations. Like the speed, how quickly can you get from bubble to bubble.
Paragraph 1807
Yeah.
Paragraph 1808
So there's the open loop, and then it goes to the closed loop.
Paragraph 1809
Closed loop.
Paragraph 1810
The closed loop can already start giving you a sense, 'cause you're getting feedback of like how good the model is.
Paragraph 1811
Yeah, so closed loop is when I first get cursor control and how they've described it to me, someone who does not understand this stuff, I am the dumbest person in the room every time I'm with- - I love the humility.
Paragraph 1812
Yeah, is that I am closing the loop. So I am actually now the one that is like finishing the loop of whatever this loop is. I don't even know what the loop is, they've never told me. They just say there is a loop. And at one point, it's open and I can't control. And then I get control and it's closed. So I'm finishing the loop.
Paragraph 1813
So how long the calibration usually take? You said like 10, 15 minutes.
Paragraph 1814
Well, yeah, they're trying to get that number down pretty low. That's what we've been working on a lot recently is getting that down as low as possible, so that way, if this is something that people need to do on a daily basis or if some people need to do on a like every other day basis or once a week, they don't want people to be sitting in calibration for long periods of time.
Paragraph 1815
I think they've wanted to get it down seven minutes or below, at least where we're at right now. It'd be nice if you never had to do calibration. So we'll get there at some point, I'm sure, the more we learn about the brain and like I think that's the dream. I think right now, for me to get like really, really good models, I am in calibration 40 or 45 minutes. And I don't mind.
Paragraph 1816
Like I said, they always feel really bad, but if it's gonna get me a model that can like break these records on Webgrid, I'll stay in it for flipping two hours.
Paragraph 1817
Let's talk business. So Webgrid. I saw a presentation where Bliss said by March, you selected 89,000 targets in Webgrid. Can you explain this game? What is Webgrid, and what does it take to be a world class performer in Webgrid, as you continue to break world records?
Paragraph 1818
Yeah.
Paragraph 1819
It's like a gold medalist, like well.
Paragraph 1820
Yeah, I'd like to thank, I'd like to thank everyone who's helped me get here, my coaches, my parents for driving me to practice every day at five in the morning. Like to thank God. And just overall, my dedication to my craft.
Paragraph 1821
The interviews with athletes, they're always like that, it's like that template.
Paragraph 1822
Yeah.
Paragraph 1823
So Webgrid is a grid that sells.
Paragraph 1824
Webgrid is, yeah. It's literally just a grid. They can make it as big or small as you can make a grid. A single box on that grid will light up and you go and click it. And it is a way for them to benchmark how good a BCI is. So it's pretty straightforward. You just click targets.
Paragraph 1825
[Lex] Only one blue cell appears, and you're supposed to move the mouse to there and click on it.
Paragraph 1826
Yep. So I like playing on like bigger grids, 'cause the bigger the grid, the like more BPS. It's bits per second that you get every time you click one. So I'll say, I'll play on like a 35 by 35 grid, and then one of those little squares cell, we call it target, whatever, will light up and you move the cursor there and you click it and then you do that forever.
Paragraph 1827
And you've been able to achieve at first eight bits per second. And then you recently broke that.
Paragraph 1828
Yeah, I'm at 8. 5 right now. I would've beaten that literally the day before I came to Austin. But I had like, I don't know, like a five second lag right at the end. And I just had to wait until the latency calmed down and then I kept clicking. But I was at like 8. 01 and then five seconds of lag, and then the next like three targets I clicked all stayed at 8. 01.
Paragraph 1829
So if I would've been able to click during that time of lag, I probably would've hit, I don't know, I might've hit nine. So I'm there. I'm really close. And then this whole Austin trip has really gotten in the way of my Webgrid playing ability.
Paragraph 1830
It's frustrating.
Paragraph 1831
Yeah, it's- - So that's all you've been thinking about right now?
Paragraph 1832
Yeah, I know. I want to do better at nine. I want to do better. I wanna hit nine, I think. Well, I know nine is very, very achievable. I'm right there. I think 10 I could hit maybe in the next month. Like I could do it probably in the next few weeks if I really push it.
Paragraph 1833
I think you and Elon are basically the same person, 'cause last time I did a podcast with him, he came in extremely frustrated that he can't beat Uber Lilith as a droid. That was like a year ago I think, I forget, like solo. And I could just tell, there's some percentage of his brain the entire time was thinking like, "I wish I was right now attempting."
Paragraph 1834
I think he did it.
Paragraph 1835
He did it that night.
Paragraph 1836
Yeah.
Paragraph 1837
He stayed up and did it that night. It's just crazy to me. I mean, in a fundamental way, it's really inspiring. And what you're doing is inspiring in that way, 'cause I mean, it's not just about the game. Everything you're doing there has impact. By striving to do well on Webgrid, you're helping everybody figure out how to create the system all along, like the decoding, the software, the hardware, the calibration, all of it, how to make all of that work so you can do everything else really well.
Paragraph 1838
Yeah, it's just really fun.
Paragraph 1839
Well, that's also, that's part of the thing is like making it fun.
Paragraph 1840
Yeah, it's addicting. I've joked about like what they actually did when they went in and put this thing in my brain. They must have flipped a switch to make me more susceptible to these kinds of games, to make me addicted to like Webgrid or something. Do you know Bliss's high score?
Paragraph 1841
Yeah, he said like 14 or something.
Paragraph 1842
17.
Paragraph 1843
Oh boy.
Paragraph 1844
17.1 or something, 17.01.
Paragraph 1845
17 dot, 17.01.
Paragraph 1846
Yeah.
Paragraph 1847
He told me he like does it on the floor with peanut butter and he like fasts. It's weird. That sounds like cheating. Sounds like performance enhancing.
Paragraph 1848
[Bliss] No, like the first time Noland played this game, he asked, "How good are we at this game?" And I think you told me right then, "You're gonna try to beat me on that."
Paragraph 1849
I'm gonna get there someday.
Paragraph 1850
Yeah. I fully believe you.
Paragraph 1851
I think I can.
Paragraph 1852
I'm excited for that.
Paragraph 1853
Yeah. So I've been playing, first off, with the dwell cursor, which really hampers my Webgrid playing ability. Basically, I have to wait 0.3 seconds for every click.
Paragraph 1854
Oh, so you can't do the clicks. So you click by dwelling, you said 0.3?
Paragraph 1855
0.3 seconds, which sucks. It really slows down how much I'm able to, like how high I'm able to get. I still hit like 50, I think I hit like 50 something trials, net trials per minute in that, which was pretty good, 'cause I'm able to like, there's one of the settings is also like how slow you need to be moving in order to initiate a click, to start a click. So I can tell sort of when I'm on that threshold to start initiating a click just a bit early, so I'm not fully stopped over the target when I go to click. I'm doing it like on my way to the targets a little to try to time it just right.
Paragraph 1856
Oh wow. So you're slowing down.
Paragraph 1857
Yeah, just a hair right before the targets. (Lex laughing) - This is like elite performance, okay. But that's still, it sucks that there's a ceiling of the 0.3.
Paragraph 1858
Well, I can get down to 0.2 and 0.1. Point one's what- - I get it.
Paragraph 1859
Yeah, and I've played with that a little bit too. I have to adjust a ton of different parameters in order to play with 0. 1, and I don't have control over all that on my end yet. It also changes like how the models are trained.
Paragraph 1860
Like if I train a model like in Webgrid, like I bootstrap on a model, which basically is them training models as I'm playing Webgrid based off of like the Webgrid data, so like if I play Webgrid for 10 minutes, they can train off that data specifically in order to get me a better model. If I do that with 0. 3 versus 0. 1, the models come out different. The way that they interact is just much, much different. So I have to be really careful.
Paragraph 1861
I found that doing it with 0. 3 is actually better in some ways, unless I can do it with 0. 1 and change all of the different parameters, then that's more ideal, 'cause obviously, 0. 3 is faster than 0. 1. So I could get there. I can get there.
Paragraph 1862
Can you click using your brain?
Paragraph 1863
For right now, it's the hover clicking with the dwell cursor. Before all the thread retraction stuff happened, we were calibrating clicks - left click, right click. That was my previous ceiling. Before I broke the record again with the dwell cursor was I think on a 35 by 35 grid with left and right click. And you get more BPS, more bits per second using multiple clicks 'cause it's more difficult.
Paragraph 1864
Oh, because what is it, you're supposed to do either a left click or a like right click? You use a different color for stuff like this?
Paragraph 1865
Yeah, blue targets for left click; orange targets for right click is what they had done.
Paragraph 1866
Got it.
Paragraph 1867
So my previous record of 7.5 was with the blue and the orange targets, yeah, which I think if I went back to that now doing the click calibration, and being able to like initiate clicks on my own, I think I would would break that 10 ceiling like in a couple days max.
Paragraph 1868
Like yeah, you would start making Bliss nervous about his 17- - You should be.
Paragraph 1869
Why do you think we haven't given him the- - Yeah, exactly. So what did it feel like with the retractions? That some of the threads retracted?
Paragraph 1870
It sucked. It was really, really hard. The day they told me was the day of my big Neuralink tour at their Fremont facility, and they told me like right before we went over there. It was really hard to hear. My initial reaction was, "Alright, go in, fix it. Like go in, take it out, and fix it." The first surgery was so easy. Like I went to sleep. A couple hours later, I woke up and here we are.
Paragraph 1871
I didn't feel any pain, didn't take like any pain pills or anything. So I just knew that if they wanted to, they could go in and put in a new one like next day if that's what it took, 'cause I wanted it to be better and I wanted not to lose the capability. I had so much fun playing with it for a few weeks, for a month. It had opened up so many doors for me. It had opened up so many more possibilities that I didn't want to lose it after a month.
Paragraph 1872
I thought it would've been a cruel twist of fate if I had gotten to see the view from like the top of this mountain and then have it all come crashing down after a month. And I knew like, say, the top of the mountain, but how I saw it was I was just now starting to climb the mountain. There was so much more that I knew was possible. And so to have all of that be taken away was really, really hard.
Paragraph 1873
But then on the drive over to the facility, I don't know, like five minute drive, whatever it is, I talked with my parents about it. I prayed about it. I was just like, "I'm not gonna let this ruin my day. I'm not gonna let this ruin this amazing like tour that they have set up for me. Like I wanna go show everyone how much I appreciate all the work they're doing.
Paragraph 1874
I wanna go like meet all of the people who have made this possible, and I wanna go have one of the best days of my life." And I did, and it was amazing, and it absolutely was one of the best days I've ever been privileged to experience. And then for a few days, I was pretty down in the dumps. But for like the first few days afterwards, I was just like, I didn't know if it was gonna ever gonna work again.
Paragraph 1875
I made the decision that even if I lost the ability to use the Neuralink, even if I lost, even if I like lost out on everything to come, if I could keep giving them data in any way, then I would do that. If I needed to just do like some of the data collection every day or body mapping every day for a year, then I would do it, because I know that everything I'm doing helps everyone to come after me. And that's all I wanted.
Paragraph 1876
I guess the whole reason that I did this was to help people, and I knew that anything I could do to help, I would continue to do, even if I never got to use the cursor again, then I was just happy to be a part of it. And everything that I'd done was just a perk. It was something that I got to experience, and I know how amazing it's gonna be for everyone to come after me. So might as well just keep trucking along.
Paragraph 1877
Well, that said, you were able to get to work your way up, to get the performance back. So this is like going from rocky one to rocky two. So when did you first realize that this is possible, and what gave you sort of the strength and motivation, determination to do it? To increase back up and beat your previous record?
Paragraph 1878
Yeah, it was within a couple weeks.
Paragraph 1879
Again, this feels like I'm interviewing an athlete. (laughs) This is great. I like to thank my parents.
Paragraph 1880
The road back was long and hard, fraught many difficulties. There were dark days. It was a couple weeks, I think. And then there was just a turning point. I think they had switched how they were measuring the neuron spikes in my brain. Bliss, help me out.
Paragraph 1881
[Bliss] Yeah, the way in which we're measuring the behavior of individual neurons.
Paragraph 1882
Yeah.
Paragraph 1883
So we're switching from sort of individual spike detection to something called spike band power, which if you watch the previous segments with either me or DJ, you probably have some content.
Paragraph 1884
Yeah, okay, so when they did that, it was kind of like, a light over the head, like light bulb moment. Like, "Oh, this works." And this seems like we can run with this. And I saw the uptick in performance immediately. Like I could feel it when they switched over. I was like, "This is better. Like this is good."
Paragraph 1885
Like everything up till this point for the last few weeks, last like whatever, three or four weeks, 'cause it was before they even told me, like everything before this sucked. Like let's keep doing what we're doing now. And at that point, it was not like, "Oh, I know I'm still only at, like saying Webgrid terms, like four or five BPS compared to my 7. 5 before. But I know that if we keep doing this, then like I can get back there."
Paragraph 1886
And then they gave me the dwell cursor, and the dwell cursor sucked at first. It's not, obviously, not what I want, but it gave me a path forward to be able to continue using it, and hopefully, to continue to help out. And so I just ran with it, never looked back. Like I said, I'm just kind of person that roll with the punches anyways.
Paragraph 1887
What was the process? What was the feedback loop on the figuring out how to do the spike detection in a way that would actually work well for Noland?
Paragraph 1888
Yeah, it's a great question. So maybe just describe first how the actual update worked. It was basically an update to your implant. So we just did an over the air software update to his implants, same way you'd update your Tesla or your iPhone, and that firmware change enabled us to record sort of averages of populations of neurons nearby individual electrodes.
Paragraph 1889
So we have sort of less resolution about which individual neuron is doing what, but we have a broader picture of what's going on nearby an electrode overall. And that feedback, I mean, basically, Noland described it was immediate when we flipped that switch. I think the first day we did that, you had three or four BPS right out of the box. And that was a light bulb moment for, "Okay, this is the right path to go down."
Paragraph 1890
And from there, there's a lot of feedback around like how to make this useful for independent use. So what we care about ultimately is that you can use it independently to do whatever you want. And to get to that point, it required us to re-engineer the UX, as you talked about the dwell cursor, to make it something that you can use independently without us needing to be involved all the time.
Paragraph 1891
And yeah, this is obviously the start of this journey still. Hopefully, we get back to the places where you're doing multiple clicks and using that to control much more fluidly, everything, and much more naturally, the applications that you're trying to interface with.
Paragraph 1892
And most importantly, get that Webgrid number up.
Paragraph 1893
Yes.
Paragraph 1894
Yeah. So how is the, on the hover click, do you accidentally click cells sometimes? Like how hard is it to avoid accidentally clicking?
Paragraph 1895
I have to continuously keep it moving, basically. So like I said, there's a threshold where it will initiate a click. So if I ever drop below that, it'll start, and I have 0. 3 seconds to move it before it clicks anything. And if I don't want it to ever get there, I just keep it moving at a certain speed, and like just constantly like doing circles on screen, moving it back and forth to keep it from clicking stuff.
Paragraph 1896
I actually noticed a couple weeks back, when I was not using the implant, I was just moving my hand back and forth or in circles. Like I was trying to keep the cursor from clicking, and I was just doing it like while I was trying to go to sleep, and I was like, "Okay, this is a problem." (both laughing) - To avoid the clicking. I guess, does that create problems like when you're gaming accidentally click a thing?
Paragraph 1897
Yeah, yeah, it happens in chess. I've lost a number of games because I'll accidentally click something.
Paragraph 1898
[Bliss] I think the first time I ever beat you was because of an accident.
Paragraph 1899
Yeah, I misclicked, yeah.
Paragraph 1900
It's a nice excuse, right?
Paragraph 1901
Yeah.
Paragraph 1902
You can always, anytime you lose- - You could just say- - That was accidental.
Paragraph 1903
Yeah.
Paragraph 1904
You said the app improved a lot from version one. When you first started using it, it was very different. So can you just talk about the trial and error that you went through with the team? Like 200 plus pages of notes. What's that process like of- - Yeah.
Paragraph 1905
Going back and forth and working together to improve the thing?
Paragraph 1906
It's a lot of me just using it like day in and day out and saying, like, "Hey, can you guys do this for me? Like give me this. I wanna be able to do that. I need this." I think a lot of it just doesn't occur to them maybe until someone is actually using the app, using the implant. It's just something that they just never would've thought of. Or it's very specific to even like me, maybe what I want.
Paragraph 1907
It's something I'm a little worried about with the next people that come is maybe they will want things much different than how I've set it up, or what the advice I've given the team. And they're gonna look at some of the things they've added for me. Like that's a dumb idea. Like why would he ask for that?
Paragraph 1908
And so I'm really looking forward to get the next people on because I guarantee that they're going to think of things that I've never thought of. They're gonna think of improvements. I'm like, "Wow, that's a really good idea. Like I wish I would've thought of that." And then they're also gonna give me some pushback about like, "Yeah, what you are asking them to do here, that's a bad idea. Let's do it this way."
Paragraph 1909
And I'm more than happy to have that happen. But it's just a lot of like different interactions with different games or applications, the internet, just with the computer in general. There's tons of bugs that end up popping up left, right, center. So it's just me trying to use it as much as possible and showing them what works and what doesn't work and what I would like to be better.
Paragraph 1910
And then they take that feedback, and they usually create amazing things for me. They solve these problems in ways I would've never imagined. They're so good at everything they do. And so I'm just really thankful that I'm able to give them feedback and they can make something of it, 'cause a lot of my feedback is like really dumb. It's just like, "I want this. Please do something about it."
Paragraph 1911
And we'll come back and super well thought out, and it's way better than anything I could have ever thought of or implemented myself. So they're just great. They're really, really cool.
Paragraph 1912
As the BCI community grows, would you like to hang out with the other folks with Neuralink? What relationship, if any, would you wanna have with them? Because you said like they might have a different set of like ideas of how to use the thing.
Paragraph 1913
Yeah.
Paragraph 1914
Would you be intimidated by their Webgrid performance?
Paragraph 1915
No, no, I hope compete. I hope day one, they like wipe the floor with me. I hope they beat it and they crush it. Double it if they can. Just because, on one hand, it's only gonna push me to be better, 'cause I'm super competitive. I want other people to push me. I think that is important for anyone trying to achieve greatness is they need other people around them who are going to push them to be better. And I even made a joke about it on X once.
Paragraph 1916
Like once the next people get chosen, like qubadi cot music, like I'm just excited to have other people to do this with and to like share experiences with. I'm more than happy to interact with them as much as they want, more than happy to give them advice. I don't know what kind of advice I could give them, but if they have questions, I'm more than happy.
Paragraph 1917
What advice would you have for the next participant in the clinical trial?
Paragraph 1918
That they should have fun with this, because it is a lot of fun. And that I hope they work really, really hard, because it's not just for us, it's for everyone that comes after us. And come to me if they need anything. And to go to Neuralink if they need anything. Man, Neuralink moves mountains. Like they do absolutely anything for me that they can. And it's an amazing support system to have.
Paragraph 1919
It puts my mind at ease for like so many things that I have had like questions about, or so many things I wanna do. And they're always there, and that's really, really nice. I would tell them not to be afraid to go to Neuralink with any questions that they have, any concerns, anything that they're looking to do with this. And any help that Neuralink is capable of providing, I know they will. And I don't know, I don't know.
Paragraph 1920
Just work your ass off, because it's really important that we try to give our all to this.
Paragraph 1921
So have fun and work hard.
Paragraph 1922
Yeah, yeah, there we go. Maybe that's what I'll just start saying to people: have fun, work hard.
Paragraph 1923
Now, you're a real pro athlete. Just keep it short. (Noland laughing) Maybe it's good to talk about what you've been able to do now that you have a Neuralink implant. Like the freedom you gain from this way of interacting with the outside world. Like you play video games all night. And you do that by yourself. And that's a kind of freedom. Can you speak to that freedom that you gain?
Paragraph 1924
Yeah, it's what all, I don't know, people in my position want. They just want more independence. The more load that I can take away from people around me, the better. If I'm able to interact with the world without using my family, without going through any of my friends, like needing them to help me with things, the better.
Paragraph 1925
If I'm able to sit up on my computer all night and not need someone to like sit me up, say like on my iPad, like in a position where I can use it and then have to have them wait up for me all night until I'm ready to be done using it, it takes a load off of all of us. And it's really like all I can ask for. It's something that I could never thank Neuralink enough for. And I know my family feels the same way.
Paragraph 1926
Just being able to have the freedom to do things on my own at any hour of the day or night, it means the world to me. And I don't know.
Paragraph 1927
When you're up at 2:00 AM playing Webgrid by yourself, I just imagine like it's darkness and there's just a light glowing. And you're just focused. What's going through your mind? (Noland laughing) Or you were like in a state of flow where it's like the mind is empty, like those like Zen masters?
Paragraph 1928
Yeah, generally, it is me playing music of some sort. I have a massive playlist, and so I'm just like rocking out to music. And then it's also just like a race against time, 'cause I'm constantly looking at how much battery percentage I have left on my implant. Like, all right, I have 30%, which equates to x amount of time, which means I have to break this record in the next hour and a half, or else, it's not happening tonight.
Paragraph 1929
And so it's a little stressful when that happens. When it's above 50%, I'm like, "Okay, like I got time." It starts getting down to 30 and then 20. It's like, all right, 10%, a little popup is gonna pop up right here, and it's gonna really screw my Webgrid flow. It's gonna tell me that there's like the low battery, low battery popup comes up, and I'm like, it's really gonna screw me over.
Paragraph 1930
So if I have to, if I'm gonna break this record, I have to do it in the next like 30 seconds, or else, that popup is gonna get in the way, like cover my Webgrid. After that, I go click on it, go back into Webgrid. And I'm like, "All right, that means I have 10 minutes left before this thing's dead." That's what's going on in my head generally, that and whatever song's playing. I want to break those records so bad.
Paragraph 1931
Like it's all I want when I'm playing Webgrid. It has become less of like, "Oh, this is just a leisurely activity." Like I just enjoy doing this, because it just feels so nice and it puts me at ease. No, once I'm in Webgrid, you better break this record or you're gonna waste like five hours of your life right now. And I don't know, it's just fun. It's fun, man.
Paragraph 1932
Have you ever tried Webgrid with like two targets and three targets? Can you get higher BPS with that?
Paragraph 1933
Can you do that?
Paragraph 1934
[Bliss] You mean, like different color targets, or you mean- - Oh, multiple targets, 'cause that change the thing.
Paragraph 1935
Yeah, so BPS is a log of number of targets times correct minus incorrect divided by time. And so you can think of like different clicks as basically doubling the number of active targets.
Paragraph 1936
Got it.
Paragraph 1937
So you know, you get basically higher BPS, the more options there are, the more difficult to task. And there's also like zen mode you've played in before, which is like- - Yeah. Yeah, it covers the whole screen with a grid. And I don't know.
Paragraph 1938
[Lex] Yeah, and so you can go like, that's insane.
Paragraph 1939
Yeah.
Paragraph 1940
[Bliss] He doesn't like it 'cause it didn't show BPS.
Paragraph 1941
I had them put in a giant BPS in the background, so now it's like the opposite of Zen mode. It's like super hard mode, like just metal mode. It's just like a giant number in the back counter.
Paragraph 1942
[Bliss] We should rename that. Metal mode is a much better name now.
Paragraph 1943
So you also play Civilization VI?
Paragraph 1944
I love Civ VI, yeah.
Paragraph 1945
[Lex] You usually go with Korea, you said?
Paragraph 1946
I do, yeah. So the great part about Korea is they focus on like science tech victories, which was not planned. Like I've been playing Korea for years, and then all of the Neuralink stuff happened. So it kind of aligned. But what I've noticed with tech victories is if you can just rush tech, rush science, then you can do anything.
Paragraph 1947
Like at one point in the game, you'll be so far ahead of everyone technologically that you'll have like musket men, infantry men, plane sometimes, and people will still be fighting with like bows and arrows. And so if you want to win a domination victory, you just get to a certain point with the science and then go and wipe out the rest of the world.
Paragraph 1948
Or you can just take science all the way and win that way, and you're gonna be so far ahead of everyone 'cause you're producing so much science that it's not even close. I've accidentally won in different ways just by focusing on science.
Paragraph 1949
Accidentally won by focusing on science.
Paragraph 1950
I was playing only science, obviously. Like just science all the way, just tech. And I was trying to get like every tech in the tech tree and stuff. And then I accidentally won through a diplomatic victory, and I was so mad. (Lex laughing) I was so mad, 'cause it just like ends the game. One turn, it was like, "Oh, you won, you're so diplomatic." I'm like, "I don't wanna do this. I should have declared war on more people or something."
Paragraph 1951
It was terrible, but you don't need like giant civilizations with tech, especially with Korea. You can keep it pretty small. So I generally just get to a certain military unit and put them all around my border to keep everyone out, and then I will just build up. So very isolationist.
Paragraph 1952
Nice.
Paragraph 1953
Yeah.
Paragraph 1954
Just work on the science and the tech.
Paragraph 1955
[Noland] Yep, that's it.
Paragraph 1956
You're making it sound so fun.
Paragraph 1957
It's so much fun.
Paragraph 1958
And I also saw Civilization VII trailer.
Paragraph 1959
Oh man, I'm so pumped.
Paragraph 1960
Yeah. And that's probably coming out- - Come on, Civ VII, hit me up. Alpha, beta tests, whatever.
Paragraph 1961
Wait, when is it coming out?
Paragraph 1962
2025.
Paragraph 1963
Yeah, yeah, next year, yeah. What other stuff would you like to see improved about the Neuralink app and just the entire experience?
Paragraph 1964
I would like to, like I said, get back to the like click on demand, like the regular clicks. That would be great. I would like to be able to connect to more devices. Right now, it's just the computer. I'd like to be able to use it on my phone or use it on different consoles, different platforms. I'd like to be able to control as much stuff as possible, honestly. Like an Optimus robot would be pretty cool.
Paragraph 1965
That would be sick if I could control an Optimus robot. The Link app itself, it seems like we are getting pretty dialed in to what it might look like down the road. Seems like we've gotten through a lot of what I want from it at least. The only other thing I would say is like more control over all the parameters that I can tweak with my like cursor and stuff. There's a lot of things that go into how the cursor moves in certain ways.
Paragraph 1966
And I have, I don't know, like three or four of those parameters and there might- - Like gain and friction and all that?
Paragraph 1967
Gain and friction, yeah. And there's maybe double the amount of those with just like velocity and then with the actual dwell cursor. So I would like all of it. I want as much control over my environment as possible, especially- - So you want like advanced mode? Like there's menus usually, there's basic mode. And you're like one of those folks like- - I go the- - Power user advanced.
Paragraph 1968
Yeah, yeah.
Paragraph 1969
Got it.
Paragraph 1970
That's what I want. I want as much control over this as possible. So yeah, that's really all I can ask for. Just give me everything.
Paragraph 1971
Has speech been useful? Like just being able to talk also in addition to everything else?
Paragraph 1972
Yeah, you mean like while I'm using it?
Paragraph 1973
While you're using it, like speech to text?
Paragraph 1974
Oh yeah.
Paragraph 1975
Or do you type, or like, 'cause there's also a keyboard? That's really nice.
Paragraph 1976
Yeah, yeah. So there's a virtual keyboard. That's another thing I would like to work more on is finding some way to type or text in a different way. Right now, it is like a dictation basically and a virtual keyboard that I can use with the cursor. But we've played around with like finger spelling, like sign language, finger spelling. And that seems really promising.
Paragraph 1977
So I have this thought in my head that it's going to be a very similar learning curve that I had with the cursor, where I went from attempted movement to imagined movement at one point. I have a feeling, this is just my intuition, that at some point, I'm going to be doing finger spelling and I won't need to actually attempt to finger spell anymore, that I'll just be able to think the like letter that I want and it'll pop up.
Paragraph 1978
That would be epic.
Paragraph 1979
Yeah.
Paragraph 1980
That's challenging, that's hard. That's a lot of work for you to kinda take that leap. But that would be awesome.
Paragraph 1981
And then like going from letters to words is another step. Like you would go from, right now, it's finger spelling of like just the sign language alphabet. But if it's able to pick that up, then it should be able to pick up like the whole sign language like language.
Paragraph 1982
And so then if I could do something along those lines, or just the sign language spelled word, if I can spell it at a reasonable speed and it can pick that up, then I would just be able to think that through and it would do the same thing. I don't see why not. After what I saw with the cursor control, I don't see why it wouldn't work, but we'd have to play around with it more.
Paragraph 1983
What was the process in terms of like training yourself to go from attempted movement to imagined movement? How long did that take? So how long would this kind of process take?
Paragraph 1984
Well, it was a couple weeks before it just like happened upon me. But now that I know that that was possible, I think I could make it happen with other things. I think it would be much, much simpler.
Paragraph 1985
Would you get an upgraded implant device?
Paragraph 1986
Sure, absolutely. Whenever they'll let me.
Paragraph 1987
So you don't have any concerns for you with the surgery experience? All of it was like no regrets?
Paragraph 1988
No.
Paragraph 1989
So everything's been good so far?
Paragraph 1990
Yep.
Paragraph 1991
You just keep getting upgrades.
Paragraph 1992
Yeah, I mean, why not? I've seen how much it's impacted my life already. And I know that everything from here on out, shit's gonna get better and better. So I would love to. I would love to get the upgrade.
Paragraph 1993
What future capabilities are you excited about sort of beyond this kind of telepathy? Is vision interesting? So for folks who, for example, who are blind, so you're like enabling people to see, or for speech.
Paragraph 1994
Yeah, there's a lot that's very, very cool about this. I mean, we're talking about the brain, so like this is just motor cortex stuff. There's so much more that can be done. The vision one is fascinating to me. I think that is going to be very, very cool. To give someone the ability to see for the first time in their life would just be, I mean it, it might be more amazing than even helping someone like me. Like that just sounds incredible.
Paragraph 1995
The speech thing is really interesting, being able to have some sort of like real time translation and cut away that language barrier would be really cool. Any sort of like actual impairments that it could solve, like with speech, would be very, very cool. And then also, there are a lot of different disabilities that all originate in the brain. And you would be able to, hopefully be able to solve a lot of those.
Paragraph 1996
I know there's already stuff to help people with seizures that can be implanted in the brain. This would do, I imagine, the same thing. And so you could do something like that. I know that even someone like Joe Rogan has talked about the possibilities with being able to stimulate the brain in different ways. I'm not sure. I'm not sure how ethical a lot of that would be. That's beyond me honestly.
Paragraph 1997
But I know that there's a lot that can be done when we're talking about the brain and being able to go in and physically make changes to help people or to improve their lives. So I'm really looking forward to everything that comes from this. And I don't think it's all that far off. I think a lot of this can be implemented within my lifetime, assuming that I live a long life.
Paragraph 1998
What you were referring to is things like people suffering from depression or things of that nature potentially getting help.
Paragraph 1999
Yeah, flip a switch like that, make someone happy. I know, I think Joe has talked about it more in terms of like you want to experience like what a drug trip feels like. Like you wanna experience what it'd be like to be on- - Of course.
Paragraph 2000
Yeah, mushrooms or something like that, DMT. Like you can just flip that switch in the brain. My buddy Bain has talked about being able to like wipe parts of your memory and re-experience things that, like for the first time, like your favorite movie or your favorite book. Like just wipe that out real quick, and then re-fall in love with Harry Potter or something. I told him, I was like, "I don't know how I feel about like people being able to just wipe parts of your memory. That seems a little sketchy to me." He's like, "They're already doing it."
Paragraph 2001
Sounds legit. Yeah, I would love memory replay. Just like actually high resolution replay of old memories.
Paragraph 2002
Yeah, I saw an episode of "Black Mirror" about that once. I don't think I want it.
Paragraph 2003
Yeah, so "Black Mirror" always kind of considers the worst case, which is important. I think people don't consider the best case or the average case enough. I don't know what it is about us humans. We wanna think about the worst possible thing. We love drama.
Paragraph 2004
[Noland] Yeah. (laughs) - It's like, how's this new technology gonna kill everybody? We just love that. Again like, yes, let's watch.
Paragraph 2005
Hopefully, people don't think about that too much with me. It'll ruin a lot of my plans.
Paragraph 2006
Yeah, yeah. I assume you're gonna have to take over the world. I mean, I loved your Twitter. You tweeted, "I'd like to make jokes about hearing voices in my head since getting the Neuralink, but I feel like people would take it the wrong way. Plus, the voices in my head told me not to."
Paragraph 2007
Yeah.
Paragraph 2008
Please never stop. So you're talking about Optimus. Is that something you would love to be able to do, to control the robotic arm or the entirety of Optimus?
Paragraph 2009
Oh yeah, for sure. For sure, absolutely.
Paragraph 2010
You think there's something like fundamentally different about just being able to physically interact with the world?
Paragraph 2011
Yeah, oh, 100%. I know another thing with like being able to like give people the ability to like feel sensation and stuff too by going in with the brain and having the Neuralink maybe do that. That could be something that could be translated through, transferred through the Optimus as well. Like there's all sorts of really cool interplay between that. And then also, like you said, just physically interacting.
Paragraph 2012
I mean, 99% of the things that I can't do myself obviously need, I need a caretaker for, someone to physically do things for me. If an Optimus robot could do that, like I could live an incredibly independent life and not be such a burden on those around me and it would change the way people like me live, at least until whatever this is gets cured. But being able to interact with the world physically, like that would just be amazing.
Paragraph 2013
And they're not just like for being, for having to be a caretaker or something, but something like I talked about, just being able to read a book. Imagine an Optimus robot just being able to hold a book open in front of me, like get that smell again. I might not be able to feel it at that point. Or maybe I could again with the sensation and stuff.
Paragraph 2014
But there's something different about reading like a physical book than staring at a screen or listening to an audio book. I actually don't like audio books. I've listened to a ton of them at this point, but I don't really like 'em. I would much rather like read a physical copy.
Paragraph 2015
So one of the things you would love to be able to experience is opening the book, bringing it up to you. And to feel the touch of the paper.
Paragraph 2016
Yeah. Oh man, the touch, the smell. I mean, it's just like something about the words on the page, and they've replicated that page color on like the Kindle and stuff. Yeah, it's just not the same, yeah. So just something as simple as that.
Paragraph 2017
So one of the things you miss is touch.
Paragraph 2018
I do, yeah.
Paragraph 2019
A lot of things that I interact with in the world, like clothes or literally any physical thing that I interact with in the world, a lot of times, what people around me will do is they'll just come like, rub it on my face. They'll like lay something on me so I can feel the weight. They will rub a shirt on me so I can feel fabric. Like there's something very profound about touch, and it's something that I miss a lot, and something I would love to do again, but we'll see.
Paragraph 2020
What would be the first thing you do with a hand that can touch? Give your mom a hug after that, right?
Paragraph 2021
Yeah, I know. It's one thing that I've asked like God for basically every day since my accident was just being able to like one day move, even if it was only like my hand. So that way, like I could squeeze my mom's hand or something just to like show her that, like how much I care and how much I love her and everything. Something along those lines. Being able to just interact with the people around me, handshake, give someone a hug. I don't know, anything like that. Being able to help me eat, like I'd probably get really fat, which would be a terrible, terrible thing.
Paragraph 2022
Also beat Bliss in chess on a physical chess board.
Paragraph 2023
Yeah, yeah. I mean, there are just so many upsides. (laughs) And any way to find some way to feel like I'm bringing Bliss down to my level.
Paragraph 2024
Yeah.
Paragraph 2025
Because- - Yeah.
Paragraph 2026
He's just such an amazing guy, and everything about him is just so above and beyond that anything I can do to take him down a notch, I'm more than happy.
Paragraph 2027
Yeah, humble him a bit, he needs it.
Paragraph 2028
[Noland] Yeah. (laughs) - Okay. As he's sitting next to me. Did you ever make sense of why God puts good people through such hardship?
Paragraph 2029
Oh, man. I think it's all about understanding how much we need God. And I don't think that there's any light without the dark. I think that if all of us were happy all the time, there would be no reason to turn to God ever. I feel like there would be no concept of good or bad. And I think that as much of like the darkness and the evil that's in the world, it makes us all appreciate the good and the things we have so much more.
Paragraph 2030
And I think, like when I had my accident, one of the first things I said to one of my best friends was, and this was within like the first month or two after my accident, I said, "Everything about this accident has just made me understand and believe that like God is real and that there really is a God, basically. And that like my interactions with him have all been real and worthwhile."
Paragraph 2031
And he said, if anything, seeing me go through this accident, he believes that there isn't a God. And it's a very different reaction. But I believe that it is a way for God to test us, to build our character, to send us through trials and tribulations, to make sure that we understand how precious he is, and the things that he's given us and the time that he's given us. And then to hopefully grow from all of that.
Paragraph 2032
I think that's a huge part of being here is to not just have an easy life and do everything that's easy, but to step out of our comfort zones and really challenge ourselves, because I think that's how we grow.
Paragraph 2033
What gives you hope about this whole thing we have going on, human civilization?
Paragraph 2034
Oh, man. I think people are my biggest inspiration. Even just being at Neuralink for a few months, looking people in the eyes and hearing their motivations for why they're doing this, it's so inspiring. And I know that they could be other places, cushier jobs, working somewhere else, doing X, Y, or Z that doesn't really mean that much.
Paragraph 2035
But instead they're here, and they want to better humanity and they want a better, just the people around them, the people that they've interacted with in their life, they wanna make better lives for their own family members who might have disabilities, or they look at someone like me and they say, "I can do something about that so I'm going to." And it's always been what I've connected with most in the world are people.
Paragraph 2036
I've always been a people person and I love learning about people, and I love learning like how people developed and where they came from. And to see like how much people are willing to do for someone like me when they don't have to, and they're going out of their way to make my life better.
Paragraph 2037
It gives me a lot of hope for just humanity in general, how much we care and how much we're capable of when we all kind of get together and try to make a difference. And I know there's a lot of bad out there in the world, but there always has been and there always will be.
Paragraph 2038
And I think that that is, it shows human resiliency and it shows what we're able to endure, and how much we just want to be there and help each other, and how much satisfaction we get from that, because I think that's one of the reasons that we're here is just to help each other. And I don't know, that always gives me hope. It's just realizing that there are people out there who still care and who wanna help.
Paragraph 2039
And thank you for being one such human being and continuing to be a great human being through everything you've been through and being an inspiration to many people, to myself, for many reasons, including your epic, unbelievably great performance on Webgrid. I'll be training all night tonight to try to catch up.
Paragraph 2040
You can do it.
Paragraph 2041
And I believe in you, that you can once you come back, so sorry to interrupt with the Austin trip, once you come back, eventually beat Bliss.
Paragraph 2042
Yeah, yeah, for sure. Absolutely.
Paragraph 2043
I'm rooting for you. The whole world is rooting for you.
Paragraph 2044
Thank you.
Paragraph 2045
Thank you for everything you've done, man.
Paragraph 2046
Thanks, thanks man.
Paragraph 2047
Thanks for listening to this conversation with Nolan Arbaugh, and before that, with Elon Musk, DJ Seo, Matthew MacDougall, and Bliss Chapman. To support this podcast, please check out our sponsors in the description. And now, let me leave you with some words from Aldous Huxley in "The Doors of Perception." "We live together. We act on and react to one another, but always, and in all circumstances, we are by ourselves.
Paragraph 2048
The martyrs go hand in hand into the arena. They are crucified alone. Embraced, the lovers desperately try to fuse their insulated ecstasies into a single self-transcendence; in vain. By its very nature, every embodied spirit is doomed to suffer and enjoy in solitude. Sensations, feelings, insights, fancies, all these are private and, except through symbols and at second hand, incommunicable.
Paragraph 2049
We can pool information about experiences, but never the experiences themselves. From family to nation, every human group is a society of island universes. Thank you for listening, and hope to see you next time.