埃隆，你好，欢迎，你是在帕洛阿尔托吧，据我了解，是的，首先，嗯，一个位于帕洛阿尔托的全球工程总部，很好，嗯，非常感谢你加入我们，呃，实际上我要从一个和预想稍微不同的地方开始，呃，因为我刚刚在Twitter上，在Twitter上看到，呃，一则小公告，呃，我很想听听你的，你的想法，我看到你，呃，明天早上要采访朗达·桑蒂斯，对吗，在Twitter Spaces上，我看到，是的，时间，呃，我需要确切的时间，但，哦，是按你的时间明天早上，我想可能没错，好，呃，那么，呃，是的，我，呃，将采访，呃，罗恩·桑托斯，而他有一项相当重大的消息要宣布，呃，我们将会，我们是第一次有这样的事情在社交媒体上发生，而且

采用实时问答，呃，不是照稿的，呃，所以这将是直播，让、让、让我们看看会发生什么。外国的。而且过去几天你一直在发一些关于蒂姆·斯科特的推文，嗯，关于你支持谁，我们应该怎么想？显然这次采访向我们透露了一些信息，你能让我们了解一下你目前是怎么想的吗？是的，我的意思是，我目前，嗯，并不打算支持任何特定的癌症，嗯，但我确实有兴趣让，呃，你知道，X／Twitter在某种程度上成为一个公共市民广场，并且，呃，让越来越多的组织在Twitter上托管内容并发布公告，嗯，它是、它是互联网上唯一真正能获取，呃，实时新闻的地方，精确到分钟和秒，而且它，呃，是的，所以我认为它

这相当具有开创性，竟然会有，嗯，这类重大公告在社交媒体上发布，嗯，而我们是否应该期待，嗯，抱歉，我不想在这件事上说太久，但是，嗯，在你作为采访者而不是受访者的新角色中，呃，我们是否应该期待，嗯，呃，我们是否应该期待更多这样的事情？我的意思是，如果这是市民广场，你会采访其他候选人、民主党人吗？你对此有什么、有什么想法？人们愿意来吗？你会在那里让市民广场覆盖整个政治光谱吗？是的，当然，嗯，所以正如我、正如我承诺的那样，当我进行一系列媒体采访时，我接受了各种采访，嗯，而且，呃，我想这也算是一次媒体采访，嗯，范围从某种，你知道，左翼、温和派到被认为是

右翼，嗯，而且我确实认为，重要的是Twitter要，呃，同时具备一个公平竞争环境的现实和行列，呃，成为一个，呃，所有声音都能被听到的地方，并且具备，呃，那种你确实在其他任何地方都看不到的动态互动，嗯，比如我今天会在Twitter上见到你，嗯，呃，AOC陷入了一场，而特德·克鲁兹在Twitter上发生了争论，这，嗯，是的，不管你赞同哪一方，嗯，它仍然非常有娱乐性。那么，嗯，我确信明天也会很有娱乐性，我们都会关注，不过，嗯，当你、当你着手进行那样的采访，而且显然是像即将到来的这场非常重要的选举时，你能、你能稍微谈谈，真正关键的问题是什么吗，这些问题

在这个关键时刻对你真正重要？你可以说对作为个人的我重要，或者对作为个人的你重要，嗯，就谁来领导这个国家而言，但也可以，你知道，比这更广泛一些，你知道，对这个国家以及对、对你的企业而言。我的意思是，你能谈谈你认为这里真正的问题所在吗？嗯，我曾公开说过，嗯，我的偏好，而且我认为也是大多数美国人的偏好，实际上是让一个相当正常的人任职。我想我们大家实际上都会对此相当，呃，相当满意，嗯，你知道，我认为是一个，呃，代表温和观点的人，而我认为实际上这个国家的大多数人都持有这种观点，嗯，而且，嗯，但、但是现在的设置方式是，我们、我们的确有一个似乎会把事情推向

边缘的制度，因为有初选。所以为了赢得初选，你显然必须赢得你所在政党的多数选票，在两种情况下，这往往会导致向左和向右摆动，尽管我认为在初选期间，事情比简单的左和右更复杂，然后，呃，然后让我们在《华尔街日报》选举中转向中间，嗯。至于，呃，我、我认为的是，是的，所以我真的只是希望有一个，你知道，相当正常、容易接近的人担任总统，那就太好了。那么，如果我们逐一说说，如果我们逐一说说目前框架中的4个名字，你能不能简单地用“是”或“否”回答他们是否正常、是否明智，嗯，那么，嗯，我们有乔·拜登。我、我认为，你知道，对这些说法要谨慎，嗯，所以我，呃，我们也许

得先喝几杯，我才会把答案告诉你。我会，嗯，我期待那一天，也期待，呃，明天的对话，而且显然未来几个月还会有更多这样的对话，所以，嗯，那很好，非常感谢，嗯。那么我想先谈的是，你刚刚飞抵，我想是在过去20分钟内，嗯，你过着相当忙乱的生活，嗯，但你说过唯一真正的货币是时间。你能不能让这个房间里那些日程被安排得密不透风的人了解一下，大致上你、你，埃隆·马斯克一天的生活是什么样的？那看起来是什么样？嗯，婚礼非常漫长而复杂，正如你可能想象的那样，是的，嗯，而且，嗯，而且存在大量、存在大量的上下文切换

所以，嗯，并不是说我的生活叫作类似《毁灭战士》的东西，其中就像，恐惧不是思想杀手，上下文切换才是，呃，所以切换上下文是、是相当痛苦的，但是，嗯，我通常确实会尽量按公司划分，主要是一天只处理一家公司。所以今天是Tesla日，例如，嗯，尽管我今晚晚些时候最后可能会去Twitter，嗯，然后明天也会有一部分是Tesla日，还有Twitter，而且，嗯，然后星期四会算是半天SpaceX、半天Tesla，但这些事情在某种程度上相互交织，所以时间管理极其困难，嗯，而且这听起来会很奇怪，但我，嗯，我只有1位，呃，兼职助理，嗯。那每周工作多少天，是兼职的？我的意思是，我想、我想她是怎么工作的，谁是

严格来说是全职的，但就其本身而言，我、我大部分日程安排，呃，都是自己做，嗯，而我这么做的原因是，别人不可能知道优先事项是什么，嗯，所以，而且既然我拥有的最宝贵的东西是时间，我大部分时候就自己安排。那么，如果你今天来到Tesla，你是已经安排好了一系列会议，还是，嗯，你进来时，你知道，脑子里想着某件事，然后进去见人？我的意思是，这有多么结构化？或者如果你在Twitter购物，就为你工作的人而言，他们怎么、他们怎么处理这件事？嗯，是的，所以今天我在Twitter有几个小时已经安排好的会议，嗯，所以有许多、有许多事情是按每周节奏运作的，嗯，因此那些会议已经安排好了，而且

然后我在一天结束时还有补充会议，呃，尤其是我，我，我，我大概要到凌晨2点才会睡觉。

M或者类似的东西，而且几乎全程都在工作。而如果你亲自摇晃这个，未来几年AI会有助于帮助你，嗯，做这件事吗？你会使用技术来帮助你管理它吗？我想我们所有人都会使用技术。我本人日常并不大量使用AI。我的意思是，Tesla AI实际上在现实世界AI方面非常先进，它是最先进的远程世界AI，嗯，遥遥领先，嗯。事实上，如果职位互换，让Microsoft和OpenAI负责创造，呃，谁能创造出最好的大语言模型，基本上，如果测试互换，让Tesla承担制造最具竞争力的大语言模型的任务，而Microsoft、OpenAI承担自动驾驶任务，Tesla会赢。好的

我认为人们不了解描绘Tesla AI系统能力的程度。所以，尽管我个人不怎么使用AI，Tesla使用一个变压器支架。我们等一下会谈到这个，如果可以的话。但就管理你的人生所做的事情而言，还有最后一件事：你经营着3家非常大的公司，你在其中至少2家拥有非常大的权益以及，你知道，所有权控制，嗯。如果你突然打电话并执行你正在做的事，你的继任计划是什么？既包括谁来管理这些公司，也同样重要地包括谁来对那些股份行使投票权，就，嗯，你知道，从更长期和战略层面会发生什么而言，你对所有这些事情有什么计划？是的，嗯，继任是最棘手的古老问题之一，嗯，你知道，它、它

它是、它困扰着，嗯，你知道，国家、国王、总理和总统，以及CEO，你知道，从历史之初便是如此，嗯。没有显而易见的解决方案。我、我的意思是，已经确定了一些特定的人选，我已经告诉人们，说听着，如果我意外出了什么事，这就是我的、这就是我建议的接任者。所以在所有情况下，董事会都知道，我推荐的人选是谁，他们可以选择采纳，这当然取决于他们，他们、他们也可能选择不同的方向，但我，他们，在最坏的情况下，有一个人，这就是应该管理公司的人，呃。控制权问题则要复杂得多，它是一个要棘手得多的问题，嗯，也是我正在努力解决的问题，而且坦率地说，我愿意听取意见，因为确实

我已经创办和正在创办的公司，嗯，合在一起拥有，呃，巨大的能力，因此对它们的管理极其重要，嗯。我想确保这种管理最终，呃，积累为对人类的益处，这个理念是文明的宽容，嗯。但并不是说我们总能在这方面成功，而是说这在愿景上是我们的目标，嗯。所以，嗯，我、我有一个、一个想法，它算是已经部分落实了，就是创建某种，一个、一个算是教育机构的东西，由它控制，嗯，我的大部分投票权，嗯。但这、这不是，嗯，自动地，我、我绝对不属于自动给我的孩子们，你知道，一些公司股份的那一派，即使他们并不，即使他们

没有兴趣或意愿，你知道，也没有能力去管理这些公司，我认为那是个错误，嗯，所以，但但这是一个非常难解决的问题，对吧，所以，然后谁应该进入教育机构的董事会，也是一个非常非常难解决的工作，嗯，所以我认为，或许在某种程度上分散控制权是合理的，我其实只是在这里创造性地边想边说，但听起来这是你需要，我是说，你需要开始规划那些人会是谁，因为正如你所说，无论是看 SpaceX、看 Tesla，还是看 Twitter，这些这些对社会都非常重要，而由合适的人来就它们未来走向何方以及资金花在哪里进行那些投票，是相当

重要的。是的，绝对如此。现在，这些公司的目标，它们实现这些目标的难度差异相当大，嗯，你知道，Tesla 最初的目标是加速可持续能源时代的到来，事实上我认为我们已经在很大程度上做到了这一点，嗯，而且实际上已经，这有点是，呃，这有点，这有点是一个订单行业的 CEO 过于频繁地承认 Tesla 在加速电动汽车发展方面的作用，嗯，而且，嗯，是的，所以所以我觉得这件事已经形成了很大的势头，嗯，他们仍在解决自动驾驶问题，而我们，你知道，满怀希望地期望今年做到，呃，而且，嗯，所以它还有很长的路要走，但但执行计划相对清晰，嗯，而且，呃，接下来，这项执行计划将产生大量正向的

公司现金流，所以这就像是，如果说这是件相当显而易见该做的事，那么 SpaceX 的问题更难，因为，呃，长期目标是让生命成为多行星物种，在火星上建立一座自给自足的城市，而这起初很可能会非常随意或者是负面的，从长期来看，我们就说火星上的目标市场很小。你必须从长远考虑。是的，我想，你还必须和与你一起去的那些人相处得非常好。呃，是的，当然，嗯，你知道，理智将是前往火星的首要要求，呃，还有稳定性，你不会希望有人发疯，在半夜打开气闸舱门，嗯，对，所以 SpaceX 是一个更难的问题，因为它是，它是，呃，一个更加长期得多的目标，而且，和，呃，更多的钱、更少的

在这个过程中，所以必须确保那件事发生，嗯，而那是否是这样一种事情，继续谈 SpaceX，我们马上会转到 Twitter 并请你告诉我们，但，嗯，那是否是你希望锁定在 SpaceX 目标中的事情，即无论发生什么，火星始终是这一项目的终极，嗯，抱负？那对你有那么重要吗？是的，我是说，是让光成为多行星的，以便，嗯，而多行星化的关键门槛是，如果来自地球的补给船因任何原因停止抵达，火星不会灭绝，那是，那是关键的大过滤器。如果用费米悖论来谈论这些事情，大过滤器就是，嗯，火星能够在没有任何来自地球的补给展示的情况下自给自足。在我们达到那一点之前，我们

实际上仍然只是一个单行星文明，只是，呃，有一个延伸，但当各颗行星都能自给自足、阿曼能自给自足时，那么即便在最坏的情形下，地球文明要么轰然消亡，要么悄然消亡，嗯，那么火星存续下去的机会就会大得多，嗯，所以这里的总体意图是确保意识之光延续，而它看起来只是一片广袤黑暗中的一根微小蜡烛，嗯，我经常，我经常被问到，我是否见过外星人存在的任何证据，而我，我，我没有，嗯，你知道，除了我在申请我的，呃，绿卡时，曾一度没有外侨登记卡这一事实，呃，那是一种外侨登记，嗯，确实可能是稍微不同类型的外星人，但，嗯，所以你认为你会，将会

你能活着看到火星这件事发生吗？我，我希望能活着看到第一批人类登上火星，嗯，但我认为在那之后还需要一段时间，才能让我实现自给自足，所以据我猜测，从第一次造访到让我实现自给自足至少需要 20 年，也可能是 40 或 50 年。

而且那还是假设你真的全力以赴，对吧，所以这是一件难事，嗯，但怎么说呢，所以我认为这对于提高文明的生存能力很重要。那谁来为此买单？我是说，你的投资者会拿出资金来做这件事吗？你会期望政府资助它吗？这些钱从哪里来？因为正如你所说，你可以从星链获得回报，可以从为其他人发射卫星和太空旅游中获得回报，但我是说，那是，那是更难获得的回报，不是吗？是的，我认为，我认为从长期来看，呃，它的价值会高得不可思议。我只是会说，它只是超出了大多数人或大多数投资者的规划周期。所以，嗯，我是说，显然，如果，嗯，火星上有一座繁荣的，嗯，城市，而且有大量行星际

贸易，而 SpaceX 是其主要服务提供商，它将极其有价值，嗯，所以，嗯，但但你知道，有一点是，要成为这个自给自足的，你知道，殖民地，嗯，而且，嗯，我认为我们，我们，我认为我们，呃，通常过度假设文明是稳固的，嗯，没有什么真能让它垮掉，呃，这种情绪在历史上各个帝国宣告结束前不久都很常见。所以，而且你知道，我不得不说，你知道，现在确实有一点帝国晚期的氛围，嗯，呃，是的，当然。你是否，嗯，是的，嗯，呃，在你看来，AI 是否会加速这种风险，或者增加出现那种结果的风险？我认为会，是的，嗯，我是说，我们肯定能建造一座

自给自足的火星城市，而不需要，而不需要 AI，也不需要所谓的 AGI，后者通常指通用人工智能或超级智能。我有，所以我认为那，呃，对我们正在做的任何事情都不是必需的，嗯，但它正在发生，而且发生得非常快，所以，呃，先进 AI 确实存在一种风险，嗯，它可能会消灭人类或限制人文领域的发展。我更多想到的是相反的情况：它是否会增加这个星、这个星球自行输入内爆、进而让那些事情成真的概率？你，我是说，你目前有多担心这些发展会加速你的，你的糟糕情境？嗯，呃，我是说，人工数字式超级智能的发展，呃，在很大程度上是一把双刃剑，所以它是，如果你有，如果你有

一个能满足你任何愿望的精灵，他们也能做任何事情，嗯，那必然会带来危险，而且我预计 AI 最早的用途将是，或者至少政府最早对 AI 的用途将是武器技术。所以只要在战场上拥有，嗯，更先进的武器，如果你回头看，它们的反应速度可以，呃，快过任何人类，那，那正是 AI 将能做到的事情。我是说，先进国家之间未来的战争，或者至少拥有强大无人机能力的国家之间的未来战争，将在很大程度上是无人机战争。所以我想回到 AI，因为这是，这是大事，我也想更详细地讨论，但我确实想从遥远的未来回到，就回到当下，嗯，你刚刚聘请了一位新 CEO，呃，琳达·亚卡里诺，一位广告业老将

嗯，进入 Twitter。你通常专注于招聘工程师，你知道，琳达是一个非常不同的人。你能不能快速讲讲你们接洽的经过，那是怎么发生的？嗯，我们在数月间进行了交谈，呃，只是与广告有关，嗯，然后，呃，琳达觉得，嗯，让广告商亲眼见到我会非常有帮助，于是邀请我去迈阿密参加一场会议，嗯，这非常有帮助，而且我亲自会见了许多广告商，向他们保证，你知道，向他们保证 Twitter 是一个适合投放广告的地方，而且总的来说，事实上，仇恨言论已经减少，确实如此，而且系统的质量，特别是在诈骗犯和垃圾信息发送者方面，比过去显著提高了，嗯，我们已经

清除了，到目前为止，远超 90% 的那些诈骗和诈骗，那些诈骗以及 Twitter 上的垃圾信息。现在应该相当少见了，在 UCS 营地，嗯，所以，嗯，我们还推出了，呃，某种，所以说清楚一点，当你说你已经清除了 90 的诈骗时，这和机器人账号是同一回事吗？还是说这是泛指诈骗，而机器人账号在这里是另一回事？它们通常会使用机器人账号进行诈骗，但你们并没有清除 90% 的机器人账号。没有，我认为我们，我们实际上清除了。我觉得你们清除了。好的，是的，是的，我认为我们清除了。是的，也许超过 19。

现在你至少说过，嗯，现在要在 Twitter 上运营一个平台并让它产生任何、任何，呃，优势，已经困难得多得多了，嗯，所以，呃，在机器人账号方面有了显著改善，在检测，呃，某种试用军团的能力方面也有了显著改善，这有点不同，那种情况是，比如说，哦，你知道，嗯，在一个低工资国家的一间仓库里有100个人，每个人都坐在一张放着100部手机的桌子前，所以你就有了10 000个真实的人，然后他们会协同行动，围攻某个特定话题，或者让某件并不受欢迎的事情显得非常受欢迎，嗯，而我们已经能够击败其中几乎所有，我们认为实际上仍然能够运作的已经非常少了，所以系统的质量已经好得多了。好，那么，嗯，如果你对琳达说过

无论这会造成什么商业影响，你都会继续直抒己见，而她同意这一点了吗？她对此满意吗？你们意见一致，呃，是的。好，那么，以她的 CEO 身份，她对内容审核有任何决定权吗，还是那归你管，或者你们会一起做？嗯，总体原则是，我们当时，我们当时在这里贴近法律。因此，对于任何一个特定国家，我们都会尽量贴近低限度。我们的法律在各国之间非常，而我们不能简单地在，嗯，在另一个国家对抗法律，因为他们会直接切断我们，嗯，所以总体原则是，尽我们所能让人们之间能够进行自由、开放的交流，前提是他们没有像我说的那样违反。她认同，认同那个计划，是的，那个重点。

嗯，是的，有一件重要的事，就是，这显然并不意味着，比如说，广告商应该被迫出现在任何内容旁边。所以我们还开发了邻接控制功能，以确保如果你所宣传的是，嗯，比如 Disney，Disney 就是一个大广告商，它不使用儿童电影的广告，他们，你知道，不会希望附近的内容是某种适合全家观看的内容，这完全可以理解，嗯，所以，呃，所以并不是说广告商必须出现在他们、他们不认同的内容旁边。那么你能，嗯，有些人说你发推时可能有点反复无常，或者至少，嗯，会发布范围广泛的、的内容，嗯，有没有人说我不想出现在埃隆·马斯克旁边

那是、那是平台上发生过的事情吗？我还从未听说过。嗯，但是，呃，嗯，这件事有没有向琳达提起过，比如你发什么推文，以及那是否会影响广告商？她有没有就此问过你？呃，她确实问了，事实上是在我们于迈阿密举办的会议上。所以我们……言论至高无上。好。嗯，我想稍微问问你对Twitter作为一个社区、作为一场对话的愿景。嗯，你谈到过希望最大化无悔时间。你能不能解释一下这是什么意思，以及你如何衡量它？呃，是的，所以，呃，以前Twitter主要关注一个叫作……他们称为mDAU，即可变现日活跃用户的数字。嗯，但问题

是，当你仔细观察时，会发现那里面有一大批用户，嗯，甚至从未访问过Twitter。他们会去……呃，会在手机上看到一条关于某条推文的通知，但实际上不会点击进入网站。所以真正重要的是用户真正使用屏幕的秒数。嗯，所以那就是我们现在追踪的指标，它依据的是，呃，呃，iOS、Android和浏览器向我们报告的屏幕使用时间。所以，嗯，它必须是……必须是应用在前台运行的时长，对吧？这是评估这一点最严谨的方式。所以，所以当你说“无后悔……”抱歉，请继续。没错。所以就“评分不足”而言，这要稍微难衡量一些，嗯，但我们当然可以通过轶事方式收集，也就是

说，如果你昨天在Twitter上花了，你知道，半个小时，那么其中有百分之多少、那段时间的百分之多少是你后悔的？嗯，而记录反馈按钮一直非常积极，呃，他们、他们认为这些信息有用、有娱乐性、很有趣。嗯，所以据我所知，我们似乎正朝着正确的方向前进。我当然愿意听取在座各位的任何批评。好吧，让我就此问你一个问题，那就是，嗯，你知道，你最近发了一条关于乔治·索罗斯的推文。你说，嗯，让我回去……好吧，让我直接说原话，因为我挺想知道你对此怎么看：“他想侵蚀文明的根基本身。索罗斯憎恨人类。”这显然在Twitter上引发了双方大量回应，许多不同的

观点。那算是不会后悔的时间、无悔时间吗？你引发的那场争论，你认为它属于那一类别吗？嗯，我是说，我说的是，索罗斯让我想起万磁王，你知道。呃，好吧，你知道，你说的可不止这些。但还是不讨论关于索罗斯的那条推文本身，你知道，你显然是Twitter上的一个重要人物，而你正在定下一种基调和一个名字。所以我只是想知道，这类被触发的争论，是否符合你试图在那个新市民广场中建立的定义？嗯，我的意思是，我认为重要的是，比如听着，我所说的并不是……就是我会说的话。嗯，你知道，这有点像一个市民广场。我不会缓和自己所说的话，因为那会构成抑制

至于你的言论自由，这并不意味着，为了算作言论自由，你就必须同意我说的话；如果有人说完全相反的话，也不意味着他们、他们不会在 Twitter 上得到支持。嗯，重点是要有一套转移性的观点，而言论自由只有在这种情况下才有意义：如果是你不喜欢的人发表言论，说了你不喜欢的话，这被允许吗？如果、如果允许，你就有言论自由，否则就没有。对于那些主张审查的人，我想说，如果、如果你成功做到了，那么审查转而针对你只是时间问题。我同意，我是说，你可以——那是你所说的言论自由定义，但我只是好奇，在不可后悔的那部分，你、你、你、你试图进行什么类型的对话

试图实现什么，以及那是否属于可以接受、但也许并不是你希望更广泛的对话走向的东西。嗯，我确实澄清过，你知道，我对 Taurus 的一些担忧在于，他资助了全国各地数量非常庞大的、规模较小但有影响力的竞选，尤其是地方检察官竞选。我们发现，例如 alien 以及旧金山地方检察官竞选，呃、呃，有 chess boarding，还有那个家伙，我总是、我总想叫他《美女与野兽》里的 Gaston，开车，还有、还有，这基本上是，他、他，嗯，当然有大量当选的 ASP 对犯罪非常宽松，而且经常会拒、拒绝起诉。所以你当时基本上是想借此表达一个更深层的观点

简短回答，是的。嗯，我能不能快速转到——因为我不想在那个、那个兔子洞里走得太远，因为那场辩论已经在 Twitter 上进行过一些——嗯，你知道，你们现在又接近盈利了吗？Twitter 还没完全到那里，但我们、我们不像，你知道，收购刚完成时，我会说那就像被传送进一架正向地面俯冲、引擎着火、控制装置失灵的飞机里。嗯，所以至少可以说，这让人安心。嗯，现在我们必须采取一些相当强硬的，呃，公交削减公司健康，但目前我们正朝着——如果运气好，我们下个月也许会随意、随意地转正，但还有待观察。现在的员工配置是你想要的水平吗，还是你打算

从这个水平重新开始增加？我想它已经从 8 000 人降到了大约 1500 人之类的，是不是很疯狂？正确，是的。嗯，我认为，你知道，确实，我们会开始为公司增加人员，嗯，而且我们已经开始为公司增加了一些人员。嗯，而且，嗯，但仍然、仍然有很多变化必须发生。所以，不过我、我认为 1500 可能是个合理的数字。这是否展示了你在大型科技公司削减成本方面能做什么？我的意思是，当你环顾硅谷的其他大型科技公司时，嗯，根据你的经验，你会说那些公司也有空间进行大幅得多的改变吗？是的，我认为 Twitter 也许有些异常，因为，嗯

那里有很多人在做似乎没有多少价值的事情。嗯，我认为大多数硅谷公司可能都是这样，嗯，也许没有 Twitter 那么严重，但，呃，仍然，是的，我认为那些公司有可能在不影响生产力的情况下进行大幅削减，事实上还能提高生产力。嗯，所以，你知道，嗯，在任何一家公司，都有人帮助推动事情向前发展，也有人算是在试图猛踩刹车。Twitter 当时的情况是，你会开一个 10 人会议，你知道，其中 1 个人踩油门，9 个、9 个人各有一套刹车。嗯，所以你走不了多远，对吧。嗯，而且，嗯，所以现在、现在我们要回家了，大约

发布功能，即使会给网站稳定性带来一点风险，只要不是太严重就行。我认为就目前而言，可以比较公允地说，我们过去，呃，6 个月推出的功能比 Twitter 过去 6 年推出的还多。至于宕机，早期出现过一些宕机，你、你确信现在一切稳定了吗？嗯，宕机并不罕见，例如 Instagram 最近就发生过宕机，颇具讽刺意味的是，这是在 Twitter 上报道的。嗯，所以，呃，我们发生过宕机，但不是、不是大规模的，而且通常时间很短、范围有限。好的。嗯，你后悔收购它吗？你曾试图退出交易，还是说现在你很高兴买下了它？嗯，结局好就一切都好。现在已经有好结局了吗，还是我们仍然要等着看？嗯，我

认为我们正处在、有希望处在，呃，回升弧线上。好的，所以，我是说，你谈过的一件事是，你以 440 亿买下了它，你谈到它有朝一日价值会达到 250，我想是在内部会议上。你能不能谈谈你要如何达到那里？更宏大的愿景是什么？我的意思是，你现在希望把广告商带回来，顺便问一下，他们回来了吗？是的，是的。嗯，就你们失去了谁、谁又在回来而言，你能不能说明一下回归的规模？呃，我认为会非常可观。嗯，所以广告代理商在这一点上都，嗯，取消了对 Twitter 的警告。因此，我很感谢，例如 Group M，嗯，移除了对 Twitter 的那种，呃，担忧标签，这是一件非常重大的事。嗯，所以我

认为在这一点上，呃，我预计几乎所有广告商都会发起攻击。好的。嗯，我们也——我做了更多工作，让广告，呃，与用户更加相关，嗯，这样我们就会向用户展示他们更有可能有兴趣购买的东西。听起来很显然，但，对，这往往就是会发生的事。是的，也没有那么显然。我的意思是，就、就是一些基本的事情，比如如果你以前在 Twitter 上搜索，这是搜索横幅应用，嗯，它不会考虑搜索词，这相当疯狂。嗯，所以它只会展示一个随机广告。好的。而显然应该展示一个，你知道，与你的搜索相匹配的广告。听、听起来值得做。那么快速问一下，你谈过那种既能通讯、又能处理金融和其他事务的单一应用。我的意思是，你能不能稍微

详细谈谈你如何实现它，以及为什么美国想要它？显然，这将由人们决定他们是否想要它。嗯，关键在于，我们是否做出了足够有用的东西，嗯，让你想更频繁地使用它。嗯，很好，这就是我们的目标。嗯，所以我们不会做任何事情来阻止人们离开应用，也不会试图在应用里追踪他们，但我们要提供足够的编译功能，让人们对这个平台的使用量随着时间增长。那么 10 年后，广告仍将主导 Twitter 吗？我认为广告始终会发挥作用。嗯，在某个时候，比如 10 年后，它可能不会发挥最大的作用，但至少在未来几年里，它会发挥最大的作用。所以我想快速进行一轮，嗯，快速的远轮次

一些问题，就是，你知道，想象一下，深夜时你坐在那里，针对一些事情快速连续发几条推文回应。嗯，我要问你几个问题，如果你能只给我简短回答，然后我想转到 AR，稍微谈谈你、你刚才开始提出的一些更深层的观点。嗯，所以第 1 个问题，Twitter 会在 5 年内再次上市吗？我不知道。好的。嗯，你认为总部还会在旧金山吗？好的，目前不太顺利，我们再试几个。嗯，我们会在哪个十年攻克通用人工智能？我认为这个、这个十年。好的，所以会这么快。好的。嗯，你自己会参加一次 SpaceX 旅行吗？我会的，在某个时候，是的。不确定什么时候，但你会很好。嗯，哪一个是最

目前是一个令人兴奋的、适合建设Tesla工厂的国家。嗯，是的，我们确实宣布过，墨西哥将成为我们在美国以外的下一个，呃，地点。

S，呃，而且选好了一个地点以及所有东西，所以有那个，然后，嗯，我们大概会在今年年底前后再选一个地点，印度有吸引力吗，当然，嗯，你仍然是加密货币的粉丝吗，嗯，我的意思是，我并不建议任何人购买一种加密货币，或者把全部家当押在，你知道，狗狗币或类似东西上，好吧，在狗狗币方面，可能一直在想，也许你应该这么做，但让我建议你，那会是一个有一些方式，嗯，好吧，所以它确实很酷儿，是我的，是我某种最喜欢的加密货币，因为，呃，它的幽默感最好，而且，呃，有狗，嗯，不过我昨天确实看了它的价格，它，嗯，它比以前低了，我想，嗯，我不知道，也许，你知道，就像我的一个朋友有种说法，最具讽刺意味的解释最有可能，而

货币最具讽刺意味的结果将是，那个被当作，呃，一个笑话制造出来、用来取笑加密货币的东西，呃，最具讽刺意味的结果会变成，它成为全球货币，好吧，我们拭目以待，最后一个问题，嗯，你能分别按满分10分给美国和中国的人工智能发展排名吗，嗯，我的意思是，美国说非常拥有，呃，最先进的人工智能，所以这是你说的，就像，就像中国肯定紧随其后，并且拥有扩大规模和优化所需的资源，嗯，人工智能领域最大的单项进展仍然来自美国和欧洲，但是，嗯，好吧，所以很难给出确切的数字评分，因为它更像是，但仍然存在很大差距，确实存在一个，存在一个差距，嗯，那个差距看起来大约是12个月，对吧，鱼是在缩小还是全都在扩大

很难说，我猜它会在某种程度上缩小，好吧，嗯，你能谈谈吗，你自己创建了一家新的人工智能公司，嗯，显然这个领域有大量的能量和活动，或者至少人们一直在谈论，我的意思是，除了Tesla和，还有你之前谈到的那些东西之外，你自己想在这个领域做什么，那个新东西是什么，嗯，我认为这场竞赛中应该有一匹重要的第3匹马，呃，我们有OpenAI和微软、Google DeepMind，而且这场竞赛中大概应该有第3匹马，嗯，很快会再多说一点，但它会与Twitter的数据和Tesla的能力进行交互吗，它是否试图把你之前谈到的那些东西在

能力方面结合起来，成为那第3个参与者，那是你所说的吗，在某种程度上是，我不想在公告方面抢跑，但是，呃，嗯，你知道，开放中的人工智能与微软有一种关系，看起来运作得非常好，相当好，所以有可能，嗯，xAI和Twitter以及Tesla会有某种类似的关系，有可能，嗯，你谈到过监管的重要性，而且你呼吁实行这个，这个暂停，我的意思是，监管科技的历史好坏参半，监管机构一直很难跟上科技，更不用说走在科技前面了，你认为实际上需要发生什么，在这个领域切实可行地发生什么，才能尝试改变这种情况，因为显然这段历史并不令人鼓舞，是的，我的意思是，我认为

应该有，你知道，我长期以来一直在大力推动这件事。我会见过多位，嗯，中心人员和国会议员，在国会、在白宫，倡导对 AI 进行监管，呃，首先成立一个洞察委员会，由独立各方组成，也许还包括行业领导者中的参与者，而那个，那个那个监督委员会，呃，获得，嗯，或者应该说，让那个洞察委员会了解各家公司正在做什么，嗯，而且，呃，你知道，在存在某种竞争动态的情况下，你显然可以，呃，你会将那些或许存在利益冲突的董事会成员隔离开来，呃，但无论如何，如果你设立某种监管委员会，而，呃，而他们先从获取洞察开始，然后我提出了

制定规则，嗯，然后那个，你知道，会收到行业提出的意见，而，呃，然后希望我们能有某种改善安全性的监督规则，就像我们通过 FAA 对飞机进行监管、对航天器和汽车通过 Nitza 进行监管、以及通过食品药品管理局对食品和药品进行监管一样，对吧？而对于我们所谈论的这样一种全球性事物，这将如何运作？AI 以及各国之间的相对，呃，进展会非常非常重要，这种东西能够全球化吗？这是，这是，嗯，真正的关键问题是，呃，中国会不会，呃，你知道，与西方合作，这仍有待观察，嗯，但我仍会主张进行某种程度的监督。我的意思是，例如，我们对飞机有监管监督，可是

美国在飞机方面仍然做得非常好，所以它比其他地方，比任何其他地方都投入更多努力。所以，仅仅因为存在，你知道，FAA 法规，并不意味着，呃，它必然大幅减缓了发展。好的，嗯，那么，所以你的看法是，如今 AI 的变化会巩固科技巨头，也就是这个世界上的微软和谷歌的地位；它是否也会，是否也存在这样一种情景：它实际上有助于引入新参与者，并且并且并且并且并且改变这种动态？还是说这是一个可能性低得多的结果？嗯，确实有，嗯，很多 AI 初创公司。正变得棘手的是，要竞争你确实需要3样东西：你需要，呃，人才，有才华的人，嗯，你需要，嗯，大量算力、昂贵的算力，而且你需要

获得数据。所以，无论是谁拥有，无论是谁在这3方面取得成功，谁就会获胜，嗯。所以现在算力成本已经变得极其高昂，因此现在，你知道，某种程度上的最低入场筹码，我会说，呃，最低是价值2.5亿美元的白银硬件，最低，那就像只是2个，对吧，以任何方式保持相关性。所以你的意思是，初创公司更可能借助其他公司正在做的事情，而不是自己直接竞争。对，要训练一个大型的，嗯，你自己会进行一次 SpaceX 旅行吗？要训练一个规模大概相当于 GBT 5 的模型，如果他们使用至少30、30 000，也许50 000个 h100s，我不会感到意外；它们是最新的，呃，gpus，不确定这个词是否完全准确，但这是 Nvidia 的最新技术，嗯，所以，而且

然后你还需要运行推理，嗯，所以有很多那个，那个 G fuser，嗯，目前得到它比得到毒品难得多。其实这在旧金山真的算不上多高的门槛。你之后可以多跟我们说说这个，不过，是的，好的，好的，嗯，那么，嗯，我只是想深入谈谈 AI 的几件事，我很喜欢你对此的观点，嗯，这会，这对社会意味着什么？就它是否会把财富和权力固化在一个非常小的群体中，并导致不平等大幅扩大而言；还是会实现民主化并产生相反的结果？你觉得，就商品和服务的获取而言，这会走向何方？我认为 AI 将开启一个富足时代，前提是我们处于一种善意 AI 情景中。我认为 AI 将能够制造商品

和提供服务，嗯，而且成本非常低廉，所以对于任何一种不存在人为制造的稀缺性的产品或服务，比如说，我想恰好住在这个，你知道，社区的房子里，那就像，好吧，那里只有100栋房子，所以，你知道，那仍然会存在稀缺性，嗯，或者一件独一无二的艺术品会有稀缺性。但凡不存在我们有意设计出来的稀缺性的东西，在善意情景下，每个人都将充足拥有。而在 unbeline 情景下，嗯，会有各种各样的……但当你审视这件事时，你最担心的是什么？你知道，你多年来一直在谈论监管的必要性，真正让你夜不能寐的是哪种情景？嗯，我不

我不认为 AI 会试图摧毁全人类，但它可能会把我们置于严格控制之下，嗯，我的意思是，而且它变成终结者的概率并非为零，不是0%，但我认为，它毁灭人类的可能性很小，但并非为零。我们希望它完全为零，尽可能接近零。然后就像我说的，呃，AI 掌控一切，或者保障所有人的安全，并接管地球上所有计算系统和武器系统，实际上就像某种超级保姆。但难道不是，不是还有另一种情景吗？如果你说，你说是的，比如，你知道，比如说你是，呃，你知道，我必须，好吧，接触同意参赛者，假设性地说，嗯，不太可能，我们面对现实吧，嗯，而且而且

你知道，你说，你想要什么，于是我想要世界和平，嗯，而，呃，这就像，好吧，你知道，实现世界和平的一种方式是拿走人类的所有武器，让他们再也无法使用武器，并且惩罚任何从事，嗯，你知道，域外活动的人。但更，更可能出现的糟糕结果，难道不是与其由 AI 接管并成为让我们所有人都做它希望我们做的超级安全之事的终极保姆，实际上是某个邪恶的人利用这种力量来实现社会控制兼军事优势，嗯，并且实际上世界上某个国家决定以不同方式使用它？呃，是的，那就是我所说的，比如 AI 被用作武器，对吧，嗯，而且笔

比剑更有力量。因此，我们首先必须警惕 AI 被使用的领域之一，就是在社交媒体上操纵公众舆论。所以，呃，Twitter 将转向以订阅为主的系统，原因是创建一个拥有可信运营商验证过的电话号码、有一张信用卡、并且每月支付少量费用的账户，要困难得多，就像困难10 000倍，呃，而且要让这些信用卡和电话号码高度分散而非集中，难度高得不可思议，嗯。所以过去，呃，某个人可以以每个1美分的成本创建100万个虚假账户，然后进行操纵，让某件事看起来非常非常受公众喜爱，而实际上它

并没有，或者看起来被推广和转推，而实际上并没有；它的受欢迎程度并不真实，实质上是在钻系统的空子。所以，转向，呃，一种基于订阅的验证，我认为是非常有力的，而你确实将无法信任任何不这样做的社交媒体公司，呃，因为它必然会被机器人以极其极其严重的程度占据。那么，如果我们回到一开始的话题，看看即将到来的选举，过去几个月 AI 能力的这种巨大转变会发挥多大作用？这种转变显然会持续到明年。你认为，这会对信息传递方式，以及人们被告知的候选人不同形象，产生多大的影响？我认为那是某种

我们需要着重去研究的是，确保我们正在尽量降低人工智能操纵的影响，嗯，我们当然非常认真地对待这件事，在 XX Twitter，你知道，Twitter，嗯，我认为我们正在部署所有保护措施，以尽量减少系统遭到的大规模操纵，并且在看到这种情况时肯定能检测出来。好，但除了 Twitter 之外，你是否担心这会影响总体上的选举？是的，嗯，可能会有人试图利用人工智能操纵公众，其中一些尝试会成功，嗯，即使不是这次选举，也肯定会是下次。好，如果你有时间，我还有2个关于人工智能的问题，然后再稍微谈谈中国和 Tesla，如果可以的话。嗯，第1件事是，嗯，我们经常谈到

人工智能时，会谈未来5到10年，以及它会对就业和其中一些事情产生什么影响。如果着眼于长得多的时间跨度，考虑到变化的速度和规模，再看看你的孙辈和曾孙辈，你能否让我们大致了解一下，做人会是什么样子？到那时，这会在多大程度上改变我们作为一个种族运作方式的根本性质？我认为它会改变很多，嗯，尤其是如果你展望更遥远的未来。我的意思是，一切都会自动化。我的意思是，会有家用机器人，你完全可以像与人交谈一样和它们交谈，嗯，它们可以在家里帮助你，或成为陪伴者，或做任何可能需要做的事。呃，会有人形

机器人遍布，你知道，工厂，嗯，而且，嗯，我们的也将全部自动化，任何能够运用智能的事情，嗯，即使只需要有限的智能，也会实现自动化。比如说，如果你说从现在起10、20年后，我们会不会通过某种 Neuralink 类型的神圣……与那项技术连接起来？我的意思是，很显然，在你看来，这就是它的发展方向吗？嗯，从大脑皮层到那个的高带宽接口，所以也就是计算或人工智能的第三层，它其实已经存在了，呃，你知道，只是我们没有高带宽连接。嗯，我们有一个，嗯，边缘系统，它算是一种基础要素，也就是我们的本能和欲望之类的东西；呃，它们不是位于其上、负责思考的皮层

部分；在我们的脑中，还有一个第三层数字层，目前以我们的手机、电脑、笔记本电脑之类的东西，以及所有应用程序的形式存在。而实现更好的、更好的融合所面临的限制，或者说这种构造，你知道，让人类利益和机器利益保持一致所面临的限制，是带宽，尤其是输出带宽。所以如果你选择，你能以多快的速度向计算机输出？现在是使用语音或手指，而它们的移动速度非常慢。所以你说的可能是每秒10比特，或者某种相当低的数据速率。嗯，所以借助 Neuralink，你可以把它提高，你知道，可能提高100万倍。嗯，所以一切都只是加速了，它加速了。是的。好，嗯，我的意思是，这显然是在一个相对

良性的情景中，因为问题不只是，呃，假设这是一个布莱恩情景，嗯，我们甚至要如何领会或理解计算机正在做什么，对吧？嗯，我们要如何，我们要如何搭上这趟车？嗯，如果我们有一个更好的，如果你有一个快上100万倍或更快的大脑机器接口，我不知道，那我们就能更好地搭上这趟车，相比之下，如果我们是用2根移动缓慢的肉棍操作手机，嗯，如果你要这么说的话。嗯，说到这方面，你有很多孩子，这个房间里很多人也有孩子，你知道，他们需要具备什么，他们需要具备哪些技能？你认为最重要、并试图传授给他们的3项技能是什么，才能让他们做好准备并处于有利位置

来面对这个新世界？嗯，我认为，对许多不同学科拥有广泛的理解，这很重要。所以我认为常识很重要，嗯，这样你至少能大致知道自己在不同领域里不知道什么，然后深入钻研你的孩子有浓厚兴趣和能力的领域。所以要找到那个重叠之处，也就是我的孩子在哪里，对这个感兴趣，并且具备某种取得成功的能力，那么，呃，你知道，找到，如果你能找到那个维恩图的重叠区域，那么显然鼓励这一点是好事。嗯，而且我们显然正迈向一个高科技世界，所以对计算机、软件和人工智能有一些基本了解，或许是个好主意。好，但实际的大方向，嗯，我的意思是，工作会发生变化

但更多会是由人工智能赋能，让工作变得更好、更容易，而不是你所需技能发生全面而彻底的改变。原则上，我们这里说的是什么时间跨度？所以，如果你说的是20、30年的时间跨度，我认为一切会发生超乎想象的转变。你大概会，你大概认不出30年后的社会。嗯，比如我确实认为我们已经相当接近了。你问过我通用人工智能，我认为我们可能只需要3年，也许6年，就能实现它，就在这个10年里。嗯，所以可以说，我们正处于社会超级智能这个黑洞的事件视界上。好，我要问最后1个问题，然后看看你是否有2分钟时间回答几个现场提问，以及

这个问题，它又回到了中国，你刚才稍微谈过这个问题。Tesla 在中国有很大的业务，嗯，而且显然，你知道，你正处在那条正变得，嗯，可能很有意思的地缘政治断层线上。这在多大程度上影响了你的决策，比如你如何在当地配置资产和其他东西？作为一个商人，你对此有多担忧？这个房间里很多人在中国都有业务，这会不会让我们的处境变得非常、非常困难？嗯，从根本上说，台湾有一个正在走向紧要关头的问题，嗯，究竟什么时候布什会现身还不清楚，但看来很可能会发展到非动真格不可的地步，事情正朝那个方向发展。嗯，我倒是认为，那会发生什么

其结果对全球经济而言将是绝对灾难性的，嗯，但是，嗯，你知道，中国已经非常清楚地说明了它在中国问题上的目标，以及，呃，某种程度上，嗯，把台湾纳入中国。因此，人们无需揣摩字里行间的意思，只需读那些字句本身。他们曾是手镯，而且他们没有得到……而最大的担忧，除了，你知道，冲突本身本身本身本身的前景之外，嗯，显然，全球很多高端芯片来自台湾。我的意思是，如果这种供应被切断，后果会有多么灾难性？嗯，还有更多东西来自中国，所以这种转运很多。全球制造业中如此多的繁重工作都由那里承担，尤其是当制造工作是，你知道，纯粹的苦活，而且可以说并不是特别光鲜时

嗯，指控做了大量艰苦工作，大多数人根本不知道他们做了多少艰苦工作，所以，嗯，你有点从台湾来的担忧要少得多、少得多，不如从中国被覆盖那么令人担忧。现在，中国当然也会相应受损，嗯，因为我要说，中国和台湾的经济，它们就像连体双胞胎，与西方经济、与世界其他地区连在一起。所以从经济角度来看，中国、中国、西方、世界其他地区成为连体双胞胎，将意味着分离的后果确实会非常严重。好吧，如果那发生了，我希望它不要发生。所以，嗯，这里没有简单的解决办法，但如果有任何、如果有任何通往外交解决方案的路径，我们确实应该认真对待。

很好，埃隆，你有时间回答几个现场提问吗？当然。有人想提问吗？我们先请这里，然后这里。迈克，在你身后。谢谢，埃隆，谢谢你参加，嗯，我是英格兰东北部纽卡斯尔一家房地产企业的创始人，嗯，我们在东北部制造并出口的汽车比整个意大利还多，呃，你愿意让我在英格兰东北部为你建一座 Tesla 工厂吗？谢谢你的提议，所以未来选址建造一座盖革工厂时，可能会考虑，呃，英格兰。谢谢。我会联系你。所以你会、你会考虑那里。你们正在积极考虑吗？呃，我们、我们、我们目前没有在考察新地点，但我们会在今年年底前后推进。我会给你发一些方案。

好的，嗯，呃，请大卫提问。感谢你抽出时间。呃，聚变，许多科学家说它可以改变世界、地球，它就是太阳。这个星球以及火星上的所有生命都依赖聚变，也就是太阳本身。我能问一下，为什么像你这样拥有卓越头脑、资源和才华的人，实际上没有专注于聚变？我认为它可能会改变社会的游戏规则，而不是专注于 Twitter，那里有许多正派的媒体公司能够做这件事，而且我要说，那几乎是一种微不足道的方式。嗯，我认为我们已经在天空中拥有一个巨大的聚变反应堆，叫作太阳，它每天都会出现。所以，嗯，我一直说，如果你想知道站在未来聚变反应堆前是什么感觉，那就出去，去站在太阳前面，你知道。

直接走到外面去，那就是、那就是一个巨大聚变反应堆给人的感觉，因为太阳就是这样的，嗯，它每秒把大约450万座城镇、450万吨质量转化为能量，而且不需要维护。太神奇了，嗯，你不必、你不必给它补充燃料，也不必维护它，它就在那里。所以我对聚变的建议是太阳能和电池，我们完全可以只用光伏和电池为整个地球供能，不，我是说，并不容易，但确实有一条非常清晰的路径可以做到，嗯，而且、而且，呃，没有奇迹是安静的，只管工作。有意思。我也支持风能和、和核裂变、地热、水力发电等等。我们再回答几个问题，这里一个，然后那位在……的女士。

后面。谢谢。在你考虑对中国的敞口时，尤其是，嗯，在电动汽车领域以及电池供应链方面，你评估近期和中期政治风险的流程是什么？我想我就是和我的团队谈谈，你知道，看看新闻，呃，我不知道，评估、评估一下 Twitter 上的意见吧，嗯，你可以在 Twitter 上看到一些对特定主题具有世界级专业水平的人所作的非常深入的分析。所以，嗯，我不知道，我认为我们只是、我们努力做最坏的准备，抱最好的希望，嗯，而且，嗯，你知道，确保我们在全球不同地理区域设有工厂，并让供应链尽可能本地化。但这对于强制 Azure 情况也很重要。所以，如果发生地震、野火。

骚乱、革命，呃，冰暴，呃，热浪，呃，诸如此类，呃，我想到了这个时候我已经全都见识过了，所以你会想要拥有，你、你会想要一条不会承受来自整个地球各处不可抗力影响的供应链，嗯，因为某件事总会在某个地方发生，这是一颗大行星，所以、所以这就是为什么我认为拥有，嗯，本地化的供应链，并在许多地理区域设有工厂很重要。呃，最后一个问题，由后面的那位女士提问。谢谢。是的，你曾发过一条著名的推文，说你认为人口崩溃对人类构成的风险远大于气候变化。你认为各州、家庭，甚至公司可以做些什么，来确保我们当中有更多人愿意生更多孩子？嗯，是的，我的意思是，我认为如果你看看出生率，这非常能说明问题，而这些出生率

只是，你知道，公开可得的，嗯，你可以查看，呃，比如去年每个国家的出生率，嗯，网上就有，嗯，而且你可以看看两种比率的趋势，非常清楚的是，这一趋势一直在大幅下降，嗯，而且我们最近创下了历史新低。所以，嗯，你会想，如果，你知道，在秘密期间，你知道，你还能做什么呢，还不如生个孩子，但，嗯，实际上并没有发生这种情况，我们在秘密期间的出生率大幅下降，离婚也有所增加，因为和他们的另一半相处了很多时间，嗯，所以我总体上认为，嗯，所以只要改变人们对于生孩子是一件好事的想法，为了、为了延续文明，生孩子是非常重要的，嗯，而且我认为有时候

这被视为，呃，你知道，孩子被视为强加给世界的负担，我完全不认为是这样，或者有时人们认为世界上的人口太多了，那、那肯定不是事实。你可以、你可以让地球上的所有人都挤在纽约市的一层楼面上，你知道，那会很不舒服，但、但这只是为了让你了解地球上由人类占据的横截面积有多么小。如果你身处大城市，它看起来才会很大，嗯，但对于地球上绝大部分地区而言，如果、如果给你一个任务，让你从前面的飞机上扔下一颗保龄球，并且、并且、并且你必须砸中某个人，你会、你会扔偏，我几乎从来砸不中任何人，嗯，所以，呃，那是什么，你坐飞机时很少会飞越某个人的上空，嗯，你从洛杉矶飞往

纽约，大片陆地区域完全没有人。所以无论如何，我认为我们总体上希望社会鼓励人们生孩子。我认为公司当然需要支持，呃，有孩子的员工，嗯，我认为就政府激励措施而言，呃，应该有，呃，我想，为生孩子提供一些税收减免，你知道，或者干脆让养育孩子在经济上不构成负担，嗯，而且始终值得记住，比如先把自主性放在一边，嗯，如果某个人没有孩子，你实际要求的就是，呃，让别人的孩子在你年老时照顾你，而那、那似乎不太对，你知道，嗯，因为、因为那就是、那就是在没有自动化的情况下，他们将被迫去做的世界，也就是别人的孩子将不得不

在你年老时照顾你，而且，嗯，你知道，所以我认为，无论如何，我们都需要解决这个出生率问题，否则文明将会趋近于不存在。那不正是 AI 发挥作用的地方吗？它会替我们完成所有工作，这样我们就能、我们就能应对一个，呃，可能更少的人口，或者说你讲的情况。我认为会出现非常自信的机器人保姆，所以这会有所帮助。

Elon hello welcome are you in Palo Alto I understand yeah first well a Global Engineering headquarters in Palo Alto great well thank you so much for joining us um I'm actually going to start somewhere a little bit differently than I expected um because I just saw on Twitter on Twitter um a little announcement um I would love to get your your thoughts on it I see your uh interviewing Rhonda santis tomorrow morning is that right on Twitter spaces I see yes it's uh I need like the exact time but oh tomorrow morning your time I think it would be probably correct okay um so uh yes I um will be interviewing um Ron Santos and he has quite an announcement to make um and we'll be we're the first time that's something like this is happening on social media and

with the real-time questions and answers uh no not scripted uh so it's Gonna Be Live and Let Let let's see what happens foreign and you've been tweeting some Tim Scott stuff in the last few days um what should we be thinking about who you're backing obviously this interview tells us something can you give us a sense of where your thinking is at the moment yes I mean I'm not at this time um planning to endorse any particular cancer um but I am interested in uh you know X slash Twitter being somewhat of a public town square and uh where more and more organizations host content and make announcements on Twitter um it's the it's the only place on the internet to really get uh real-time like down to the minute and second news and it's uh yeah so I think it's

it's quite groundbreaking that there'd be um a major announcement of this type on social media um and should we expect um sorry I don't want to go on too long about this but um in your new role as interviewer rather than interviewee uh should we expect um uh should we expect more of this I mean if it's the Town Square are you going to be interviewing other candidates Democrats what's your what's your thought of this are people willing to come are you going to be there to execute the Town Square across the Spectrum yes absolutely um so just as I as I promised when I do a series of media interviews I did a range of interviews um and uh I guess this would be also a media interview um so ranging from sort of on the you know left moderate to what's considered

right um and I do think it's important that Twitter be uh have both the reality and the procession of uh Level Playing Field of uh a place where uh all voices are heard and where uh as the kind of dynamic interaction that is you don't really see anywhere else um I'll meet you today on Twitter for example um uh AOC got into an and Ted Cruz got into an argument on Twitter which was um yeah independent of which side do you agree with um it's still very entertaining so what um and I'm sure tomorrow will be entertaining we're all going to be tuned into that but um when you when you approach an interview like that and and obviously a really important election like the one that is coming up you can you just talk a little bit about what are the key issues that

really matter for you at this pivotal moment you can matter for me as an individual or matter for you as an individual um in terms of who leads the country but also you know more broadly than that you know for the country and for for your businesses I mean can you give your sense of of where the real issues lie here well I've said publicly that um my preference and I think would be the preference of most Americans is really to have someone fairly normal in office I think we'd all be quite uh quite happy with that actually um you know I think someone that uh is representative of the moderate views that I think most the country holds in reality um and um but but the way things are set up is that we we do have a system that seems to push things towards the

edges because of the primaries so in order to win the primary you've got to win obviously majority of your parties vote in both cases that tends to cause a swing to the left and the right although I think things are more complex than simply left and right during the primaries and then uh and then let's shift towards the center for the Journal election um as far uh what what I think is yeah so I would really just like someone you know fairly normal accessible to be the president that would be great so if we go through the if we go through the four names in the frame at the moment can you just give a sort of yes no whether they're normal and sensible um so um we've got Joe Biden I I think you know be careful about these statements um so I uh we'd maybe

have to have a few drinks before I would give you the answers I will um I will look forward to that and I look forward to the uh to the conversation tomorrow and obviously a lot more of those to come over the coming months so um that's great thank you very much um so what I wanted to start with you've just flown in I think in the last 20 minutes um you live a pretty hectic lifestyle um but you've said that the only true currency is time can you give a sense to the people in this room who are scheduled within an inch of their lives sort of how you you what is a day in the life of Elon Musk what does that look like well weddings are very long and complicated as you might imagine yeah um and um and there's a great there's a great deal of context switching

so um doesn't mean my life called like relating to Doom where it's like fear is not the mind killer context switching is uh so switching context is is quite painful but um I do generally try to divide uh companies that's predominantly one company on one day so today is a Tesla day for example um although I might end up at Twitter late tonight um and then tomorrow would be partly a Tesla day as well have Twitter and um and then Thursday would be sort of a half SpaceX have Tesla day but these things are somewhat intertwined so the time management is extremely difficult um and this is going to sound pretty strange but I um I only have one uh part-time assistant um how many days a week is that the part-time I mean I suppose I suppose how she works who were

technically full-time but as such I I do most of the scheduling uh myself um and the reason I do that is because it's impossible for someone else to know what the priorities are um so the and since the most valuable thing I have is time I schedule it myself for the most part so if you come into Tesla today do you have a series of meetings set up or do you um come in with you know something on your mind and you go in and see people I mean how structured is this or if you shop at Twitter in terms of the people working for you how do they how do they handle that um yeah so today I have several hours of scheduled meetings at Twitter um so there are a number of of things that are operated on a weekly Cadence um and so those meetings are already set up and

then I have supplemental meetings at the end of the day um particularly I I I I won't be going to sleep until probably 2 A.

M or something like that and be working almost the entire time and if you're shaking this yourself is AI going to be helpful over the next few years to help you um do this and you're going to be using technology to help you manage that I guess we'll all be using technology I don't use a lot of AI myself day-to-day I mean Tesla AI is actually very Advanced for real world AI it's the most advanced remote World AI um by far um and in fact if positions were swapped and it would say up to Microsoft and open AI to create uh who could create the best large language model basically if the tests were swapped Tesla was given that the task of making the most competitive large language model and Microsoft open AI were tasked with self-driving Tesla would win okay

I don't think people understand the degree of depict the capability of Tesla's AI system so while I don't use AI a lot personally Tesla uses a trans mount we'll get on to that in a second if that's okay but one final thing in terms of just the management of what you do with your life you're running three very big companies you have very big stakes and you know ownership control of of two of those at least um what is your succession plan if you suddenly call and execute what you're doing both in terms of who runs the companies but as importantly who votes those shares in terms of um you know what happens longer term and strategically what have you got a plan for all of those yeah um succession is one of the toughest age-old problems um you know it's it's

a it's plagued um you know countries Kings Prime Ministers and presidents for and CEOs for you know since the dawn of History um there is no obvious solution I I mean there are particular individuals identified as that that I've told people would look if something happens to me unexpectedly this is who my this is my recommendation for taking over so in all cases the board is aware of of who my recommendation is which they may choose to it's up to them of course they they may choose to go a different direction but I they there is a in in worst case scenario this is who should run a company uh the control question is a much more it's a much tougher question um and something that I'm wrestling with and I'm frankly open to ideas because it certainly is true

that the companies that I have created and are creating um collectively possess uh immense capability and so the stewardship of them is incredibly important um I want to make sure that the stewardship is ultimately uh accrues the benefit of humanity that's the idea is the tolerance of civilization um but they're not that we're always successful in that but that is aspirationally our goal um so um I I have one one idea which is sort of partly in place which is to create kind of a an a sort of a educational institution that that would control um most of my vote um but this is this is not a case of um automatically I I'm definitely not not of the school of automatically giving my kids a you know some share of the companies even if they are not even if they

have no interest or inclination you know or ability to to manage the companies I think that's a mistake um so but but it's a very hard problem to solve right so and then who should be on the board of directors of the educational institution is also a very very hard job to solve um so I think probably some disaggregation of control would make sense I'm really just kind of thinking out loud creatively here but it sounds like it's something that you I mean you need to get planning of who those people are going to be because as you've said whether when you look at SpaceX you look at Tesla you look at Twitter these these matter to society a lot and having the right people to take those votes on the future of where they go and where money gets spent is is fairly

important yes absolutely now the goals of the companies they're the achievement of those goals varies considerably in difficulty um you know the the original goal of Tesla was to accelerate the Advent of sustainable energy which actually I think we've done done that to a significant degree um and have actually it's kind of of uh it's kind of it's kind of a order industry CEOs too often acknowledge Tesla's role in accelerating electric vehicles um and um yeah so so that that I feel has a lot of momentum um they're still solving self-driving which were you know aspirationally hoping to do this year uh and um and so it has got a long way to go but but the execution plan is relatively clear um and uh the next that execution plan will generate a lot of positive

cash flow for the company so it's like if it's a fairly obvious thing to do with SpaceX it's a harder problem because uh the long-term objective is to make life multi-planetary with a self-sustaining City on Mars which is likely to be very casual or negative uh at first a long term let's just say the target market on Mars is small you got to think long term yeah you're also going to have to get on very well with those you go with I would imagine uh yes definitely um you know sanity will be a prime requirement for uh and stability for traveling to Mars you don't want someone going nuts and opening the airlock in the middle of the night um right so SpaceX is a harder problem because it's it's uh much a long-term goal and and with uh a lot more money less

along the way so gotta make sure that that happens um and is that the sort of thing just to stay on SpaceX and we'll go to Twitter and tell us in a second but um is that the sort of thing that you'd like to lock in to the goals of SpaceX that Mars Remains the ultimate um ambition of this come what may is that is that that important to you yeah I mean it's to make light multi-planetary such that um and the key threshold for multi-planatory is that if the supply ships from Earth stop coming for any reason that a Mars does not die out that's that's the that's the critical great filter if you talk about things in terms of the Fermi Paradox the great filter is um Mars being self-sustaining without any resupply shows from Earth until we reach that point we're

really just a one planet civilization with uh an extension but the point in which the planets are self-sustaining Oman is self-sustaining then even in a worst case scenario of of Earth civilization either dying with a bang or a whimper um then Mars would have a much better chance of surviving um so the intent here overall is to ensure the light of Consciousness which appears to be just a tiny candle and a vast Darkness um I frequently I frequently get asked have I seen any evidence of aliens and I I I'm not um you know apart from the fact that I didn't at one point have an alien registration card when I was getting my uh green green card uh it's an alien registration um indeed possibly a slightly different type of alien but um so do you think you'll will

you live to see Mars happen I I hope to love to see the first humans on Mars um but I think it'll take some period of time beyond that to make my self-sustaining so it's at least 20 years from the first visit to make my sustain as my guess and it may be 40 or 50.

and that's assuming you really go for it right so that's a tough one um but like so I think important for improving the survivability of civilization and who's going to pay for that I mean are your investors gonna put the money up to do that are you going to expect government to fund that where does that money come from because as you say you can make a return on starlink you can make a return on launching satellites for other people and space tourism but I mean that's a that's a tougher return Isn't it yeah I think I think long term uh the value of it will be incredibly high I would just it's just beyond the planning Horizon of most people or most investors so um I mean obviously if um if there's a thriving um City on Mars and there's a lot of interplanetary

Commerce and SpaceX is the primary provider of that it would be immensely valuable um so um but but you know the one thing is that at the be this self-sustaining you know Colony um and um I think we we I think we uh generally operate with too much of an assumption that Civilization is robust um and nothing could really take it down uh a sentiment that has been common throughout history among Empires shortly before they called it so and you know I have to say that you know there's a little bit of late stage Empire Vibes going on right now um uh yeah for sure are you um yeah um uh is is AI something that in your view accelerates the risk of that or increases the risk of that at that outcome I think it does yeah um I mean we could definitely make a city

of Mars self-sustaining without without AI or without sort of AGI which is generally artificial general intelligence or super intelligence I have so I think that that is uh it's not necessary for anything we're doing um but it is happening and happening very quickly so uh there is a risk that uh Advanced AI um either eliminates or constrains Humanities growth I was more thinking the opposite does it increase the chance that plant the planet self input implodes and those things come true you I mean how concerned are you about these developments right now sort of accelerating your your bad case scenario here well uh I mean the development of artificial digital sort of super intelligence uh is very much a double-edged sword so it's if you have if you have

a genie that can grant you anything they can also do anything um that necessarily is presents a danger and I expect the first uses of AI to be or certainly the first government uses of AI to be weapons technology so just having um more advanced weapons on the battle if you go back that can react uh faster than any human could that's that's really what AI will be capable of I mean future Wars between Advanced countries or at least countries that have significant drone capability will be very much the Drone Wars so I want to get back to AI because this is this is big stuff and I'd like to to talk about in more detail but I do want to come just come back to the present from a long way in the future um you've just hired a new CEO uh Linda yacarino an ad veteran

um into Twitter you usually focus on hiring Engineers you know Linda is a very different person can you just quickly tell us about your courtship how did that go down well um we had conversations over a number of months uh just relating to advertising um and then uh Linda felt that um it would be very helpful for the advertisers to see me in person so invited me down to a conference in Miami um which was very helpful and met with a number of advertisers personally to assure their you know assure them that that Twitter is a good place to advertise in a and generally that in fact that that hate speech is declined which it has and that the quality of the system especially with respect to scammas and spammas is dramatically better than it used to be um we've

gotten rid of at this point well over 90 percent of those the scams and scams the scams and spam on Twitter it should be quite rare at this point at UCS camp um so um we've also rolled out uh sort of so just to be clear when you say you've got rid of 90 of the scams is it but is that the same thing as the Bots or is this scams in general and Bots is a different animal here they were typically used Bots for scan but you haven't taken the Bots down 90 percent no I think we we have actually I think you have okay yeah yeah I think we have yeah maybe more than 19.

now you've said at least um is now much much harder to operate a platform on Twitter and have it uh yield any any uh advantage um so uh dramatic Improvement in Bots dramatic Improvement and ability to detect uh sort of trial armies which is a little different that's where you've got say oh you know um a hundred people in a warehouse in a low-wage country Each of which are sitting at a desk with 100 phones so you've got 10 000 actual people and they will then act together to Brigade a particular subject or make something seem very popular when it is not um and we've been able to defeat almost all of them we think very few of them are actually still able to to operate so quality of the system has gotten a lot better okay so um if you said to Linda that

you are going to keep speaking your mind whatever the commercial impact of that and has she agreed to that is she happy with that you aligned uh yeah okay and in her role as CEO does she have any say over moderation or is that under you or you do you do that together well the general principle is that um we were we were here close to the law so for any given country we will try to adhere as closer to the low as possible our law is very between countries and we can't simply fight the law in in um in another country because they will simply cut us off um so the general principle is do whatever we can to enable free and open Communications with between people um provided they're not like I said breaking she's aligned on on that that plan yeah that Focus

um yeah there is an important thing which is like that that obviously doesn't mean that say advertisers should be forced to appear next to any content so we've also developed adjacency controls that ensure that if what you're advertising is um like Disney Disney for example is a big Advertiser it doesn't use advertising a children's movie they you know it won't want the the contents nearby to be sort of family friendly that's totally understandable um so uh so it's it's not like advertisers have to appear next to content that they that they don't agree with and can you um so some people say you're you can't be a little erratic with your Tweeting or at least um tweet a broad range of of content um does anybody say I don't want to be adjacent to Elon Musk

is that is that something that's happened on the on the platform I've never heard that yet um but uh um and did it did that come up with Linda at all sort of what you tweet and and whether that was something that could affect advertisers did she did she ask you about that uh she did it in fact at the conference that we did in Miami so we speech is Paramount fine um I wanted to ask you a little bit about your vision for Twitter as a as a community and as a as a conversation um you've talked about your desire to maximize unregretted time can you just could you explain what that means and how you measure that uh yeah so uh previously Twitter was mostly uh focused on this number called uh they called mdow monetizable daily active users um but the problem

is that when you look closely at that a bunch of those users um never even went to Twitter they would go to um uh there would see a notification on their phone about a tweet but they wouldn't actually click through the site so but what really matters is uh true user seconds of screen time um so that's that's the figure we track right now and that's based on the the screen time as reported To Us by uh uh iOS Android and the browser so um it would be it would have to be the time the amount of time the app is in the foreground right is the most rigorous way to assess this so so when you say unregret it sorry please keep going exactly so in terms of undergraded it's that's a little harder to measure um but we can certainly gather it anecdotally which is to

say that if you spent you know half an hour on Twitter yesterday what percentage of the percentage of that time do you regret um and journaling the feedbackup button has been very positive uh that they they find it's the information to be useful entertaining funny um so we seem to be heading the right direction as far as I can tell I'm certainly open to any critiques from the room well let me let me ask you one on that which is um you know you recently tweeted about George Soros you said um let me get back well let me just get the words because I'm kind of interested in what you think about this he wants to erode the very fabric of civilization Soros hates Humanity that obviously generated a huge amount of response on Twitter on both sides lots of different

viewpoints is that unregrettable time unregretted time that that debate that you created does that fit into that category do you think well I mean I said like Source reminds me of Magneto you know uh well you know you went a little further than that but again without going into the Soros tweet itself you know you're obviously a big figure on Twitter and you're setting a tone and a name so I'm just curious as to whether that sort of debate which which gets triggered is that does that fit into the definition that you're trying to create in that New Town Square well I mean I think the important thing is that like look what I say is not uh is what I'd say um you know it's sort of a Town Square I'm not going to mitigate what I say because that would be inhibiting

for your freedom of speech that doesn't mean you have to agree with what I say in order to mean if somebody says the total opposite that they're that they won't be supported on Twitter they are um the point is to have a diversion set of views and free speech is only relevant um if it's a speech by a speech by someone you don't like who says something you don't like is that allowed if if so you have free speech otherwise you do not and for those who would Advocate censorship I would say it is if if you succeed in that it's only a matter of time before the censorship gets turned on you I agree I mean you can that's your free speech definition which you said but I'm just curious as to on the unregrettable part what what what what type of conversation you're

trying to achieve and whether that's something that is acceptable but maybe not where you want the broader conversation to go well I mean I did clarify that you know some of my concerns about Taurus are that he's funded um a very large number of uh small but influential races around the country especially with District Attorneys um and we find that for example the alien uh and and San Francisco district attorney races uh uh with uh chess boarding and um the guy always I always want to call him Gaston from Beauty and the Beast driving um and and the the it's basically he's he's um of course a large number of the ASP elected who are very easy on crime and will often fuse refuse to uh prosecute so you were basically trying to make a deeper point with that

short yeah um can I just move on quickly to um because I don't want to go too far down that that rabbit hole because that debate has played out on Twitter a bit is um you know are you back near profitability now Twitter is not quite there but we we're we're not like you know when I first acquisition closed I would say it's analogous to being teleported into a plane that's plunging to the ground with its engines on fire the controls don't work um so it's comforting say the least um now we have to do some pretty heavy-handed uh bus Cutting Company healthy but we're at this point we're training towards if we get lucky we might be casual casual positive next month but it remains to be seen and is the Staffing the level you now want it or are you going to

start taking it back up again from this it's gone from I think 8 000 to about 1500 or something like that is that crazy correct yeah um I think there's you know there's um there's definitely we are going to start adding people to the company um and we have started adding some number of people to the company um and um but it's still there's still a lot of change to have to happen so but I I think 1500 is probably a reasonable number and does this show what you can do um in a big tech company in terms of cost reduction I mean when you look around other big tech companies in Silicon Valley um would you say from your experience that there's room for much more significant change at those as well yeah I think Twitter maybe somewhat of an outlier in that um

that there were a lot of people doing things that that didn't seem to have a lot of value um and that's I think that's true probably at most the Silicon Valley companies um maybe not to the degree to which it was a Twitter but uh it's still yeah there's a potential for significant Cuts I think out of the companies without affecting their productivity in fact increasing their productivity um so you know um at any given company there are people who help move things forward and and people who've sort of tried to slam the brakes on and Twitter was in a situation where you'd have a meeting of 10 people you know and one person with an accelerator and nine nine with a set of brakes um so you didn't go very far right um and um so now now we're going home about

releasing functionality even with a little bit of risk to side stability as soon as it's not too serious and I think this point is probably fair to say we were introduced more functionality in the last uh six months than Twitter has in the last six years and in terms of outages there were some outages early on are you are you confident things are stable now well outages Are Not Unusual Instagram recently had an outage for example it was reported on Twitter ironically um so uh we've had outages but not not massive ones and they've generally been brief and limited in scope okay um do you regret buying it you tried to get out of it or are you now happy you bought it well all's well that ends well has it ended well yet or we still got to wait and see um I

think we're on the hopefully on the uh comeback Arc okay so I mean one of the things you have talked about you bought it for 44 billion you've talked about it one day being worth 250 I think in internal meetings can you just talk about how you get there what is the what is the bigger Vision I mean you want to bring back advertisers now and are they coming back by the way yeah yeah um can you give any idea of the scale of the comeback in terms of who you lost and who's coming back uh well I think it'll be very significant um so the advertising agencies this point of all um lifted their warnings on Twitter so appreciate the fact that that group M for example um removed the sort of uh their concern label over Twitter which is a very big deal um and so I

think at this point uh I expect almost all advertisers to attack okay um we've also I've done a lot more to make the advertising uh more relevant to users um so that we show users things that are they're more likely to be interested in buying sounds obvious but right that's what tends to happen yeah not super obvious I mean just just basic stuff like if you do a search on Twitter previously this the search Banner app um that did not take the Search terms into account which is pretty insane um so just show a random ad okay whereas obviously show an ad that is you know matches your search sounds sounds worth doing so just quickly what if you've talked about the sort of single app that does messaging and does finance and other things I mean can you just

enlarge a little bit on sort of how you get there and why America wants that obviously it'll be up to people to decide if they want it um it's like do we make something that is useful enough um that you want to use it more frequently um great that's our goal um so we're not going to do anything to stop people leaving the app or try to track them in the app but let's provide enough compiling functionality that over time uh people's usage of the platform growth so in 10 years time is advertising still going to be dominant on Twitter I think advertising will always play a role um at some point take 10 years from now it may not play the largest role but it will play the largest role for at least a few years to come so I want to do a quick um quick far round

of questions just you know imagine that you're late at night you're sitting there tweeting a few rapid fire responses to stuff um and I'm just going to ask you a few questions and if you can just give me short answers then I want to go on to AR and talk a little bit about some of the deeper points that you you started making earlier um so first one will Twitter be public again in five years I don't know okay um do you think the HQ will still be in San Francisco okay not good so far let's try a couple more um which decade are we going to crack artificial general intelligence I think this one this one okay so it's that soon okay um are you going to take a SpaceX trip yourself I will at some point yeah not sure when but you'll be nice um which is the most

exciting country to build a Tesla plant in right now um well we did make an announcement that Mexico would be our next uh location outside the U.

S uh and picked a site and everything so there's that and then um we'll probably pick another location towards the end of this year is India interesting absolutely um are you still a fan of crypto um well I mean I'm not advising anyone to buy a crypto or bet the farm on you know Dogecoin or anything like that okay on Dogecoin might have been thinking maybe you should but let me advise you that would be a have some ways um okay so it does queer is my is my sort of favorite cryptocurrency because uh it has the best humor and uh has dogs um I did however look at the price of it yesterday it's um it's lower than it was I think well I don't know maybe you know it's like friend of mine has a saying that the most ironic explanation is the most likely and the

most ironic outcome for currency would be that the thing that was made at uh as a joke to make fun of cryptocurrencies uh most ironic outcome would become that it becomes the Global Currency okay we'll wait and see final one um can you rank the US and China on their development of AI each out of ten um I mean the US says very much has the uh most advanced AI so this is you say like like China's close behind certainly and has the resources to scale and to optimize um the the biggest single advances in AI still come from the US and Europe but um all right so it's hard to give an exact number score because it's more like but there's a big gap still there is a there's a gap um that Gap looks like it's on the order of 12 months right fish a narrowing all expanding

it's hard to tell I suspect it will narrow to some degree okay um can you talk a little bit about you've created a new AI company yourself um obviously there's a huge amount of energy and activity in this space or at least it's been talked about I mean what do you want to do yourself in this space beyond Tesla and and the the stuff you talked about earlier what is that new thing well I think there should be a significant third horse in the race here uh we've got open Ai and Microsoft Google deepmind and probably there should be a third horse in the race um so a little bit more on that soon but is it something that will interact with the data of Twitter and the capability of Tesla is it something that tries to bring what you've talked about earlier in

terms of capability together and become that third player is that what you're talking about to some degree I don't want to jump a gun here on announcements but uh um you know the opening AI has a relationship with Microsoft that seems to work very well fairly well so it's possible that um xai and Twitter and Tesla would have something similar possible um you've talked about the importance of Regulation and you call for this this moratorium I mean the history of regulating Tech has been checkered it's been very hard for Regulators to keep up with tech let alone get ahead of it what do you think actually needs to happen that practically could in this space to try to change that because obviously the history of this is not encouraging yeah I mean I think

there should be you know I've been pushing hard for a long time I met with a number of um centers and Congress people in Congress in the white house uh to advocate for AI regulation uh starting with an Insight committee that is formed of independent parties as well as perhaps participants from the leaders in industry and that uh that that oversight committee uh gains um or should say get that Insight committee gains insight into what various companies are up to um and uh you know to the degree that there's okay competitive Dynamics there you can obviously uh you would um sequester board members who are perhaps have conflicts uh but anyway if you figure out some sort of regulatory board and uh and they start off gaining insight and then I have proposed

rulemaking um and then that you know will get comments that are on by industry and uh and then hopefully we have some sort of oversight rules that improve uh safety just as we do with uh aircraft with the FAA and spacecraft and cars with Nitza and Food and Drugs with the Food Drug Administration right and how would that work in such a global thing as we're talking about where Ai and the relative uh Advance between countries is going to be very very important is that something that is is globalizable is that is that well really the key question is uh we'll China uh you know cooperate with the West that remains to be seen um but I would still Advocate like some degree of of oversight I mean we have a regulatory oversight of aircraft for example and yet

the the US is still very much doing great on the aircraft so it makes more effort than the rest than any in any other place so just because you have you know FAA regulations doesn't mean that uh it's necessarily slowed down very much okay um does so so your view would be that the AI changes today lock in the tech Giants the microsofts and the Googles of this world does it also is there also a scenario where it actually helps to bring in new players and and and and and change that dynamic or is that a much more unlikely outcome well that there are um a lot of AI startups the thing that's becoming tricky is that in order you really need three things to compete you need uh Talent talented people um you need um a lot of compute expensive compute and you need

to access the data so whoever's got whoever's succeeding on those three will win um so now that the cost of compute has gotten astronomical so it's now you know kind of sort of minimum ante I would say uh minimum would be 250 million dollars of silver Hardware minimum that's like just two right relevant in any way so the startups are more likely to piggyback off what the others are doing rather than compete directly themselves is what you're saying yeah to train a big one um are you going to take a SpaceX trip yourself to train a model of probably gbt five size I wouldn't be surprised if they use at least 30 30 000 maybe 50 000 h100s which are the latest uh gpus are not sure it's not quite the right word but the latest technology from Nvidia um so and

then you need to run inference as well um so there's a lot of the the G fuser um this point considerably harder to get than drugs actually that's really not a high bar in San Francisco you can tell us more about that later but yes okay okay um so um a couple of things I just wanted to go into on on AI which I love your perspective on um is this gonna what does it mean for society in terms of is this going to embed wealth and power in a very small subset and create a big widening of inequality is it going to democratize and create the opposite what is your sense of of where this heads terms of access to goods and services I think AI will be ushering an age of abundance assuming that we're in a benign AI scenario I think the AI will be able to make goods

and services um very inexpensively and so in anything that is a product or a service where there's not artificial scarcity created such as like I want to live exactly in and this you know neighborhood houses it's like okay well there's only 100 houses there so you know that that would still have scarcity um or a unique artwork would have scarcity but anything that does not have scarcity that we def that we deliberately designed to be scarce will be plentiful for everyone in a benign scenario and in the unbeline scenario well there's a wide range of but what's the thing that you're most worried about when you look at you know when you've been talking for years about the need for regulation what is the scenario that really keeps you up at night well I don't

I don't think the AI is going to try to destroy all Humanity but it might put us under strict controls um I mean and there's no non-zero chance of of it going Terminator it's not zero percent but it's it's I think it's it's a small likelihood of of annihilating humanity but it's not zero we wanted at all going to be zero close to zero as possible and then like I said the uh of AI assuming control or the safety of all the humans and taking over all the Computing systems and weapon systems of Earth and effectively being like some sort of uber nanny but isn't isn't another scenario if you say you say that yeah like what you know like let's say you're uh you know I must well contact consent contestant hypothetically um it's unlikely let's face it um and and

you know you say what what do you want and so I want World Peace um and uh it's like okay well the you know one way to achieve well peace is to take all the weapons away from the humans so they can no longer use them and and to punish any humans that engage in um you know extra territorial activity but isn't the more isn't the more likely nasty outcome that rather than AI taking over and being the ultimate naddy that keeps us all doing stuff that is super safe and it wants us to that actually somebody nefariously harnesses that power to achieve societal control stroke military superiority um and that actually some country around the world decides to use it in a different way uh yes that that's what I mean by like AI uses as a weapon right um and the pen

is mightier than the sword so one of the first places we have to be careful of AI being used is in social media to manipulate public opinion so the reason that uh Twitter is going to a primarily uh subscriber based system is because it is dramatically harder to create it's like or 10 000 times harder uh to create a an account that has a verified phone number from a credible carrier that has a a credit card and that pays a small amount of money per month um and have those credit cards and phone numbers be highly distributed not clustered incredibly difficult um so whereas in the past uh someone could create a million fake accounts for a penny a piece and then manipulates have have something appear to be very very much liked by the public when in fact it

is not or promoted and retweeted when in fact it is not is popularity is is not real essentially gain the system so the device towards uh a is a subscription-based verification I think is is very powerful and that really you won't be able to trust any social media company that does not do this uh because it will simply be overrun with Bots to such an extreme extreme degree so if we take it back to where we started if you look at the election that's coming up how big a role will this big shift in AI capability over the last few months which will obviously continue through the next year how big an impact is this going to play do you think in the messaging and the way that people get told the different pictures of of the candidates I think that's something

we need to go and look at for in a big way is to make sure that this we're minimizing the impact of AI manipulation um we're certainly pretty much taking it Taking that seriously at XX Twitter you know Twitter and um and I think we're putting in place all of the protections to minimize and certainly detect when we see large-scale manipulation of the system okay but beyond Twitter are you worried about this for the election in general yeah um there probably will be attempts to use AI to manipulate the public and some of it will be successful um and if not the selection for sure the next one okay I've got two more questions on AI if you've got the time and then just a little bit on China and Tesla if that's okay um the first thing is um we talk a lot in

terms of AI about the next five to ten years and what the impact is going to be on jobs and some of these things if you look out on a much longer time frame given the speed and scale of the change and you look to your grandkids and great grandkids can you just give us a sense of what what it's going to be like to be human how much is this going to change the fundamental nature of how we operate as as a race at this point I think it's gonna change a lot um especially if you go further out into the future I mean there will be everything will be automatic I mean they'll they'll be household robots that you can fully talk to as though there are people um that can help you around the house or be a companion or whatever the case may be uh there will be humanoid

robots throughout you know factories um and um ours will also be all automatic and anything that that we're intelligence can be applied um even modest intelligence will be automated say like so if you say like 10 20 years from now and we will we be connected to that technology through a neuralink type Divine I mean is that is that where this in your view obviously is that why this heads well A high bandwidth interface from cortex to the so you're sort of computing or AI tertiary layer which already exists uh you know it's just that we don't have a high bandwidth connection um we've got a um limbic system which is a sort of foundational element that's sort of our instincts and desires and whatnot uh they're not cortex on top of that which is a thinking

part of our brain and about a tertiary digital layer which is currently in the form of our phones and computers and laptops and whatnot and all the applications and the the constraint on better a better merging or the construct you know the constraint on on on um having human interests and machine interests be aligned is the bandwidth especially the output so if you select at what speed can you output to a computer it's using Voice or your fingers which will move very slowly so you're talking about maybe 10 bits per second or some some fairly small data rate um so with with the neural link you can increase that by you know increase that by a million probably um so everything just speeds up it's up yeah okay um I mean this is obviously in a relatively

benign scenario because there's a question of not just uh let's say it's a Brian scenario um how do we even appreciate or understand what the computer is doing right um how do we how do we go how do we go along for the ride um and if we have a better if you have a brain machine interface that's I don't know a million times faster than or more like we'll go along for the ride a lot better that then if we're interfacing with a phone using two slow-moving meat sticks um if you put it like that um and in terms of you have a lot of kids many in this room have kids you know what do they need to what skills do they need to have what are the three skills that you think are most important for them that you're trying to give them to be prepared and well positioned

for this new world well I think it's it's important to have a broad range of of understanding in many different subjects so I think general knowledge is important um so you at least have some clue of what you don't know in different areas and then go deep in areas where your child is has a strong interest and ability so finding that that overlap of where is my child what interested in this and has some ability to be successful then uh you know Finding if you can find that Venn diagram overlap then obviously encouraging that is a good thing um and we are obviously headed to a high-tech world so some basic understanding of computers and software and artificial intelligence is probably a good idea okay but the actual broad thrust of um I mean jobs will change

but it'll be more AI enabling and making it better and easier rather than wholesale complete change of the skills you need Which principle what time frame we're talking about here so if you say like over 20 30 year time frame I think things will be transformed beyond belief what you probably think you probably won't recognize Society in 30 years um like I do think we're fairly close you asked me about artificial general intelligence I think we're perhaps only three years maybe six years away from it this this decade um so the impact arguably we are on the Event Horizon of the black hole that is social super intelligence okay so I'm going to ask one final question I'm going to see if you've got two minutes to take a couple of questions from the floor and

it's it comes back to China which you talked about a little bit you have a very big business in China in Tesla um and obviously you know you're on that geopolitical fault line that's getting um potentially interesting it to what extent is this affecting your decision making around sort of how you put assets and and stuff on the ground and how concerned are you about that as a business person and a lot of people in this room have business in China about that getting very very difficult for us well there is fundamentally um an issue that's coming to a head with Taiwan um and it's unclear when exactly Bush will come to show but it seems that there's a good chance push will come to shove it's trending in that direction um I'd rather think what what that would

happen the results would be for the global economy would be absolutely catastrophic um but um you know China has been very clear about its goal on China and uh sort of um including Taiwan um as part of China so one does not need to read between the lines one can simply read the lines they were bracelet and they're not getting and is the biggest concern despite you know the the the the prospect of conflict itself um obviously a lot of the world's high-end chips come out of Taiwan I mean how catastrophic would that be if that was cut off well there's even more that comes out of China so the transit is a lot so much of of the world's heavy lifting on manufacturing especially if the manufacturing is you know simply hard work and say not not particularly glamorous

um charges does an immense amount of hard work that people most people have no idea how much hard work they do so um you kind of from Taiwan was much less less of a concern than being covered from time from China now China would reciprocally suffer of course um because I would say that the economy of the economies of China and Taiwan are they're like conjoined twins with the what the Western economy with the rest of the world so China China the West the rest of the world being Conjuring twins from an economic standpoint will mean that the separation is going to be dire indeed okay that happens I hope it does not happen so um and there's no easy solution here but if there's any if there's any path to diplomatic solution we should really take that seriously

great do you have time to take a few questions from the floor Elon sure does anybody have a question they'd like to pose we'll go here and then here Mike behind you thanks Elon thanks for joining us um I'm the founder of a real estate business in Newcastle in the northeast of England um we export manufacture and Export more cars from the Northeast than the whole of Italy uh would you let me build you a Tesla Factory in the northeast of England thanks for the offer so probably consider uh England uh for a future location of a geiger Factory thank you I'll get you so you will you will consider it are you actively considering it uh we we we we're not currently uh looking at new locations but we will pull towards the end of this year I'll send you some plans

okay um uh David here please thank you for your time uh Fusion lots of scientists say it could change the world Planet it's the sun all live on this planet and on Mars depends on Fusion the sun itself can I ask you why a man of your brilliant brain resources and talent it's not actually focusing on Fusion which I think could be a game changer for society and rather than on Twitter where there are many media decent companies that can do it and I would say it almost in a trivial way well I'm I think we already have a giant Fusion reactor in the sky that called the Sun that shows up every day so um which always said like if you want to know what's standing in front of a future Fusion reactor feels like just go out go and stand in front of the sun you know

just walk outside that's what that's what a giant Fusion reactor feels like because that's what the sun is um it converts about four and a half million towns four and a half million tons of Mass to energy every second and requires no maintenance it's amazing um you don't have to you don't have to refuel it you don't have to maintain it just there so my recommendation for Fusion is solar power and batteries and we can easily power all of Earth with uh just with photovoltaics and batteries not I mean not easily but there's just a very clear path to do so um and and uh no Miracles are quiet just work interesting I've also an advocate of wind and of of nuclear fission uh geothermal hydro and whatnot uh we'll take a couple more one here and then the lady at

the back thank you as you're considering exposure to China and particularly um in the EV space and with the battery supply chain what's your process for evaluating political risk in the near and Midterm I guess I just talked to my team you know read the news uh I don't know assess assess the opinion on Twitter I suppose um there's a very deep analysis you can get on Twitter from people that are World experts on a particular subject so um I don't know I think we just we try to prepare for the worst hope for the best um and um you know make sure we have factories and uh geographically Diversified regions of the world where the supply chain is as localized as possible but this is important uh also for forced Azure situations so if there are earthquakes wildfires

riots revolutions uh ice storms uh heat waves uh you name it uh I think I've seen it all at this point so you want to have you you want to have a supply chain that does not inherit Force Major from all of Earth um because something that's going to happen somewhere it's Big Planet so so that's why I think it's important to have um localized Supply chains with factories in in many geographies uh last question from the lady at the back thank you yeah you famously tweeted that you thought the population collapse was a much bigger risk to humanity than climate change what do you think States families even companies can do to ensure that more of us want to have more children well yeah I mean I think it's it's very telling if you look at the birth rates which

are just you know publicly available um you can look at uh say the birth rate last year for every country um it's available online um and you can look at the trend in both rates and it's just very clear that the trend has been strongly downward um and that we've recently hit all-time lows so um you think if you know during covert you know what else you got to do you might as well have a kid but um it didn't happen actually we had a big drop in birth rate during during covert an increase in divorces too since make a lot of time with their significant other um so I think generally um so it's simply changing people's mind about the goodness of having kids it's like very important to have kids in order to to continue civilization um and I think sometimes

it's viewed as uh you know kids are viewed as an imposition on the world I don't think that's the case at all or that people sometimes think there are too many people in the world that's that's certainly not the case you can you can fit all of the humans on Earth on one floor in the city of New York you know it would be uncomfortable but but just to give you a sense of the cross-sectional area of earth that is human is very tiny it just seems big if you're in a big city um but for the vast majority of the earth if if you're given a task of from front plane of dropping a bowling ball and and and you have to hit someone you'd you'd Miss I almost never hit anyone um so uh what is that you very rarely go over a person um in an aircraft you fly from LA to

New York the vast areas of land with no one at all so anyway the I think we want to just generally have it be socially encouraged to have kids I think certainly companies need to support uh employees that have kids um I think in terms of government incentives uh there should be uh some I think tax breaks for having kids you know or make it just financially not burdensome to have children um and it's always worth bearing in mind like autonomy aside um if someone doesn't have kids what you're actually asking is that uh someone else's kids take care of you when you're old and that that doesn't seem like quite right you know um because because that's that's the world before they'll be forced to do absent automation is is that someone else the kids will have

to take care of you when you're old and um you know so I think anyway one way or another we need to solve this birth rate issue or civilization will go a little to nothing isn't that where AI comes in it'll do all the jobs for us so we can we can handle a uh potentially lower population or what you're talking about I think there will be robot nannies that are very confident so that will help