今年早些时候,中国的发电量超过了美国、欧洲和印度加起来的总和。
这句话在访谈里出现得特别轻,马斯克把它当成一个“事实陈述”抛出来,语气平得像在报天气。可这个数字值得停一下。它意味着全球电力版图上出现了一个新的量级,而电力正是眼下 AI 竞赛最硬的约束条件。
这场对话发生在特斯拉全球工程总部,主持方是 CGTN。话题跨度很大:他对中国的观感、上海工厂为什么能跑起来、Grok 4.7 和 Opus 5.5 的差距、中国 AI 公司的单位算力表现、AI 安全怎么做跨国监管、十亿台人形机器人的时间表、火星、脑机接口,以及他会带一个第一次来中国的美国企业家去看什么。
把这些话按主题拆开看,会发现它们咬合得很紧。制造、算力、电力,三件事最后指向同一个结论:谁能把物理世界的东西造出来,谁就握住了下一轮的入场券。
“上海工厂的魔力,是因为中国人”
主持人问的是,上海超级工厂给特斯拉的全球版图带来了什么优势。马斯克没有先谈政策、成本或者供应链,他先讲了一个跨度三千年的判断。
中国有把制造业做好的固有天赋。这件事不只现在成立,过去两三千年一直成立。
然后是那句被剪成短视频到处传的话:坦白讲,特斯拉上海的魔力,是因为中国人。
说实话,特斯拉上海工厂的魔力,来自中国人。(马斯克)
他接着列举了具体品质:中国团队极有才华、极其勤奋、非常可靠。质量和效率都优秀。工厂本身是块宝石,很漂亮,去过现场的人都会这么说。
这段里还有一个细节容易被忽略。他说他们很在意照顾好工人,确保医保、确保好的食物,让大家乐意来上班,这一点非常重要。
很多外资工厂把中国当成成本洼地,马斯克的说法是能力高地外加组织能力。这两件事的差别很大。前者意味着随时可以被更便宜的地方替代,后者意味着很难复制。
主持人顺着说,看得出来上海工厂的员工工作得很享受。这句对话其实把“魔力”的谜底补全了:高效率和员工状态好,在他那里不是一组矛盾。
那台没有方向盘、没有踏板的车已经在德州跑商业运营
访谈里这段问答有点错位。主持人问 Cybertruck 的市场潜力,马斯克答的是另一台车。
他说的是一台没有方向盘、没有踏板、没有侧后视镜、没有后视镜的车,已经在德州商业化运营,很快会到佛罗里达和内华达,明年年中大概到加州。
语音识别稿大概率把车型弄混了,但这段话透露的信息比 Cybertruck 的销量预测重要得多:无人驾驶出租车在美国已经进入多州商业运营阶段,时间表从“测试”切换到了“运营”。
回到 Cybertruck 为什么长这样,马斯克给的理由挺有意思。他说不希望街道的美学停滞。1950、60、70、80、90 年代,时尚演化都很快,现在该让东西看起来更像未来。他甚至把服装也拉进来了,认为衣着审美也该往前走一步。
车不只是交通工具,也是一座城市每天被看见几百万次的公共视觉。他用的词是“进化”。这个视角放在一个工程师身上有点反常,但特斯拉的产品逻辑里一直有这条线。
AI 快到自己人都头晕
主持人说现在每天都要刷社交媒体看 AI 有什么新进展。马斯克的回应很直接。
AI 软件改进的速度,还有那些重大发布,快得让我头晕。我就在 AI 行业里,它还是让我头晕。(马斯克)
他描述了自己的日常:上床睡觉的时候有个大突破,醒来又有一个,到午饭时间再来一个。
这个说法听着像夸张,但它解释了后面很多判断的底色。一个认为技术正在以“天”为单位迭代的人,对时间表的估计天然会比行业外的人激进。十亿台机器人、两三年解决光刻、明年追上前沿,这些数字都要放在这个尺度里读。
Grok 4.7 不如 Opus 5.5,他把这句话说得很干脆
主持人提到 Grok 4.7 是最近两天的焦点,也提到马斯克在社交平台上说 Grok bot 已经装进特斯拉车里。
马斯克给了一个相当坦率的定位。Grok 4.7 是一台“结实的干活模型”,不如刚发布的 Opus 5.5。
然后他补了一个时间账:xAI 做 AI 只有大约三年,Anthropic 做了六年。所以他预计大概明年某个时候能追上前沿。
这个对比有信息量。模型能力的差距经常被描述成技术路线之争,实际上很大一部分是时间积累之争。三年的团队和六年的团队,差的不是聪明程度,是踩过的坑、攒下的数据管线和工程直觉。
另一个数字更值得记:他们大约每一到两个月发布一个新模型。Grok bot 作为个人数字助理很受欢迎,每月增长约 100%。
月更,加月度翻倍的用户增长,这两个节奏叠在一起,说明 AI 产品的竞争单位已经从“代”变成了“月”。用年来评价一个模型的位置,很快就会失真。
真实世界工程这块空地,目前还没人占住
主持人问了一个很准的问题:Grok 4.7 已经用 SpaceX 的真实工程数据训练,这会不会成为突破 AI 天花板的秘密武器。毕竟其他模型现在都在抢数据。
马斯克的回答先缩了一下:目前只用了一点点。
然后他给出了自己真正的判断。Anthropic 把一件极其重要的事做成了,让 AI 在软件工程上表现出色。但到现在为止,还没有人把 AI 在真实世界工程上做到极好。
Anthropic 做得极好的是让 AI 擅长软件工程。但还没有人让 AI 擅长真实世界的工程。这可能是 Grok 的机会,用 SpaceX 和特斯拉的训练数据。(马斯克)
这句话的分量在于它划出了一块还没被占领的地。
软件工程的语料在 GitHub 上,在开源社区里,谁都能爬。真实世界工程的语料在火箭的失败记录里、在产线的良率数据里、在材料在极端条件下的表现里。这类数据不上网,不外流,也不可复制。
模型会不会解 LeetCode,和模型能不能判断一个阀门在低温下该选什么材质,是两种完全不同的能力。后者的经济价值可能大得多。
单位算力性能,中国可能是全球最好的
主持人问中国模型最近有什么让人印象深刻的突破。马斯克给了一个很高的评价,而且给了限定条件。
中国 AI 模型整体上非常出色。在算力受限的前提下,性能表现很优秀。他判断,按单位算力产出算,中国大概是全世界做得最好的。
然后他给了一个时间表。他公开说过,中国会用比大多数人预期更快的速度解决算力约束。粗略猜测,光刻和芯片制造这块,两到三年。
这个判断背后的逻辑其实不难懂。算力约束是一个工程问题,工程问题有工期。设计、验证、量产、良率爬坡,每一步都有历史参照,能估算。工程问题会让人不舒服,但它不会无限期悬着。
真正难估的是组织效率和资本投入的决心,而这两样恰恰是他在中国反复看到的东西。
AI 的尽头是电,而电的门槛在审批和并网
主持人提到他在 G20 虚拟会议上的发言,关于算力和电力短缺的严重担忧,并请他展开讲讲中国电力基础设施的能力。
马斯克的回答有四层。
第一层:中国今年的发电量超过美国、欧洲、印度之和。
第二层:这个数字还在显著增长。
第三层:如果中国达到和美国相当的人均发电量水平(他认为很可能),那么中国的总发电量会从现在的约三倍美国,变成约四倍美国。
第四层:四倍正好对应两国的人口倍数。
这个推演方式很马斯克。他不谈能源结构、不谈碳排、不谈政策,他只做人均对齐。人均发电量是生活水平的物理刻度,一个社会要生产多少东西、让多少人过什么样的日子,最后都落在千瓦时上。
对做 AI 的人来说,这段值得反复看。GPU 可以设计,可以下单,可以堆机房。电厂的审批、建设、并网,周期是另一套时间尺度,短则三五年,长则十年。芯片的瓶颈可以用钱缓解,电的瓶颈只能用时间缓解。
AI 安全光靠一国监管没意义
主持人问 AI 安全最关键的担忧是什么。马斯克给出的方案很具体:搞一个工作委员会。
理由是他认为安全这件事没法只靠美国监管,也没法只靠中国监管。规则要能公平地适用于任何国家生产的 AI,但真正重要的玩家就是中美两家。
他的原话是,这会是对全人类 AI 安全真正有好处的一件事。
这里有一个隐含前提值得点出来。他谈的“安全”是跨国公共品,而公共品的特征是没人愿意单独买单。两家最大的玩家如果不坐下来对齐规则,任何单边监管都只是把风险推到监管边界之外。
机器人打拳击为什么值得看
访谈中间有个轻松的段落。录制前,团队和 Optimus 玩石头剪刀布,人类赢了前两局,机器人赢了后面所有局。
马斯克的解释是物理层面的:机器人反应时间快。人类受生物条件限制,机器人可以改进执行器、改进传感器、提高相机帧率、增加 AI 算力,反应速度会天然优于人类。
他说他喜欢看机器人比赛。机器人打拳、摔跤、跑步、做体操,挺酷。他还补了一句很有意思的话:如果机器人只会摔倒,那就不好看了。
这句玩笑点出了一件事。观赏性本身就是能力证明。一个能打完整场比赛的人形机器人,背后是平衡控制、力矩控制、实时感知、跌倒恢复这一整套东西。观众觉得“好看”,是因为技术真的到了那个位置。
主持人说美国社交媒体上全在传这些视频。技术传播有时候不靠白皮书,靠一段十几秒的打架视频。
十亿台只是保守估计,二十年后是一千亿台
马斯克重申了他的预测:十年内至少十亿台人形机器人。他补了一句,说这是个很容易做的预测,最晚不超过十年。
然后是那个让人倒吸一口气的复利推演。如果数量大约每年翻倍,那么三到四年就是十倍。十五年是百亿台,二十年是千亿台。
他说这话时的语气是“这只是算术”。
复利式的推演最容易出错的地方是它假设了持续翻倍。但换个角度想,他的意思可能是:只要年增长率维持在接近翻倍的水平,终点就会远远超出直觉。人对指数曲线的估计一向偏差极大,十亿台听着已经很多,二十年后是一千亿台这个数字,才真正说明他的模型里机器人是什么量级的东西。
地球上现在有八十亿人。一千亿台机器人,是人均十几台。
九成概率的好结局长什么样
主持人问得很实在:十年后机器人那么多,普通人怎么在这样的世界里生活。
马斯克先把话说清楚。他只想谈那九成的、繁荣的好结局。剩下那一成的坏结局不能掉以轻心,所以 AI 安全必须盯紧。
然后他描绘了他认为大概率会发生的那一幕。
每个人都有自己的机器人伙伴,有点像《星球大战》里的 C-3PO 和 R2-D2,友善、有用。而且比电影里更好,更有生产力。它可以照顾年迈的父母,可以看护孩子,可以做家教的活,给孩子当一对一的老师,可以做你能想到的任何工作。
再往后是企业形态的变化。会出现这样的公司:一个人,配上几百甚至几千个机器人,有物理的,也有数字的。一个人的产出被极大放大。
他的结论是,实质上的全民高收入会实现。他甚至说,钱在未来是否还有意义,他自己也不确定。
推理链是这样的:如果把商品和服务的产出规模扩大到远超地球人口能够消费的水平,就进入了一个富足时代。这个时代对我们来说很难想象,因为人类整个进化史都是在约束下完成的。食物不够,取暖不够,什么都缺。
他做了一个对比,我觉得是整场访谈里最锋利的一段。
人类大约要花二十年长大,最后二十年退休。中间那段生产性的时间大概只占一半寿命。而且人还要睡觉、吃饭、上厕所。所以大多数人实际能工作的时长,每周大约四十到五十小时。
机器人乐意每周工作一百六十八小时,连续不断。它一开机就是工作年龄,也没有退休这回事。
四十到五十小时对一百六十八小时,这个倍数摆出来,很多关于“机器换人”的抽象讨论就不需要再绕了。
机器人会不会被少数巨头囤起来
主持人的追问非常到位:如果钱不再是交换工具,怎么防止少数科技巨头控制所有资源,普通人怎么分到自己那一份。
马斯克的回答是一个“需求饱和”的论证。
他认为机器人不会被少数人囤积。原因是机器人会饱和人类的需求,它们会做人类能想到的所有事,甚至做得更多。然后它们会无事可做。因为人类能想到的需求在合理范围内是有限的。等到为人类做完了,机器人就会转而为机器人自己做事。
他举了个例子。你想要一座城堡,机器人会给你建一座城堡。
这个论证挺有说服力,但只回答了一半的问题。它论证的是产能过剩会导致价格坍塌,物质极大丰富之后囤积失去意义。它没有回答制度层面的分配:谁拥有这些机器人,谁拥有它们的产出,初始产权怎么定。
他描绘的是一个产能极大化的世界,物理规律保证东西足够多。而从产能到每个人碗里,中间还有一段路要社会自己走。这一段他没有展开。
每次发射扔掉一架中型客机
话题转到太空。主持人问中美在航天领域有哪些合作空间。
马斯克的回答很务实。他认为最重要的合作是卫星发射时的轨道协调,避免碰撞。两家每年往近地轨道送那么多卫星,协调不好就是公共风险。
然后是那个他重复了十几年的核心观点:太空发射需要的根本突破是完全、快速的可重复使用。
他打了一个类比,这个类比很精确。猎鹰 9 能回收助推器和整流罩,但回收不了上面级。等于每一次飞行都扔掉一架中型喷气式客机。如果航空业每次飞完就得把飞机丢掉,航空运输根本不可能成为现实选项。火箭行业长期以来就是这个状态。
他自己都说,这一点他强调过太多次,怎么强调都不为过。
Starship 的时间表也给了:希望下个月底能接住飞船,今年底或者明年初复飞。这是关键突破。他也很坦率地说,别人迟早会解决这个问题,中国在某个时点也会解决。
为什么可复用是所有事情的前提,他讲得最清楚的一次在这里。只有火箭像飞机一样能反复飞,才可能在月球和火星上建起能自我生长的城市。一旦这样的城市建成,意识的预期寿命会大幅提高。鸡蛋不再只放在一个篮子里。月球出问题、火星出问题、地球出问题,人类文明在三个地方都有备份,存活概率就高得多。
两条论证线:人寿保险,和早上想起床的理由
马斯克把他所有的太空论述分成两类。
一类是防御性的。意识分布在多个天体上,文明的预期寿命会显著提高。可以理解为给生命整体买了一份保险。
另一类是激励性的,他说这一条更能驱动他自己。
生活不能只是解决一个又一个悲伤的问题。总得有些事让你早上愿意兴奋地醒来,让你期待未来,让活着这件事值得。(马斯克)
这两条论证的重量不一样。第一条是逻辑,第二条是情绪。他在公开场合讲第一条多,但他说得很清楚,真正推着他往前走的是第二条。
这也解释了他为什么总把“超级酷”挂在嘴边。一个把“让未来变得令人期待”当作核心驱动的人,会本能地用“酷不酷”来评价项目,而不只是“赚不赚钱”。
一百比特每秒,对比一万亿比特每秒
主持人提到他以前列过的五个关注领域:可持续能源、太空、互联网、AI、基因编辑。问他未来十年会不会有新条目。
他给了一个:Neuralink 带来的生物增强。
他解释的角度是带宽。人类有严重的带宽约束。就算在最好的情况下,AI 超级友善,一心一意想让你开心,通信速率也太慢了。
具体数字是这样的。人类输出带宽的峰值大约是每秒一百比特。输入带宽高得多,因为视觉通道强,眼睛本质上就是两台相机,每秒可能有几兆比特。但计算机之间的通信是每秒万亿比特。
他的结论是,任何能提高人类通信速度的东西,都会改善人机对齐。否则,一个每秒万亿比特的 AI 和人类说话,就像人类对着一棵树说话。
这个比喻值得多想两秒。树不笨,也不坏。问题在于它回应得太慢,慢到你无法和它协作。对齐问题经常被描述成价值观冲突,很多时候其实是速率不匹配。你和一台比你快十亿倍的东西共事,最难的可能不是让它认同你,是让它愿意等你把话说完。
给二十岁年轻人的建议:把知识的面摊到最宽
主持人提到马斯克办公室里那块“你可以回到特斯拉上学”的牌子,然后问了一个很现实的问题:二十年后的年轻人,怎么在这么大的转型里找到自己的价值,不被落下。
马斯克的建议是尽可能宽基础的教育。艺术、科学、工程、广博的常识,都要有。
理由非常实用:为了知道该向机器人要什么,你得能提出那个问题。
他顺手把这件事和 prompt engineering 联系起来。你得把自己的愿望准确地传达给机器人。而这一能力的天花板,取决于你的知识面有多宽。你知道的东西越多,你能想到的问题就越多,你能描述的需求就越具体。
他还说了句听起来很狂的话,但他自己也知道它听起来很狂:未来机器人和 AI 会多到这种程度,AI 会急切地想知道你还有什么请求。因为任何请求都会被立刻满足。
这段话如果只读一半,像是某种乐观主义的宣传。读全了其实是另一个意思。当满足需求变得极其廉价,稀缺的东西就变成了需求本身。知道自己想要什么,能把想要的东西讲清楚,这件事会变得比执行能力值钱得多。
带一个美国企业家第一次去中国,他会推荐去哪
访谈最后这个问题很轻,马斯克的回答倒是很具体。
他的建议是能待多久就待多久,中国有太多东西可看。
最低限度:上海和北京。然后他主动加了西安,兵马俑。他自己坐高铁从北京去的西安,说那是很好的体验,火车站非常棒。兵马俑是世界奇迹之一。
还有一个细节,他说得特别认真:记得往窗外看,别玩手机。路上有太多东西值得看。
主持人说他一直鼓励大家亲自去中国看看。马斯克的回应是:
任何我能说的话,都不足以描述中国那种令人震撼的壮丽。(马斯克)
这句话从一个以“第一性原理”和“物理极限”为语言习惯的人嘴里说出来,分量不太一样。他给不出一个具体参数来描述这件事,只能说“自己去看”。
把这些话串起来看
整场访谈里,马斯克没有讲什么宏大叙事,他给的全是数字和工程判断。发电量三倍、两三年解决光刻、一百比特每秒、一百六十八小时、十亿台、一千亿台。
串起来是一条清晰的路径:制造业决定谁能把东西造出来,电力决定 AI 能跑多远,数据决定 AI 能干什么活,机器人决定产出规模,可复用火箭决定这套系统能不能备份。
他反复强调的东西也有共同点。上海工厂的“魔力”是人的品质。中国 AI 的强项是单位算力的效率。真实世界工程数据的价值在于稀缺。这些判断背后是同一个标准:真正难替代的东西,才值得下注。
对读者来说,这场访谈最实用的部分不在那些惊人的预测,而在他给出的一堆可以检验的时间点。两三年后的光刻突破,明年某个时候追上前沿模型,明年年中加州的商业运营,十年内十亿台机器人。这些都能拿来对照。预测的价值不在于它准不准,在于它给你一个可校准的坐标。
- 盯电,别只盯芯片。看一个地区的 AI 潜力,去查它的发电量和新增装机,比看它有多少张卡更靠谱。电力项目的周期以年计,一旦开工就很难加速,这是最硬的约束。
- 在自己的行业里找那块“还没人把 AI 做到极好”的空地。软件工程已经被占住了,真实世界工程还空着。你的行业里如果有大量不上网、不外流的经验数据,那就是你的护城河。
- 把知识面摊宽,练提问。当执行变便宜,描述需求的能力就成了瓶颈。多学一门不相关的学科,比多刷一门本专业的课,长期回报可能更高。
- 关注人形机器人的公开进展,但别只看演示视频。看它能不能连续工作、能不能恢复跌倒、能不能在非结构化环境里完成任务。这些都是可观察的指标。
- 如果你关心 AI 安全,把注意力放在跨国规则上。单边监管的效果有限,中美之间的技术对话才是真正影响结果的地方。
- 有机会就去中国走一趟,别只去北上广深。马斯克推荐的高铁加西安路线值得抄:北京到西安的车上别玩手机,看窗外。
- 对“钱不再重要”这类预测保持兴趣,但别当成计划依据。产能极大化和分配机制是两件事,后者需要社会自己设计。你可以在前者到来之前,先想清楚自己在后者里的位置。
0:00Thank you so much Elon for joining me and also having us here at Tesla Global Engineering Headquarters. So amazing to see kind of new technology innovations all talented engineers here as well. Yes.
0:12In [music] this May you were also on the delegation to China as well. You have met with President Xi. What's your impression about President Xi as a leader?
0:19Well, he seems like a great leader. I mean he's [music]
0:22under his tenure China has prospered immensely. I mean I've seen it firsthand like you you if you visit China and you see the the the amount of progress [music] that happens from one year to the next or even within a year you could it's very visible like you can see the the new buildings, the [music] new infrastructure and and it's also obvious to see that I just walking around the streets [music] the increased prosperity of the average Chinese citizen has increased tremendously. And that's have well that's you know it's happened under [music] President Xi's you know while he's been the leader of China. So so he's clearly done a very you know
0:57done an excellent job and is very capable.
1:00Tesla Gigafactory Shanghai has always done [music] an incredible job. You praise your teams many times. So what kind of advantage or breakthroughs do you think that [music] to making the Shanghai factory a very important position in Tesla's global blueprint and what [music] ultimate kind of confidence do you think this will you know unfold especially in the market?
1:19Well I mean China just has a very strong inherent ability [music] to do manufacturing well. And this has been true not just in present days but for the past two or [music] three thousand years. I think that if you totally frank the magic of Tesla Shanghai is because of Chinese. [laughter] To be totally honest. in China at least right?
1:41Yes we we what we have is the Tesla [music]
1:46China team is just incredibly talented and hard working and trustworthy. and they have the quality is [music] excellent and the efficiency is excellent. So, [music]
1:59the Tesla Shanghai factory is a is a gem, you know, it's it's and it's beautiful it's beautiful actually if anyone's ever been there. [music]
2:06and we also care a lot about taking care of the workers and making sure they, you know, they get health care and good food and enjoy coming [music] to work. It's very important. So,
2:16Absolutely. I can I I I think we can tell like, "Oh, your [music]
2:19stuff in Shanghai factory they're very enjoying the working there." So, Talking about the next hot topic recently, we probably we can call September is the Cybertruck month. Have you seen the videos like all the people gather around that? I guess it's it's it goes very very quickly.
2:35Yeah. I mean, it's pretty cool cuz we designed the car to look futuristic. Yeah.
2:39Cuz I think you know, from a fashion standpoint, we don't we don't want as the the aesthetics of the street to stay stagnant. You want to evolve the what what do cars look like? And also I think maybe even clothing. Like we should be Like there was there was rapid fashion evolution in the '50s, '60s, '70s, '80s, '90s and probably time to make things look more futuristic. Yeah.
3:07And then how [music] you see the kind of market potential for Cybertruck, you know, it's it's it's like, you know, it's really future futuristic, but how long it can really come to the reality or real life?
3:18Well, we have the Cybertruck which has as you know, has no steering wheel or pedals or side mirrors or rearview mirror [music] operating commercially in Texas right now. And it will soon be operating commercially in [music]
3:33Florida and Nevada and a number of other states. And probably by middle of next year it'll be operating commercially in in California.
3:41Absolutely. I think all the changes is happening really fast. You know, it's just like everyday we have to to look social to see the kind of improving of workflows about AI or other tech topics. So,
3:53Yeah, I mean AI is just so crazy. It's I mean AI the rate of AI software improvements I saw and and major announcements is so rapid that [music] it's it's makes my head spin. You know, and I'm in, you know, I'm in AI and it still makes my head spin.
4:07this industry but you still have to
4:08It's still shocking. [music] Yeah.
4:09you know, sometimes I'll like when I go to bed there'll be some big AI breakthrough and then when I wake up there's an AI breakthrough and by the time [music] it gets to lunch there's another AI breakthrough. Exactly. Yeah.
4:19So that's why even during this break you have to check some other that is how it is. Yeah. So talking about like AI models Tesla have the latest what version, right? The latest release about the Grok 4.7. [music] Yes.
4:32I think this is your most focused these two days. I I just saw your post on social saying that you know, the Grok bot is already in your Tesla. Tell us like what's the reactions what's the latest you have [music] received from?
4:45Yeah, I mean well we just we were we we released a new model, you know, roughly every month or two. Yeah. so [music]
4:53and it it's a cool Grok 4.7 is a solid workhorse of a model. [music] it's not as good it's not as good as say Opus 5.5 that just got released but [music]
5:06also we're we've only been around for about 3 years doing AI [music]
5:11whereas Anthropic's been around for 6 years. So so I I would expect that we probably catch up to Frontier sometime next [music] next year. Most likely. In the meantime I think Grok is very useful for many things. and Grok bot that's kind of like a personal digital assistant [music] has been very popular and is growing
5:40about 100% a month. So it's growing very rapidly.
5:44I I you are very humble in saying that like Grok 4.7 is still not the perfect one, but we know that although have experienced a little bit delay, but you know that people are already very fascinated or interested to know because Grok 4.7 already been trained by SpaceX real world engineering data. So, do you think
6:03Only a little bit so far.
6:04A little bit, okay. But so, do you think that this will be one of your secret weapon to kind of break through the AI ceiling cuz currently you know other models are have the challenge for data.
6:13Yes. [music] Well, I think we've got a real opportunity with the SpaceX training data and maybe with the Tesla training [music] data to make AI extremely capable in real world engineering. So, what Anthropic did extremely well was make AI excellent at software engineering. but [music] no one has yet made AI excellent at real world engineering.
6:37And I think that's an opportunity for Grok using the training data from SpaceX and Tesla [music] to be excellent at real world engineering.
6:45You mentioned many times like previously Chinese models also doing a fascinating job. So, currently what's the most impressive breakthroughs for Chinese models in your understanding?
6:55Well, they've they've been the Chinese AI models are generally outstanding actually. I mean, you can see on the sort of [music]
7:04intelligence rankings that especially given the limited amount of compute that [music] the Chinese AI companies have, the the performance is excellent. I mean, probably China's doing by far the best in terms of performance per unit of compute. And as I've said publicly you know, I think [music] China's going to address the compute limitation in I think faster than most people expect.
7:30my rough guess would be that China addresses the compute constraints with lithography and chip making in about [music] two or three years.
7:37Two or three years. We will really expecting to this time. So, what are the challenges we're facing for AGI in early September, you know, the G20 virtual session, you have raised a very serious concern about the compute and electricity shortages. You mentioned that probably China have way more capability about electricity infrastructure. Yes.
7:58And can you elaborate a little bit more about this gap and you know, what areas do you think the Chinese industry and capability can help to get a new breakthrough for the specifically on the electricity infrastructure for the whole world?
8:12I was just a statement of fact that China I think is a China like earlier this year, China passed the electricity output of the United States, Europe, [music] and India combined. Wow. Yeah. Exactly.
8:23And and continuing to increase [music] significantly. so if if China reaches the same GDP [music]
8:32electricity output of the US, which I think is likely, then China would would go from where it is now roughly three three times the US electricity production to probably four times, which would be proportionate to the population difference.
8:44Talking about the topic like AI safety, what kind of crucial concerns does do you expect? [music]
8:49Yeah, I think maybe there could be a working committee to figure out safety because it's I think in order for [music]
8:57safety to work, it's it can't just be regulation in the US or regulation in China. It's got [music] to be we you know, how do we how how do we apply apply regulation fairly to AI produced in any country, but really China and the US are the the the ones that really [music] matter here.
9:15so I think a really important [music] thing that would be good for the safety of AI for all of humanity.
9:22Yeah. So, let's talk about the human right robotics cuz I think this is also a kind of industrial field US-China can still also working together leading the world. And also I can give you one example for the Optimus. It's really amazing, really impressive. So, before interview, we have a a kind of a interactions. We play rock, paper, scissors with Optimus. Oh, yeah. Yeah.
9:43Yeah, we just [clears throat] won the first two rounds, and Optimus won all the remaining. That's really impressive.
9:48Well, robots have a very fast reaction time. So, it's going to be very difficult to beat a robot for reaction time. cuz we are we are somewhat constrained by our biology. But a robot, you can just improve the actuators, improve the sensors, improve the the frame rate of the cameras, in- increase the image amount of AI computer is, and it's going to just fundamentally be much better and have faster reactions than humans.
10:15I enjoyed watching the the robot games. [laughter]
10:18I mean, everyone in America was fascinated by that, by the way. It was all over US social media.
10:22Yeah, you said it is very entertaining to watch, right? Yeah, but other than that
10:26Well, but it's it's entertaining to watch, but it also shows you the progress of humanoid robotics. Exactly.
10:32Like, it wouldn't be entertaining to watch if the robots just fell over. Of course.
10:36so, but you know, the robots are like boxing, and then like Wrestling, yeah.
10:39wrestling, running, doing gymnastics. it's pretty cool.
10:44So, I mean, the future's going to have a lot of robots. Yeah. Like, a lot a lot.
10:48Yeah. You mentioned that in just in another 10 years, right? There will be Yeah.
10:52the population for humanoid robots will surpass 1 billion. Yes. So, that's That's
10:56that's an easy prediction to make.
10:58Easy prediction. Maybe within 10 years.
11:00I'd say no later than 10 years at the latest.
11:02yeah. So, what Let's talk about what's the real life will look like in 10 years, you know, when there are so big you know, number of population of humanoid robots. And also, you mentioned that one humanoid robots will have the five times of production productivity for human. Pretty much.
11:18Yeah. So, what the life will be? And then, I mean, it's specifically like like ordinary people, how we can cope with the the living, you know, with all AI kind of dominated world.
11:29Well, I I my So, I let me focus on the [music]
11:33what's the 90% likely outcome, which is the prosperous, the good outcome. Mhm.
11:37We don't want to get complacent regarding the 10% but may probably a bad outcome. So, we need to it's important for us to [music]
11:44pay close attention to AI safety. But let me tell you about the 90% positive outcome. The 90% positive outcome is that everyone's got robots that are their their buddies. and if you think of like Star Wars where you had C-3PO and R2-D2, and they were friendly, helpful helpful robots. [music]
12:04well, wouldn't it be great if everyone could have their own personal C-3PO R2-D2? Mhm.
12:09But even better than that.
12:10Yep. Like more productive. Mhm.
12:12that could take care of elderly parents or be a good guardian and watch over the children or or be a teacher, like a individualized tutor to your child. Hm. [music]
12:25and and taking care of any jobs that that that you can think of. So, that would be pretty cool. Exactly.
12:33everyone would like like to have a future where that they can have their own personal robots. And And then I think you'll see many companies where it's one person with hundreds, maybe thousands of robots, but like physical robots as well as digital robots. so, the productivity of one person would be greatly amplified. [music]
12:54and and my prediction is that is that there will be effectively universal high income. Mhm.
13:00in fact, I it's not clear to me that money will even matter in the future. like if you if you scale the economy the output of goods and services to a level much greater than could possibly be consumed by the number of humans on the planet, then you you have an age of abundance that is that is difficult for us to wrap our minds around because our entire evolutionary history has been an evolution of constraints of of
13:29where we didn't have enough food, or we didn't have enough you know, heat, or you know, so, that's certainly going to be true if the output of goods and services is constrained by the number of humans. Because you know, humans take about 20 years to grow up, and then they are generally retired for the last 20 years of their life. Yeah. So, there's about 40 years, maybe half the life of a human that's productive, [music] but then humans also need to sleep, and they need to do the eat and go to the restroom, you know, all that basic stuff. [music]
14:00and and and so, the actual number of hours that that most humans can work is maybe around 40 to 50 hours a week, something like that. whereas the robot will be happy to work 168 hours a week. Continuous. and this and and the robot is instantly at working age. And does not have retirement. Exactly. Yeah.
14:24So, you can draw us a very beautiful picture about how the life, or how highly efficiency about the productivity by the humanoid robots. But, for example, you just mentioned that maybe money will not be that kind of concern at that time. If Yeah, if money will not be a for example, like a tool for exchange, in that kind of world, do you how do you think that we can try to prevent you know, the only a handful few tech giants [music] in you know, controlling the all the resources? And is specifically, how ordinary people do you think they can claim their share in this kind of full [music]
14:56abundance? Yeah, what I'm trying to say is that I think that there will be a level of abundance that is more than anyone could possibly consume. I mean, I think it's going to get pretty wild, like where if you want a if [music] if you want a castle, you know, the robot will build you a castle.
15:16so, it's not it's not like it's it's not as though like the robots will be hoarded by a small number of people. It like actually the robots I think will be they will saturate on human demand, meaning that they will [music] be they will do more than anything humans could possibly think of. Mhm.
15:34and and [music] and and they'll run out of things to do for humans. Yeah.
15:40because there is a finite amount of things you could do for humans [music]
15:43within reason. so I think the robots will actually run out of things to do [music] for humans and then they will do things for themselves because they they've run out of things to do with for humans. So and and probably probably there will be I mean I I my prediction is like as I talked about at least a billion humanoid robots within 10 years. Probably less than 10 years. [music]
16:08I think that continues to grow and probably if if if you if you if you assume that things roughly double per year then then that's about every three or four years [music] you have a 10x increase. Yes.
16:23So you know, maybe that would mean say in 15 years there are 10 billion robots.
16:3010 billion just [music] 15 years. In 20 years 100 billion.
16:34Exactly. So, this that's why we we always call it like AI world is changing improving exponentially. Continue about the cooperation between US China or what areas we can do together. The space exploration. cuz currently many Chinese commercial space companies they are strive to you know, to explore more use after the successful land recovery. what kind of areas or avenues do you think possible avenue do you think between communications or corporations can can get between US and China?
17:06Well, I mean in space I think the most important thing to achieve cooperation with is to make sure that when satellites are launched from the US or China that we we closely coordinate where what orbits the satellites are going to to avoid a potential collision. I think that's the most important area of cooperation.
17:23and then in in in general I would I would say that the important thing for space launch, the fundamental breakthrough that's needed, is full and rapid reusability. which we're hoping to achieve with the Starship design. [music]
17:38by catching the the ship hopefully around the end of next month and then re-flying the ship either late this year or early next year. and that would be a very important breakthrough and I'm sure others will solve that that problem as well and I'm pretty sure China will at will at some point solve that problem, too. [music]
17:55but it's it's the full and rapid reusability of a rocket is the fundamental breakthrough needed to make life multi-planetary.
18:02Continue to talk about multi-planetary species. You always You have been always believe that human should become a multi-planetary species. So, tell us what's the current crucial step for your Mars mission?
18:13Well, like I said, the the most profound and and fundamental in invention necessary for making life multi-planetary is rapidly and completely reusable rockets. Mhm.
18:23otherwise the costs are impossible. you know, with with our Falcon 9 rocket, we [music]
18:30we recover the booster stage and the fairing, but not the upper stage. And this is like throwing away a medium-sized jet on every flight. and you can imagine if in air flight, if you had to throw away the planes after every flight, you're it would be far too expensive to and and and air flight would not be a realistic option. Well, that's been true in in rockets for [music] for a very long time.
18:53So, I can't emphasize enough and I've said this so many times over the years. [music]
18:58full and immediate reusability of of rockets, like aircraft, [music]
19:03is the is the fundamental breakthrough that's needed to to extend life beyond Earth, to make life multi-planetary and to [music]
19:11create self-growing cities on the moon and Mars. Mhm.
19:16And and if we're the as soon as we're able to create a self-growing city [music]
19:19on the moon and Mars, the probable lifespan of consciousness is dramatically greater. because now [music]
19:26we don't have all our eggs in one basket. so if something were to something wrong were to go wrong with say the moon or or Mars or Earth, but we've got human civilizations on all three of those places, then our chances of surviving are much greater.
19:42Of course. So pretty different planets. So that's why we all have to do our part. Follow
19:48And also it's just super cool.
19:49Exactly. We have to follow all the tech giants' steps, try to make our own self, try to be the become multi-planetary species.
19:58Yes, exactly. So I mean I classify these generally as like these two lines of main lines of argument. One is the defensive line, which is that it that having having consciousness on [music]
20:07multiple heavenly bodies greatly improves the probable lifespan of civilization. that's the defensive argument. It's You can think of it as life insurance for life as a whole. the but the one I find personally [music] that drives me more is actually the the inspiration argument.
20:25because you see that that like life cannot [music] just be about solving one sad problem after another. There must also be things that make [music] make you excited to wake up in the morning. Exactly.
20:36you look forward to the future. Mhm. [music]
20:40that make life worth living. And being a space-faring civilization is one of those things.
20:45I totally agree with that. So you know, you had you once list like five areas you care about most. Like sustainable energy, space, and also the internet, AI, and gene editing. Looking ahead for another 10 years, are there any new frontiers you want to that list?
21:04Well, I I think biological enhancement with like Neuralink, this company I created with some other people, where where we we want to try to solve the bandwidth constraint that humans have. so, so so cuz you can imagine like even in the most benign scenario where the AI is super friendly and wants to do everything you can wants to make you happy, just our rate of communication is too slow.
21:35So, our like average the peak human bandwidth is only about 100 bits per second, roughly. and I our input bandwidth is much higher because of vision. So, there we maybe have it's it's a debatable point, but I'm you know, a megabit per second or few a few megabits per second from our eyes, which are effectively like cameras.
22:00But our output bandwidth is really low, it maybe 100 bits per second, whereas computers are communicating at trillion bits per second. So, a- anything [music] we can do to increase the communication speed of humans will improve the I think the alignment of [music] humans and machines.
22:19cuz otherwise the the the to an AI that is communicating at at a terabit a second, [music] talking to a human will be like talking to a tree. Exactly.
22:31you talk to a tree, it's it's takes a long time to respond.
22:34Let's talk about the challenge for young people. You You mentioned [music] that You also talk about education like in the AI world. And also I I watched that in your headquarters here, you have a sign like you can come back to school in Tesla as well. Means which means [music] that in every age you can have the opportunity to come back to to learn always to follow the latest kind [music] of bricks breakthrough in AI world. But what what about kind of suggest suggestions or your predictions? At that time, how for example like a 20-year-old young people [music] they can get their fulfill their self-worth rather than being left left behind, you know, during this massive transition?
23:12My My recommendation for education [music] would be to get as broad-based an education as possible. And that includes arts and sciences, engineering, wide general knowledge. Like in order to [music]
23:26know what to ask the robots for, you need to be able to formulate the question. And and with [music] the broadest possible general knowledge, you'll be better able to formulate the questions or what they call prompt engineering, I guess. [music]
23:39You got to convey your what your wishes to the robot. And as crazy as it sounds, we will run out the There will be so many robots and so much AI that the AI will actually be eager to hear any request that you could possibly have. Mhm.
23:56because that request will be fulfilled immediately.
23:58Yeah. Your last visit to China is in May. If you you got to have one opportunity to bring one American entrepreneur for visiting China for the very first time, what will be the most experience you want him to have first?
24:11Well, I I'd generally recommend I mean, as much time as as as that person can devote to China, they should go because China has got so many things to see. so, I don't mean [music] to I mean, but like usually people don't have They have They don't have infinite time, so at a minimum they they should see [music] obviously Shanghai and Beijing. I recommend Xi'an, the Terracotta Warriors. I took the bullet train from Beijing to Xi'an. That was a great experience. Train stations are amazing.
24:40[music] the Terracotta Warriors is one of the wonders of the world. Mhm. and [music]
24:47and just and just make sure you look out the window and and you're not on your phone [laughter]
24:51cuz there's so much to see.
24:52Yeah, like you always recommend everyone to experience [music] to go and see China by themselves. Yes.
24:57Because any words I I any words I say do not do justice to the incredible majesty that is China. [music] It's awesome.
25:06Really appreciate for your words. Thank you so much Elon for sitting down with CCTV Business Channel. And Elon, we truly hope to see you again very soon in Beijing. Thank you.
资源与出处
- 视频在 YouTube 收看
- 档案silicon-valley.tech 档案页本页视频与文稿数据来自 Elon Musk Archive
- 精读本篇精读 Markdown 版