播客·中文精读

播客特辑 2:马斯克在沙特经济论坛受访

Podcast Special 2 Elon Musk interview at Saudi Economic Forum

中文精读由 AI 基于原始视频与文稿改写,非官方翻译,可能存在偏差或错误;表述以原视频为准。说明

80 亿人对数百亿台机器人。这个比例出现在沙特阿拉伯的一场经济论坛上。

马斯克的表述是,未来每个人都会有自己的 C3PO 或者 R2-D2,而且"可能还更强"。数量级他给得很具体:数百亿台个人机器人,比地球上的人多出好几倍。

说这话的场合不是特斯拉的发布会,是美沙投资论坛期间的一场对话。特朗普和王室成员都在场,台下的听众是主权基金、部长和大企业负责人。议题是沙特怎么把增长引擎从石油换成创新。

他带来的不只是 PPT。几台 Optimus(特斯拉的人形机器人)被带到现场给政要们表演,其中一段动作被现场的人戏称为"特朗普舞",后面还来了一段 YMCA。人形机器人会跳舞早就不新鲜,新鲜的是它出现在这种级别的场合,被当成外交道具用。

这场对话谈到的内容,从机器人数量、收入分配,到自动驾驶、卫星网络、地下隧道,最后落到人工智能该追求什么。跨度很大,但底层逻辑是一条线。

数百亿台机器人,这个数字意味着什么

先把这个量级摊开看。全球汽车保有量大约 15 亿辆,这是人类造过最复杂的量产工业品,花了一百多年才堆到这个数。数百亿台机器人是它的几十倍。

马斯克没有给时间表。但从他过去的表述习惯看,他指的是一个相当长周期之后的稳态,不是十年内的事情。真正有价值的信息不在数字本身,而在他把这个数字和什么东西放在一起讲。

他放的是 C3PO 和 R2-D2。这两个角色在《星球大战》里不是工具,是角色,有性格、会抱怨、会害怕。这个类比的潜台词是:这些机器不待在工厂里,它们在你家里,和孩子说话,记得你上次让它做什么。

另一层意思是价格。C3PO 是"属于某个人"的,不是公共设施。要让人人都有一台,成本必须降到家电那个区间。人形机器人现在的造价还在汽车的量级,要往下走两三个数量级,靠的是产量和供应链,而这恰好是特斯拉最擅长的事情。

这里有个反直觉的地方值得说清楚。从纯效率角度,人形并不是最优解:轮子加机械臂更稳、更便宜、更容易控制。坚持做人形的理由在环境适配。人类世界是为人体的尺寸、步态、手臂长度和抓握方式建造的,楼梯、门把手、电钻、货架,全是按这套标准设计的。做成人的形状,机器人就能直接使用这套已经存在的基础设施,不用改造整个世界。

这也是为什么马斯克提到"家电化"。家电的特点是买回来不用看说明书,坏了有人修,用久了不觉得稀奇。人形机器人要走到这一步,中间隔着的不只是硬件成本,还有可靠性、安全标准、维修网络和保险体系。

他把"基本收入"改成了"高收入"

机器人数量讲完,话题自然滑到了经济后果。马斯克在这里用了一个很讲究的词:universal high income,普适高收入。

如果机器人带来的生产力提升足够大,财富的创造方式会发生根本变化,那就不只是一张安全网,而是普遍的富裕。(马斯克在论坛上的表述,现场转述)

注意他改的那个词。universal basic income 是普适基本收入,讨论了很多年,逻辑是"机器取代工作,所以给每个人发钱保底"。basic 这个词自带一个预设:够活,但就那样。

他换成了 high。这个改动不是修辞游戏。基本收入的资金来源是税收,本质是再分配,蛋糕大小不变,切法变了。高收入的逻辑是蛋糕本身变大了,因为生产端的人均产出被抬高了几十倍,商品和服务的价格被压到很低,同样的名义收入能买到的东西多得多。

他配套给的数字是,全球经济规模可能做到今天的 10 倍。现在全球 GDP 大概 110 万亿美元,十倍就是 1100 万亿这个量级。这个数字是否成立,取决于一个前提:机器人造出来的东西有人买。如果收入集中在资本一端,产能再大也卖不出去。所以"普适高收入"在他这套叙事里,既是愿景,也是让十倍 GDP 得以闭环的必要条件。

这套推论有一个明显的空档:过渡期怎么办。从今天的就业结构走到那个稳态,中间必然有一段技能错配、岗位消失、财富分配的剧烈摩擦期。他在论坛上没有展开这一段。而这一段,恰恰是政策制定者最该问的。

他给乐观派科幻指了一本具体的书

马斯克在谈 AI 未来时,点了一本书:伊恩·班克斯(Iain M. Banks)的《文化》系列(Culture series)。

这个系列在中国的知名度不高,但在科技圈里是硬通货。它的设定是这样的:一个由超级人工智能"心智"(Minds)管理的星际文明,物质极度充裕,人类几乎不需要为生存工作,可以自由选择把生命花在艺术、游戏、探险或者纯粹的闲逛上。那些 AI 不是统治者,更像是一群极其聪明又有幽默感的监护人。

马斯克反复推荐这套书,用意很清楚。公众讨论 AI 未来时的默认画面是《终结者》:机器觉醒,人类被清算。他需要一个反例,一个"技术高度发达但人类活得挺好"的模板。《文化》提供的正是这个:AI 强到近乎全能,同时人类没有被圈养,也没有被消灭,文明的重心从生存竞争转向了别的东西。

选一本小说当政策参照,听着不严肃。但这件事本身说明了他对叙事权的判断:大众对 AI 的恐惧,很大一部分来自文化产品提供的想象材料。要改变预期,得先给出一批替代性的画面。这也是他后来反复说"我是亲人类的 AI 派"的原因。

终结者的风险他认,但方向盘朝着星际迷航

他没有回避危险。论坛上他直接提到了"终结者剧本"这个词,承认这条路存在真实的坏结局。

但他把目标定在了另一个坐标上:星际迷航式的未来。探索、繁荣、技术服务于人的进步。这套说法里,AI 的角色是飞船的引擎,不是飞船的敌人。

这个处理方式值得琢磨。工程上有个常识:风险要管,但组织不能靠躲避风险来运转,得朝着一个明确的目标推进,在推进过程中处理风险。把所有精力放在"防止坏结果"上,最后往往连好结果也一起防掉了。他把风险当成边界条件,把目标当成方向,这套位次关系在很多技术判断上都能看到。

同样的话他在别的场合讲过很多次:AI 出事概率不为零,所以要参与,要把它做对,站在旁边喊停不解决问题。这一段和后面 xAI 的定位是连着的。

车不再是消费品,是四个轮子的印钞机

话题转到交通的时候,沙特方面的热情明显上来了。机器人出租车和自动驾驶在那边被讨论得很热。

他们用的框架很有意思:把车理解成"四个轮子的机器人",而且是能自己挣钱的机器人。

这个定义变化的杀伤力比多数人意识到的要大。今天你买一辆车,钱花出去就变成了折旧资产,一下地掉价,然后每天再往里贴保险、停车、电费。全世界私家车的平均使用率是个位数,剩下 90% 以上的时间它趴在车位上,一动不动。

如果车能自己出去跑活儿,这 90% 的闲置时间就变成了现金流。买车的人从消费者变成资产持有者,车从耐用消费品变成生产资料,价格逻辑、融资方式、保险产品全都要跟着重写。

沙特对这套叙事接受得快,有它的现实原因。人口结构年轻,大城市集中,公共交通体系还在建设期,没有那么多存量出租车牌照和既得利益需要摆平。更要紧的是财政。油价和财政平衡之间的绑定正在松动,主权基金需要一个能持续产生回报的产业叙事,自动驾驶是现成的一个。

这套逻辑能不能走通,卡点不在车,在监管。无人车要拿到商业运营许可,涉及事故责任怎么判、保险怎么定、城市道路数据归谁。沙特在这类事情上的决策链条比多数国家短,这也是马斯克在那边说话格外顺的原因之一。

船和飞机拿到许可,卫星补上最后一块盲区

论坛上还提到一件事:Starlink(星链)在海事和航空使用上拿到了批准。

这件事听起来像技术细节,实际影响很直接。陆地上的通信有光纤、有基站,铺起来贵但铺得动。船和飞机不一样,它们在海面和万米高空移动,光纤根本到不了,基站覆盖不住。

这个市场长期由少数地球静止轨道卫星运营商把持。静止轨道离地面三万六千公里,信号来回一趟延迟就有半秒,价格还贵。低轨星座把卫星放在几百到一千多公里的高度,延迟压到几十毫秒,能支撑视频通话和实时数据传输。

对沙特来说,这个许可的实际意义在红海和波斯湾。那是全球最繁忙的航运走廊之一,大量船只经过,船员的通信、船舶的定位和货物追踪都需要稳定的链路。航空那一头对应的是利雅得和吉达的枢纽地位,长途航班的旅客上网,一直是航空公司想收钱又收不好的业务。

连接性这件事情常被低估。它不像机器人那样有画面感,但所有其他技术都跑在它上面。自动驾驶需要数据,物流需要数据,金融和旅游更需要数据。先铺路,再跑车,这个顺序绕不过去。

挖隧道这事儿,马斯克管它叫虫洞

Boring Company(隧道挖掘公司)也在话题里。马斯克用的说法是"虫洞效应"。

当然不是真的时空捷径。他的意思是从地表往下走,绕开地面的拥堵,把两点之间的通行时间压缩到接近直线距离。城市交通的浪费大部分发生在平面上:红绿灯、并线、路口,所有车挤在同一个二维平面里互相等待。往地下走一层,这个约束就消失了。

他们有一个能拿出来的样板:拉斯维加斯会展中心的地下环线。展会期间人流集中,地面完全堵死,把车放进地下单行道,接驳效率确实提上去了。

这套方案一直有争议,得说清楚。从运力看,隧道里的车一次载几个人,跟地铁一节车厢装一百多人完全不在一个量级。城市规划圈的主流意见是,真正解决拥堵要靠大容量公共交通。

但把它放在特定场景里看,逻辑是成立的:会展、机场接驳、园区通勤,这类短途、高频、点对点的需求,隧道的性价比确实比挖地铁高,因为直径小、不用征地、施工快。它解决的是一个细分问题,不是一个通用问题。分清这两件事,才对得上他说的那句话。

xAI 的 KPI 不是 AGI,是求真

整场对话里最抽象的一段,是讲 xAI。

多数人以为这家公司的目标是造出通用人工智能(AGI,artificial general intelligence),把能力做到最强。马斯克在论坛上的说法不太一样:目标是造出"最大程度求真"的 AI。

根本使命是理解宇宙。(马斯克在论坛上的表述,现场转述)

他给的具体问题很硬:宇宙从哪里来,要往哪里去,还有哪些问题是我们压根没想到要问的。最后这句最耐人寻味,它指向的是未知的未知,是连提问的框架都还没建立起来的领域。

这套话听着像哲学系的开场白,实际是产品定位。市面上主流大模型的训练路径是强化学习加人类反馈,模型被调教成"人类满意的样子"。这个目标有个副作用:人类满意的回答,和真实的回答不总是一回事。有时候人想听的是安慰,是和自己已有立场一致的结论,是有个明确答案而不是"我不确定"。

把优化目标换成"求真",等于把讨好用户这件事往后放。这在商业上是有代价的:一个经常说"这个我不确定""你的前提可能有问题"的助手,体验不如一个顺着你说的助手舒服。

这条路也有它自己的难题。谁来定义"真"?训练数据里的大量内容是人的观点、习惯、偏见,模型从里面学出来的东西本身就是有偏的。求真作为一个方向可以立,作为一个可测量的指标,还差得远。

沙特为什么愿意听这些

这场对话的另外一半信息,在听众身上。

长期跟踪沙特创新生态的研究者凯文·布雷克(Kevin Brake)对这类论坛的分析整体是积极基调。他的判断是,外界能从这些讨论里"提前看到沙特创新的下一步"。

他关注的重点不是某一件产品,是这些东西组合起来的效应:机器人、自动驾驶、卫星连接,三样凑在一起,指向的是生产力的一次跳升,和一个真正由创新驱动的经济结构。

沙特在这个位置上有一些别人没有的条件。主权基金有钱,可以直接买技术和买合作,不用等本土企业慢慢长大。人口年轻,对新事物的接受度高。产业结构的包袱轻,没有大量需要保护的传统岗位和旧基础设施,某些阶段可以整块跳过去。

短板也很清楚。钱能买设备,买不到生态。硬件的供应链、工程师的存量、监管细则的成熟度、大学和产业之间的转化通道,这些都要十年二十年去长。论坛上讲的是蓝图,蓝图和图纸之间还隔着大量枯燥的工作。

把这场对话放进这个背景里看,马斯克讲的每一个板块,都能对上沙特当下需要的东西:机器人和 AI 对应产业升级,Robotaxi 对应城市交通和新兴产业,Starlink 对应连接性基础设施,隧道对应城市改造。他讲的是一套完整的供给方案,对方要的正是一套完整的需求清单。这是这场对话能谈得这么顺手的真正原因。

值得带走的几件事

  • 看人形机器人的进展,别看跳舞视频,看成本曲线和故障间隔。跳舞是演示,可靠性才是产品。哪天有厂商敢公布"连续运行多少小时无人工干预",那才是真正的节点。
  • Robotaxi 的真正门槛是法规,不是技术。关注你所在城市有没有开放无人化商业运营的许可,比关注某家公司的演示视频有用得多。
  • "普适高收入"这类说法,追问三个问题:钱从哪来,怎么征,过渡期多长。愿景可以听,路径要问。
  • 卫星互联网的资费变化值得盯着,尤其是船用和航空这两块。它一旦降价,影响的是航运、渔业、远洋作业和长途旅行,是一整条产业链的成本结构。
  • 想理解 AI 乐观派的世界观,去读一本伊恩·班克斯。《Consider Phlebas》或者《Player of Games》都行,读完你会明白为什么这帮人不怕 AI。
  • 把"AI 会不会毁灭人类"这个问题往旁边放一放,换一个更能指导行动的问题:我所在的行业,哪一部分工作是信息处理,哪一部分是物理操作。前者已经在被替代,后者还有几年缓冲。
YouTube 平台字幕(自动或人工),可能有识别错误,人名与数字请以视频为准。已按语句合并成段落,并去掉 um、uh 等口头填充词。 文稿来源:YouTube(自动/人工字幕)。

0:00Okay, let's unpack this. imagine a future, right, where you're not just talking to your phone, but you've got your own personal robot. Maybe even tens of billions of them across the planet. Wow. Sounds like pure sci-fi, doesn't it? You know, think C3PO or R2-D2, but maybe even more capable, becoming everyday things, right? Like household appliances almost. Exactly. And well, this kind of vision, it was actually a big part of a really fascinating discussion recently happened at a US Saudi event. Ah yes and the focus was squarely on you know what's next technology innovation that sort of thing and what's really interesting there I think is the context you had major

0:38leaders talking about how a big economy pivots you know how it thrives when innovation is the main driver yeah the fact that President Trump and his royal highness were part of it definitely signals how important this shift is seen to be moving away from say traditional resources towards a future built on well new ideas new tech absolutely and that's our goal today really to dig into some of the most well thoughtprovoking ideas from that event. So we're drawing on Elon Musk's perspectives mainly mostly his vision. Yeah. But also weaving in insights from Kevin break you know our longtime collaborator. Hi Kevin's work that gives it a good grounding. It's a compelling mix. I

1:16think you've got these like huge predictions alongside Kevin's observations from you know on the ground and that theme of building an innovation economy. I mean, the scale of Musk's predictions is just it's kind of mind-boggling. It really is. And Kevin Brakes research adds that crucial layer, doesn't it? How these big ideas are landing, how they're being perceived within Saudi Arabia specifically. So, let's dive straight into the robots then, because apparently Elon actually brought some of the Optimus bots to this event. Showed them off to the dignitary.

1:48I did. Yeah. I even heard something about a a Trump dance. That's the story. Yes. The robots apparently showed their stuff. Some pretty engaging moves including Yeah. that sequence playfully called the Trump dance. Ah, seriously. And even a YMCA performance apparently. Wow. So, it paints a pretty vivid picture of how fast these humanoid robots are, you know, coming along. It certainly does. But, okay, beyond the the dancing, Musk made this pretty astonishing prediction. M he sees a future with tens of billions of personal robots. Mhm. Billions. That's like way

2:27more than the number of people on Earth right now. It is an incredible number. And he made that connection, didn't he, to having your own C3PO or R2-D2, but maybe even better. It really makes you stop and think about the potential. Well, yeah. And it throws up huge questions about our economies, doesn't it? What does that do? Musk himself talked about that. He suggested almost unlimited economic potential, you know, driven by that level of robotic help. Unlimited potential. Yeah. His vision includes maybe a global economy like 10 times bigger than today's 10 times. Which leads to this idea he has of universal high income. Not basic income, high income. Okay, unpack that.

3:04Universal high income. Well, the idea is if you have such a massive lead in productivity from all these robots, maybe wealth generation fundamentally changes. So, it's not just about a safety net, but widespread prosperity. That's a really expansive concept, a totally different way of thinking about work and value. It is. And he pointed to sci-fi for a positive vision. Actually, the culture series by Ian Banks. Oh, interesting. Not the usual dystopian view then. Exactly. He mentioned it as a potentially positive non-disystopian future with advanced AI, which is Yeah.

3:38Very different from the sort of Terminator fears we often hear. Right. To kill a robot. Precisely. and he acknowledged those risks, you know, the Terminator scenario as he put it, but the emphasis seemed very much on aiming for maybe a Star Trek future. Okay. Exploration, prosperity. Yeah.

3:55Facilitated by technology, using these tools for human progress basically. Okay. So, that's the big robot picture. Now, shifting gears a bit, transportation, robo taxis, and autonomous vehicles were also big topics. Yeah. Sounds like there's a lot of excitement for that in the kingdom.

4:11There seems to be. Yes. And what's interesting is how they're framing it. Thinking of cars as essentially robots on four wheels. Robots on four wheels. Yeah. And potentially as cashgenerating units. Ah okay. So not just transport but an economic engine in themselves. Exactly. It shows a very forwardthinking view on autonomy. Not just convenience but a core part of the future economy. And the positive vibe at the event suggests, you know, maybe rapid adoption is on the cards there.

4:42Makes sense. And sticking with new tech adoption, Starlink came up, too. It did. Something about approval for maritime and aviation use. That sounds like a pretty big deal for connectivity. Oh, absolutely. Getting that green light for ships and planes using Starlink, that's going to directly boost connectivity across the kingdom, especially in areas, you know, that might have struggled before. Big implications for communication data. Right. Okay. And another sort of infrastructure idea that was floated. Yeah. Tunnels from the Boring Company. yes, Elon's other venture. The whole idea of going 3D with roads to be traffic.

5:19It's definitely a different way to think about it. It is this concept of tunnels creating a kind of well he called it a wormhole effect for transport. A wormhole like shortcutting through spaceime. Huh. Well, maybe not quite that sci-fi, but the idea is you bypass all the surface level congestion by going underground, drastically cutting travel times. And they have that working example in Vegas, right? Exactly. That's a proof of concept. Shows how it can work in a real city, what it could do for urban travel. Okay. Now, let's pivot to something maybe a bit more abstract, but hugely ambitious. X AI, right? His

5:54AI venture. For those of us not, you know, deep in the AI world, what's the core mission there? What's XAI about? Well, the crucial thing to get, I think, is the goal. It's not just about building artificial general intelligence or AGI. Okay. The aim is to create AI. That is, and this is the phrase, maximally true seeking. Maximally true seeking. Yeah. The fundamental mission is quite profound to understand the universe. Understand the universe. Wow.

6:21It's about tackling the biggest questions, you know, about reality, existence. You can almost think of it as like AI aiming for enlightenment. So, trying to answer things like where did the universe come from? Where's it all going? Exactly. And even he mentioned discovering what questions do we not know to ask. Whoa. Okay. That's that's deep. The unknown unknowns. It really is. And Musk's philosophy behind it, as he put it, seems driven by this really deep curiosity, a desire to just understand how reality works fundamentally. And XAI is the tool for that. That seems to be the vision, a powerful tool to help humanity try and unravel some of those, you know,

7:00ultimate mysteries. This all paints a picture of just incredibly ambitious future focused thinking, which brings us back to Kevin Break's perspective. You've worked with him for a while. How does his research frame all this? Especially looking at those presentations at the Saudi Economic Forum. Kevin's perspective is really valuable here because he's been tracking the innovation scene in Saudi Arabia closely. Yeah. His analysis, especially of Musk's talks there, is generally very positive, very upbeat. Okay. As Kevin put it himself, it gives you a real chance to quote, "Get a sneak peek into what is ahead for innovation in Saudi." So, what were the key takeaways from

7:36Kevin's research then? What potential did he highlight based on these discussions? Well, pulling together Kevin's analysis, there's this clear sense of real excitement, you know, about how robotics, autonomous vehicles that are connectivity via Starlink. Yeah. How all that could genuinely spark a new era there, a new era of growth.

7:57Yeah. Economic growth, technological progress within the kingdom. His research really emphasizes the potential for like significant jumps in productivity and a real tangible shift towards an economy truly driven by innovation. He sees these initiatives as potentially, you know, fundamentally reshaping their economic future. Okay, so let's try and sum this up. We've touched on a future potentially teeming with billions of personal robots changing the economy, right? Universal high income robo taxes transforming how we get around cars as robots on four wheels. greatly improved connectivity from things like Starlink, even digging tunnels to reimagine roads going 3D.

8:36Yeah. And kind of underpinning a lot of the ambition is this goal of using XAI to understand the universe itself. The maximally truth seeeking AI all happening within this context of major innovation focus in Saudi Arabia which Kevin Break's research really highlights. Exactly. And these aren't just you know pie in the sky ideas. They represent potential shifts that could really genuinely impact your life, your world. Yeah. Think about those key points. The sheer number of robots predicted, cars becoming economic units, AI searching for fundamental truths. It points to a future that could look radically different and potentially hopefully very beneficial. That's the

9:15optimistic take. Yeah. So, as we sort of stand on the edge of this Mhm. this potential wave of innovation, what are the big questions we should be asking ourselves, you know, to make sure this future lines up with what we actually want? It's definitely food for thought. It really is. And look, if you're interested, if you want to explore these things more, yeah, where can people go?

9:33You can find more info on the tech itself at the official websites. You know, Tesla is tesla.com, SpaceX is spacex.com, and the Boring Company is boring company.com. Okay. Good resources to check out. Yeah. great places to dig deeper into the specifics if this has sparked your curiosity. And we genuinely love to hear what you think about all this, you know, these potential futures.

9:55Feel free to share your comments. Let us know what really stood out to

资源与出处

相关视频

同频道、同主题或相邻年份