萨提亚·纳德拉:AI 作为爬山机器

Satya Nadella: AI as the Hill-Climbing Machine

萨提亚·纳德拉 Satya Nadella · 雷德·霍夫曼 · 2026-06-05 · 约 60 分钟 · 原视频 ↗

打开互动全文版(中英对照 + 朗读 + 问答)→

本期速览 · Overview

萨提亚·纳德拉探讨工作的未来、AI 作为公司的核心,以及诗歌在理解人类体验中的重要性。

Satya Nadella discusses the future of work, AI as the core of the firm, and the importance of poetry in understanding human experience.

要点 · TL;DR

核心观点 · Key points

反共识 · Contrarian takes

本期章节 · Chapters(共 23)

全文 · Full transcript(中英对照)

未来工作与AI战略 Future of Work and AI Strategy

Satya

所以,我认为我们尚未在概念上正确理解并达成共识的是,对于一位科技 CEO 来说,工作的未来会是什么样子?你必须深入了解技术栈。AI 不是一项技术,它是公司的未来。我有一条格言:不要用前沿模型解决非前沿问题。我认为在 AI 时代,这将涵盖一切,对吧?我想,Reed,如果一年后你坐在这里,而世界还没有完全转向思考“我的 AI 供应链是什么样的”,我会非常惊讶。

So, what I think we've not yet conceptually gotten right and a shared understanding is what is this future of work going to look like for a tech CEO? You have to be deep inside of what's the tech stack. AI is not a technology. It's the future of the firm. One of the dictums I have is don't use frontier models for non-frontier problems. I think in the AI age that is going to be everything, right? I think I would be very surprised Reed if you were sitting here a year from now if the world is not completely turned on what is my AI supply chain look like.

引言与诗歌 Introduction and Poetry

Host

我非常高兴地介绍一期《Possible》特别节目,嘉宾是微软董事长兼 CEO 萨提亚·纳德拉。萨提亚和我相识已久,这期节目某种程度上关乎人类的 AI 革命。我觉得这会是一期很棒的节目。我们涵盖了各种重要话题。萨提亚一如既往地优雅、连贯、聪明、全面,最重要的是,他对我们的 AI 未来充满人文关怀。这将是一期精彩的节目。实际上,和萨提亚在这里拍摄的一个好处是,它让我想起了我们最早讨论微软和 LinkedIn 的日子。

I couldn't be more delighted to introduce a special episode of Possible with Satya Nadella, the chairman and CEO of Microsoft. Satya and I have known each other a long time, and part of in this is kind of AI revolution for humanity. I thought this would be a great episode to get out. I mean, we covered all kinds of important topics. Satya, as always, is elegant, is cohesive, is smart, is comprehensive, and above all, humanist in what is our AI future. This will be an amazing episode. Actually, one of the things that's great with Satya about filming this here is it reminds me of the earliest days when we were talking about Microsoft and LinkedIn.

Satya

没错,因为我们之前在 Greylock 办公室进行过一场非常精彩的对话,所以能回来真是太好了。

That's right, because we did one of our very, very great conversations here in the Greylock office, so it's like just awesome to be back.

Host

我想从一件我不知道有多少人了解并欣赏你的事情开始,那就是你家里有多少本诗集。你能谈谈你对诗歌的喜爱、最喜欢的诗人,以及你是如何接触诗歌的吗?

I want to start with something that I don't know as many people realize and appreciate about you, which is with how many books of poetry you have in your house. Can you say a little bit about your attraction to poetry, favorite poets, what the engagement is there?

Satya

是的,我其实是在人生的不同阶段开始接触诗歌的。我记得,中学时我们有一本标准的英语诗歌教材,我一生都在试图找回它,但不幸的是它已经绝版了。那本书里介绍了雪莱、华兹华斯,甚至还有像萨罗吉尼·奈杜这样的印度英语作家。我不知道,也许是我的注意力跨度或其他什么原因,我自然而然地被诗歌吸引,把它当作一种享受和热爱的东西。我甚至经常把它比作代码——诗歌是压缩的极致形式。所以每当我无聊的时候,我就会去读诗。其实我并不擅长深入理解,我从未研究过它,我不是那种专家。这就是为什么我懂诗的名声远远超过了我实际的诗歌知识。但我仍然继续把诗歌当作人类体验的最佳表达方式,对吧?我的意思是,如果你广义地看待文学,它比历史更能捕捉人类体验,而我认为诗歌是它的压缩形式。

Yeah, I mean, I actually got into it in different times of my life. I remember that, you know, as a middle schooler we had this standard issue English poetry book, which I've been trying to reclaim and get all my life, but unfortunately it's out of print, but it sort of had, you know, even getting introduced to Shelley or Wordsworth and it had even sort of Indian authors like Sarojini Naidu writing in English and it is I don't know. I felt maybe it was my attention span or what have you. I was naturally drawn to poetry as a thing to enjoy and love and I've always compared it even to code which is it's sort of compression in its best form. And so whenever I'm bored, I get to, you know, go read. I'm not great actually at understanding deeply. I've never studied it. I'm not like so and so. That's why my reputation of knowing about poetry is far exceeds my knowledge of poetry. But I still, you know, continue to use poetry as perhaps the best expression of the human experience, right? I mean if you broadly think about literature as sort of what captures more than even history the human experience, I think poetry is the compressed form of it.

Host

我完全同意。有没有某位诗人是你经常重温的?

I completely agree. Is there a particular poet that you go back?

Satya

是的,我真正要说的是,最能触动我的诗歌可能是乌尔都语诗歌。这是因为我在印度海得拉巴长大,乌尔都语无处不在,一些乌尔都语诗人,无论是现代的还是 17、18 世纪的,都非常杰出。特别是有一位名叫迦利布的诗人,他非常出色,还有像费兹这样的现代诗人。我在海得拉巴长大,也深受鲁米的影响。事实上,我上的高中非常有趣。我们开设了很多语言课程,显然有英语,有印地语(国语),还有梵语——大概就像这里的拉丁语吧。然后还有当地语言泰卢固语,那是我的母语。但我们还有乌尔都语和波斯语。实际上,当我们分班学习所谓的第二语言时,我的第二语言是梵语。但你知道吗,我看到我的同学们有的去学波斯语、乌尔都语、泰卢固语、印地语,这真是太迷人了。

Yeah, so the thing that I really would say the poetry that probably speaks to me most deeply is the Urdu poetry. That is sort of because I grew up in Hyderabad in India and Urdu is sort of in the air and some of the Urdu poets both modern and sort of people in the 17th, 18th century were just extraordinary. And in particular, there's this poet whose name goes by Ghalib and he's just extraordinary or even a modern poet like Faiz. I was also very having grown up in Hyderabad, I was very influenced by Rumi. In fact, the high school I went to was fascinating. In fact, we had the number of languages which were all taught was obviously there was English. There was Hindi, which is the national language, and then we had Sanskrit. Like basically, you know, like Latin here, I guess. And then we had the local language Telugu, which is my mother tongue. But we also had Urdu and Persian. And so, all of them like in fact, when we would break for what was called second language, I my second language was Sanskrit. But you know, they I could see I had classmates who would go to Persian, Urdu, Telugu, Hindi, and it was fascinating.

Host

嗯,在这些经历中,我有点语言嫉妒,因为我喜欢的一个笑话是:你会怎么称呼一个说三种语言的人?三语者。两种语言?双语者。一种语言?美国人。不幸的是,我就是这个笑话里的美国人。好了,我们刚刚经历了精彩的 Build 大会。让我们实际上从历史开始。你知道,微软的第一个产品是一个 BASIC 解释器。50 年后,在 Build 大会上,公司将未来定位在别人能用 AI 构建什么上。那么,在这个 AI 时代,新的 BASIC 解释器是什么?

Well, among the experiences I'm experiencing a little bit of language envy because I a joke that I like is you know, what do you call a person who speaks three languages? Trilingual. Two languages, bilingual. One language, American. And unfortunately, I resemble this joke. All right. So, let's we're just coming out of Build, which was amazing. Let's kind of actually start with kind of going back in history. So, you know, Microsoft's first product was a basic interpreter. And 50 years later, here at Build, framed the company's future on what others can build with AI. So, what is the new basic interpreter in this AI era?

Satya

是的。这个问题问得好。事实上,我认为这正是公司 DNA 的核心,对吧?也就是“别人能构建什么,因此我们应该构建什么?”这是我们一直问自己的两个问题。在 AI 时代,对我来说,答案就是 BASIC 解释器是爬山机器。对吧?因为你想,现在所有这些 AI,它是什么?它基本上是接受一个目标、一个结果、一个评估,然后通过学习使用数据、使用一些强化奖励来学会如何实现它。所以整个大会的重点不是“嘿,这是一个新的前沿模型”,而是帮助每个开发者、每家公司,无论是 AI 原生初创企业还是企业,构建他们自己的爬山机器,这样他们就能在前沿运作。所以,我认为就是这样。事实上,你要非常清楚你关心的评估和目标,知道如何评估它们,对吧?这在某种意义上是最重要的,而那就是新的知识产权。

Yeah. That's a great one. So, in fact, it's that I think is the core of the DNA of the company, right? Which is what can others build, and therefore what should we build? It's sort of the two questions that we always ask ourselves. And in the AI era, to me, the answer is the basic interpreter is the hill-climbing machine. Right? Because you think of you Now, bring all of this AI, what is it? It is basically taking an objective, an outcome, an eval that you have, and learning how to achieve that by learning using data, using some reinforcement reward. And so, the entire conference was about not Hey, here is a new frontier model. It was about helping every developer, every company, whether it's a AI native startup or an enterprise, building their own hill-climbing machine so that they can operate at the frontier, right? So, I think that that's it. The idea In fact, the you know, being very clear about the evals and the objectives that you care deeply about, knowing how to evaluate them, right? That it's in some sense is the most and keeping That's the new IP.

Host

是的。

Yes.

Satya

因为其他一切都很机械。但知道你想用哪组数据来训练模型以及如何奖励它,可能正是下一代知识产权产生的地方。

Because everything else is pretty mechanical. But knowing what is the set of data that you want to train a model on and how you reward it is probably where the next level of IP gets created.

Host

嗯,让我们深入探讨几个不同领域,因为你和我进行过多次关于微软战略的对话。其中一个问题是企业如何保持自身数据的优势和完整性。而微软是全世界最适合做这件事的公司。

Well, and one of the things let's dig into a couple different areas here because you and I've had a number of conversations, you know, with Microsoft strategy. And one of them is kind of this question about how enterprises keep the advantage and integrity of their own data. And Microsoft is the most natural company in the entire world to do this.

企业AI战略:数据控制与隐性知识 Enterprise AI Strategy: Data Control and Tacit Knowledge

Host

那么,请谈谈企业应该如何思考这个问题:“我们需要自己的前沿智能,但我们也需要保持对数据的控制。”

So, say a little bit about how enterprises should be thinking about, 'We need our own frontier intelligence, but we also need to maintain control of our data.'

Satya

是的,我认为这就是问题所在。未来的经济将由人力资本和所谓的“代币资本”共同塑造。这对微软如此,对初创公司如此,对存在了 100 年的银行也是如此,都一样。我们所有人都需要——实际上,你需要做的就是发挥人力资本和代币资本之间的相互作用,并让它们的回报复利增长。所以,如果你这样看问题,那么最重要的事情之一就不仅仅是考虑数据的整体概念。企业或公司的隐性知识是什么?是你能够运营、做出判断、拥有品味的独特方式。所有这些都是隐性知识,目前大部分体现在人力资本所拥有的隐性知识中,以及一些数字化的产物。那么,当谈到 AI 时,模型在某种意义上能够通过人类的行为轨迹提取这些隐性知识,然后将其编码到模型的权重中。所以,我的主张是,每个企业现在都需要更加关注人类与其数字资产之间的这种相互作用,这些轨迹本质上是在训练模型,企业要将其作为知识产权保留,而不是泄露出去。因为一旦泄露,就是一扇单向门,你在某种意义上就完了。这就是你的独特之处,你可能花了 100 年才建立起来。事实上,我们甚至不知道如何表达它。没有人在资产负债表上有一个叫“隐性知识”的条目,但我们理所当然地认为,因为我们拥有人力资本,我们就拥有它。现在,我相信它可能会泄露。事实上,它正在泄露。你看看那些模型公司是如何学习的,它们基本上是在设立带有奖励的“健身房”,雇佣那些以前在你公司工作过的员工。这应该告诉你一切,告诉你什么不应该发生。所以,这就是为什么我们想要从根本上改变这种机制或范式,让你欢迎模型进入。它们应该在你控制的机器内部进行爬山优化。你的数据就是你喂给模型的上下文。你实际上是在收集企业内部人类和智能体之间如何完成工作的这些痕迹或轨迹。但你有一个持续的循环,并且不让它泄露。我认为这是基本的操作。

Yeah, I think that is the question. This economy is going to be shaped going forward by both human capital and let's call it token capital. That is true for Microsoft, that is true for a new startup, that's going to be true for any bank that's been in existence for 100 years. Doesn't matter. All of us will now need to sort of, in fact, the interplay between human capital and token capital and compounding the returns of it is what you need to do. So, if you frame it that way, then one of the most important things is not just even thinking of data in its aggregate sense. What is the tacit knowledge of an enterprise or a firm? It's the unique ways that you're able to operate, pass judgment, have taste. All that's the tacit knowledge, mostly captured today in the tacit knowledge that is there with the human capital, and some artifacts that are digital. So now, when it comes to AI, the model, in some sense, is able to extract that tacit knowledge through human trajectories, and then code it in a set of weights in a model. So, what I claim is that every enterprise now needs to be more mindful about that interplay of humans and their digital estate working together, those trajectories training essentially the models that they keep as IP versus leak it. Because if you leak it, it's a one-way door. You're done in some sense. Which is what is unique, what you may have spent a hundred years. In fact, we don't even know how to articulate it. Nobody has a line item in their balance sheet called tacit knowledge, but we take it for granted that because we have human capital, we have it. Now, I believe that can leak. And in fact, it is leaking. If you look at even how the model companies learn, they're essentially setting up these gyms with rewards, which are employing employees who worked at your company previously. That should tell you everything about what should not be happening. So, that's one of the fundamental reasons why we want this effectively the regime to change or the paradigm to change, where you welcome the models to come in. They should hill climb inside a machine that you control. Your data is your context you feed the model. You collect, in fact, these traces or trajectories of how work gets done between humans and agents inside the enterprise. But you have a continuous loop of that and you're not letting that leak. That I think is the fundamental operation.

Host

是的。实际上,与隐性知识相关的一个平行话题是——这也是你我在微软董事会讨论过的问题——AI 员工的未来角色是什么?因为部分问题在于:“嘿,我们不仅要提供放大人类工作的工具,这对 AI 公司来说很好,我们还要提供至少是专业员工。”但挑战在于,对企业来说真正重要的事情之一是,你的员工体现了大量的隐性知识,知道如何做、做什么。你希望这些知识出现在其他公司,并在其他地方提供吗?你和微软对未来工作的看法是什么,特别是如何将 AI 也视为——甚至可能是专业员工?

Yep. And actually the related parallel to the tacit knowledge is, and this is one of the conversations you and I had at the Microsoft board, which was what is the future role of AI employees? Because part of the question is, 'Hey, we're going to provision not just tools to amplify human work, great for AI companies, but we're also going to provision at least specialist employees.' But the challenge with that is that one of the things that really matters to enterprises is your employees embody a lot of tacit knowledge, how to, what to do. Do you want that in other companies and provision other places? What's your and Microsoft's view of the future of work when it comes down to how to think about AI as also, maybe even specialist employees?

Satya

是的,我认为没错。那么,思考这个问题的方式,我有两个角度。先打个比方,我稍后会回到这个话题。如果有人在 80 年代初跑来对我们说:“你知道吗?将来会有 40 亿打字员,每天早上起来就开始打字。”我们会说:“为什么?这毫无意义。我们有打字员池,有幻灯片池,我们这样挺好的。”但除了我们发明了一个全新的东西叫知识工作,每个人都在打字、创建文档等等。所以,我认为我们在概念上还没有完全理解,我们的共识还不够清晰的是,当你拥有——以微软为例,我们有 2 万名员工,还有比如说 200 万个智能体或 2000 万个智能体,全部在一个循环中——未来的工作会是什么样子?会发生什么?会产生什么样的隐性知识?智能体之间、智能体与人类之间传递的产物是什么?所有这些都将在未来几年内展开。你可以在编程中看到早期的形式。这是一个观察社会变革的好地方。想想看:我们从传统的 IDE 开始,说:“嘿,AI 在里面。”它做代码补全。这本身就很有用,容易理解。我们一直有拼写纠正,VS Code 中的 IntelliSense 从 15 年前就有了,而且越来越好。然后我们说:“哦,你不用再跳出 IDE 去浏览器搜索 Stack Overflow 了,你现在可以把所有编程知识带到聊天会话中。”这也很好理解,你仍然在 IDE 里,有聊天功能。然后我们说:“好了,你现在有了推理模型和一些初步的智能体循环,所以你可以分配任务。”于是我们有了智能体模式。你不仅有聊天,还可以给它小任务,看着它完成,然后你可以接受它所做的并插入代码。然后来了一个重大突破:完全自主和长时间运行的智能体循环,你可以真正地“发射后不管”。你可以分配一个高级意图,它会去完成整个 PR,然后你接受这个 PR。所以,我认为这种转变将发生在所有工作中。现在我们在 Copilot 中也看到了这一点,我们有聊天和协同工作,在 Build 大会上我们宣布了 Scout 和 Autopilot。所以,编程中发生的事情也会发生在知识工作中。有趣的是,我们在 GitHub 的 Build 大会上还推出了一项新功能叫 Canvas,因为随着我们都变得非常擅长使用这些编程智能体,情况发生了变化。

Yeah, I think that's right. So, the way to think about this, there are two ways I come at this. Let's take one analogy and I'll come back to it. If somebody in the early '80s had come to us and said, 'You know what? There are going to be 4 billion typists who are going to wake up every morning and start typing.' And we would say, 'What for? It makes no sense. We have a typist pool, we have a slide pool, and we're fine with that.' But except we invented this complete new thing called knowledge work, where everybody was typing and creating artifacts and so on. So, what I think we have not yet conceptually gotten right, and our good understanding, and our shared understanding, is what is this future of work going to look like when you have, let's call it, and let's take Microsoft, we have 20,000 employees and we have, let's say, 2 million agents or 20 million agents. All in a loop. What is happening? What is the tacit knowledge that gets created? What are the artifacts that are going between agents, between agents and humans, and so on. All that's now going to be played out in the next multiple years. You can see early forms of this in coding. It's a great place to observe the social change. Think about it: We started in a good old IDE and said, 'Hey, AI is in there.' And it's doing code completion. That itself was useful, easy to understand. We've always had spelling correction and IntelliSense was there in VS Code from 15 years ago, and it just got better. Then we said, 'Oh, instead of going out of band and going to a browser and searching Stack Overflow, you can now bring all that coding knowledge to a chat session.' Well, that was also easy to understand, and you still were in the IDE, you had the chat. Then we said, 'Okay, you now have reasoning models and some preliminary agent loops, and so you can assign tasks.' So we had agent mode. You not only had chat, but you could give it small tasks and you could see it complete, and then you could accept what it did and insert it. Then came the big breakthrough of total autonomy and agentic loop working for long periods of time where you could literally fire and forget. You could assign a high-level intent and it'll go off and do the full PR and you would accept the PR. So that transition is what I think is going to happen across all work. And now we're seeing it with even in Copilot, we now have chat and we have co-work, and now at Build we announced something called Scout and Autopilot. So, what is happening in coding will happen even in knowledge work. Interestingly enough, one of the things we launched even at Build in GitHub was a new feature called Canvas, because what has happened is as we've all gotten so good at using these coding agents.

管理多智能体的挑战 Challenges of Managing Multiple Agents

Satya

事实上,我们目前最大的挑战是,我打开了 100 个 CLI 会话,试图操作这 100 个智能体。管理这些智能体带来的认知负荷非常高。这是一个命令行中的线性聊天会话,而我有 100 个这样的会话。所以,我们现在有了一个新的 IDE,我们称之为 ADE——智能体式开发环境。这就是新的 GitHub 应用。它看起来像一个收件箱,但实际上是智能体的收件箱,这些智能体跨所有仓库工作,让我能够对我赋予它们的宏观委托进行微观操控。这是一个完整的 UI,供智能体与我交互,也供我与智能体交互。我们在 GitHub 中引入了一个名为 Canvas 的新功能,其中看板可视化有助于管理 PR。这对智能体和我的交互都非常有用,相比聊天会话更高效。我认为这种创新将改变工作方式。

In fact, the biggest challenge we now have is I have a hundred CLI sessions open, trying to operate these hundred agents. The cognitive load on me managing this is so high. It's a linear chat session in a command line, and I have a hundred of them. So, we now have a new IDE, which we call an ADE—an agentic development environment. That's what the new GitHub app is. It looks like an inbox, but it's an inbox of agents working across all repos, allowing me to do micro-steering of the macro delegation I gave them. It's a complete UI for agents to deal with me and for me to deal with agents. We introduced a new feature called Canvas in GitHub, where a Kanban board visualization helps run down PRs. It's very useful for both agents and me to interact, versus a chat session. I think that type of innovation is how work will change.

Satya

另一个让我着迷的事情是 AI 研究的一个方向:模型将变得更加擅长坚持既定路线,并理解人类在可操控性方面的偏好。这就是我想要的。我希望模型不仅能遵循指令,而且真正可操控。当你拥有这一点时,你就会对它充满信心。未来的工作关乎这种互动所创造的隐性知识。就像知识工作是由数字制品和人力资本创造的一样,新的工作将是 AI 资本和人力资本协同工作,创造数字制品。

Another fascinating thing for me is a line of AI research: models will become much more tuned to learning how to stay the course and understand human preferences in steerability. That's what I want. I want models that not only follow instructions but are also really steerable. When you have that, you have confidence in it. The work of the future is about tacit knowledge created by this interplay. Just as knowledge work was created by digital artifacts and human capital, the new one will be AI capital and human capital working together, creating digital artifacts.

企业结构:安全性与可管理性 Enterprise Structure: Security and Manageability

Host

你认为人们忽略了哪些构建额外结构以赋能企业的重要性?显然,赋能人类的画布、ADE 等是一方面。但像企业安全这样的概念呢?

What do you think are the things that people miss about how important it is to build out additional structure for enabling enterprises? One is obviously the enabling humans canvas, ADEs, etc. But what about notions like security in an enterprise?

Satya

说得好。一是体验层,显然很重要。另一个是我们之前提到的爬山机器概念,需要实例化。但第三是管理性和安全性,从可观测性开始。我们构建了名为 Agent 365 的东西。我需要了解微软 2000 万个智能体的清单。我需要知道它们是什么、在做什么、推理轨迹——完全可检查和可审计。当智能体执行时,它们可以生成代码并运行,因此环境需要受策略管控。我们需要给它们身份、沙箱,并设置策略。我们扩展了 Entra 用于身份,Defender 用于安全,Purview 用于数据标记和保护。安全、遏制、管理性、可观测性是我们对这些智能体建立信心的方式。

Great point. One is the experience layer, clearly important. Another is the hill-climbing machine concept we referenced earlier, which needs to be instantiated. But the third is manageability and security, starting with observability. We built something called Agent 365. I need an inventory of the 20 million agents at Microsoft. I need to know what they are, what they are doing, their reasoning traces—fully inspectable and auditable. When agents execute, they can generate code and run it, so the environment needs to be governed by policy. We need to give them identities, sandboxes, and set policies. We extended Entra for identity, Defender for security, Purview for data labeling and protection. Security, containment, manageability, observability are how we gain confidence in these agents.

Satya

另一件重要的事情,正如 Reed 在开发者大会上提到的,是构建长期运行的智能体。在编程语言中,我们有运行时程序可验证性的模型。我们为 Foundry 添加了称为断言的功能,用于这些长期运行的智能体。这使我们能够断言边界。与其将护栏视为某种分类器,你需要在执行过程中保持路径不偏离轨道。随着我们构建平台、运行时、安全层、管理层和可观测性层,出现了许多工程上的精妙之处。

Another important thing, as Reed mentioned at the developer conference, is building long-running agents. In programming languages, we have models for program verifiability at runtime. We added something called asserts to Foundry for these long-running agents. This allows us to assert boundaries. Instead of guardrails as a classifier, you need the ability during execution to keep the path from going off rails. There's a lot of engineering sophistication emerging as we build out the platform, runtime, security layer, management layer, and observability layer.

CEO角色与AI战略建议 CEO Role and AI Strategy Advice

Host

我们来谈谈 CEO 的角色。《财富》杂志最近形容你像微软 AI 团队内部的创业 CEO。你认为 CEO 们应该在 AI 方面做些什么?AI 为许多公司带来了重新奠基的时刻。你对他们的参与方式有什么建议?

Let's talk about the role of CEO. Fortune recently described you as acting like a startup CEO inside Microsoft's AI teams. What do you think CEOs should be doing around AI? AI brings a re-founding moment to many companies. What's your advice for how they should engage?

Satya

好问题。如果你是科技 CEO,你必须深入了解技术栈。没有对未来的基本世界观,就不可能成为科技 CEO,而且在它成为共识之前,你必须对公司的发展方向做出判断。这是我们行业的一个二元转变。如果 CEO 不带头冲锋、承担风险,知道这些是艰难的转型,那就没有希望。但如果你不是科技 CEO,你需要在你所在的领域——银行、医疗等——成为一名优秀的 CEO,并且你可以拥有优秀的合作伙伴和技术顾问。然而,我正在改变我之前的看法。我注意到正在发生的事情,我认为更广泛的 CEO 群体还没有意识到这一点。他们仍然乐于与科技公司一起发布新闻稿,说他们有 AI 战略,指向八个智能体和一些成果。但 AI 不再仅仅是一项技术;它是公司的未来。

Great question. If you're a tech CEO, you have to be deep inside the tech stack. There's no way to be a tech CEO without a fundamental worldview on where the future is going, and to be long before it's conventional wisdom, you must pass judgment on where the company is going. It's a binary transition in our industry. There's no hope if the CEO doesn't lead from the front and take the shot on goal, knowing these are harsh transitions. But if you're not a tech CEO, you need to be a great CEO in your domain—banking, healthcare, etc.—and you can have great partners and technology advisers. However, I'm changing my prior on that. I notice what's happening, and I don't think the broader CEO community has woken up to this. They're still happy doing a press release with a tech company, saying they have an AI strategy, pointing to eight agents and some outcome. But AI is no longer just a technology; it's the future of the firm.

代币资本与企业转型 Token Capital and Firm Transformation

Satya

对吧?这就像说“哦,我不知道我公司的人力资本是什么”。所以,我认为关键在于你必须深刻理解你的 Token 资本。你能回答这个问题吗?你能具体地说,“哦,昨晚或者基于我们所有运营中完成的工作,我们能够将其转化为一组知识,这些知识现在某种程度上成为了我的 Token 资本的一部分。它可以是某些上下文,可以是某些技能,可以是模型中的某些权重。它不必是单一的东西,但你需要能够清楚地识别它,作为你拥有、控制、创造的东西,并且你建立了一个系统让它复合增长。”所以,我认为这可能是最艰难的变革。这种变革不同于进入移动电话时代、PC 时代或云时代,那时我只需要一个 IT 部门,知道如何与一堆供应商打交道,做一些聪明的事情来降低成本或提高效率。它应该从那里开始。我不是说那不是起点,但关键在于,当 AI 基本上了解了你所在行业需要知道的一切时,你的行业结构会发生什么变化。如果你从那里开始,那么我认为你会开始理解这不仅仅是一次技术变革。它实际上是关于公司的根本性变革。

Right? It would be like saying, 'Oh, I don't know about the human capital in my firm.' So, I think that this is where my thing would be that you have to now get a deep understanding of what's your token capital. Can you answer that question? Can you concretely say, 'Oh, last night or based on all the work that happened across all of our operations, we were able to translate that into a set of knowledge that is somehow now part of my token capital. It can be some context. It can be some skill. It can be some weights in a model. It doesn't need to be any one thing, but you need to be able to identify it clearly as something that you own, you control, you created, and you put in place a system to have that compound.' So, I think that that is going to be probably the toughest change. So, this change is unlike going to a mobile phone era or a PC era or a cloud era where all I needed to do is to have an IT department that knew how to deal with a bunch of vendors and did some smart things to reduce cost or improve my efficiency. It should start there. I'm not saying that that's not the place, but this is about what happens structurally in your industry when AI basically knows everything that it needs to know about being in your industry. And if you start there, then I think you will start understanding that this is not just a tech change. It is really about a fundamental change to the firm.

通过平台战略建立信任 Building Trust Through Platform Strategy

Host

而且我认为,考虑到公司的本质,你在领导微软方面做得非常出色。众多天才时刻和技能之一就是处理收购和合作伙伴关系,对吧?无论是 LinkedIn——显然对我们来说很亲近,还是 GitHub、OpenAI 等等。我认为微软在全球领先的一点是,你如何在这些其他生态系统中建立信任?在 AI 的世界里,这将极其重要,尤其是在公司变革之际。那么,你学到了哪些经验,你倡导哪些事情来建立和维持这种信任?

And one of the things that I think thinking of the nature of the firm is that you've done a great job in leading Microsoft. One of the many different genius moments and skills is going through acquisitions and partnerships, right? So, whether it's LinkedIn, obviously something close to our hearts. GitHub, OpenAI, etc. And part of what I think Microsoft leads the entire world on is how do you build trust in these other ecosystems? And in the world of AI, that's going to be extremely important, especially as the firm changes. So, what are some of the lessons you've learned, the things that you are advocating for in order to build and maintain that trust?

Satya

这非常关键,因为从某种意义上说,我一直在思考什么是长期稳定的,对吧?即使从微软的角度来看。长期稳定的是我们成为一家工具和平台公司,我们的根本定义是在平台之上创造的价值量,这应该远远超过平台本身捕获的任何东西,对吧?这是实现稳定的唯一途径。如果你做到了,那么你就拥有了信任。因为这样客户和合作伙伴就知道这不是零和博弈。尤其不是那种博弈论中的零和游戏,比如“我补贴这个两三年,然后抢你的饭碗”——科技公司非常擅长这种游戏。所以我一直觉得,“嘿,那不是正道。正确的思考方式是,你要非常有原则地认为你是一家平台公司,你的生死存亡取决于你在平台之上创造成功的能力,当他们成功时,你也会成功。”这就是建立信任和双方长期稳定的公式。我认为在 AI 时代,这将是一切。我会非常惊讶,Reed,如果一年后你坐在这里,世界还没有完全转向思考“我的 AI 供应链是什么样的”,它如何帮助我作为一家公司复合 AI 的回报,让我能够独特地指出这是我的价值——在一个我们知道这些学习系统按定义没有边界的世界里。它们必须有边界,主要是因为市场、社会和其他机制的结构。

So it's so key because in some sense I've always thought about what's long-term stable, right? Even from a Microsoft perspective. What is long-term stable is for us to be a tools and a platform company where we fundamentally are defined by the amount of value that gets created on top of the platform, which should far exceed anything that is captured in the platform, right? That's the only way to have stability. If you do that, then you have trust. Because then the customer, the partner knows that it's not a zero-sum game. Especially it's not a game theoretic zero-sum with these games where, 'Oh, I'm going to subsidize this until 2 years or 3 years only to then eat your lunch' type of games that the tech people are really good at. And so, I've always felt like, 'Hey, look, that is all not the way. The way to think about it is to be very principled that you are a platform company, that you will really live and die by your ability to create success on top of the platform, and when they're successful, you're going to be successful.' And that is the equation that builds trust and long-term stability for both sides. I think in the AI age, that is going to be everything. I think I would be very surprised, Reed, if you were sitting here a year from now if the world is not completely turned on, what is my AI supply chain look like, where it is helping me as a firm compound the returns of AI that I can uniquely point to as my value in a world where we know that these learning systems by definition don't have boundaries. They have to have boundaries because of mostly structures of markets and society and other mechanisms.

前沿领域:科学与创始人模式 Next Frontiers: Science and Founder Mode

Host

是的。我们谈了很多关于聊天机器人和代码的话题,这些领域刚刚经历了惊人的爆炸性进展。未来几年,LLM 的下一个前沿领域是什么?

Yep. So, we've talked a lot about chatbots and code, which have just seen amazing explosive progress. What are some of the next frontiers for LLMs over the next few years?

Satya

嗯,说到这个,事实上,我知道有一件事让你非常非常兴奋,坦白说,我为你感到兴奋,但它也让你审视自己把时间花在哪里,我让你自己来说。但看到科学领域正在发生的事情,简直难以置信。事实上,你和 Sid、Monis 创立的公司就是一个很好的例子,说明了在编码和知识工作领域发生的事情。如果它开始在科学领域发生,事实上,我甚至希望我们当初是反过来做的,对吧?因为如果我们从一些具有巨大社会效益的科学发现开始,AI 的社会许可度会高得多。那样人们就会真正认为这是一件更有帮助的事情。所以,我为你正在做的事情感到非常兴奋,Reed。显然,你花了很多时间,但也许你想谈谈这个。

Well, I mean, talking about that, in fact, one of the things that I know you have gotten very, very excited about, which I'm quite frankly, excited for you, but it also has caused you to examine where you spend your time, and I'll let you speak to it, but it's sort of unbelievable to see what's happening in science, and in fact, the company that you started with Sid and Monis, it's just a great example, I'd say of what has happened in, let's call it, coding and knowledge work. If it starts happening in science, in fact, I would even claim that I wish we did it in the reverse order, right? Because the social permission for AI would be so much higher if we had started with some nice discoveries in science that were having great societal benefits. Then people would have really thought of this as a thing that is really going to be more helpful. So, I'm really excited for you, Reed, what you're doing. Obviously, you've spent a lot of time, but maybe you want to talk about that.

Host

嗯,是的,我的意思是,这是我在过去一个月意识到的事情之一:我们看到 Manas 取得了如此大的进展,实际上我们内部有一个“化学的 Move 37”的描述,因为我们看到了可能针对非常有趣的癌症和其他疾病的化学物质。你知道,这还很早期,对吧?所以这些事情需要时间,需要所有的 ID 验证,但事实是,我们已经开始看到——我们拥有一些世界上最好的计算化学家。他们看了看说,“这非常有趣。我们以前从未见过,这可能会有效。”对吧?所以诸如此类。Sid、Uzair 和我在讨论这个,我说,“听着,我认为我需要回到创始人模式来处理这件事,所以我需要能够专注于这个。”然后,你知道,这周我们之间的对话的一部分就是,“好吧,我在微软董事会已经 10 年了。这是巨大的荣誉和快乐,不仅仅是 LinkedIn,还有 OpenAI、GitHub 等等。但是,到今年年底,我真的应该过渡到创始人模式了。”所以,就像你知道的,我们会一直合作。但现在是时候重新深入公司了。

Well, yeah, so, I mean, it was one of the things that I realized over the last month was that we're seeing such progress with Manas in actually in fact we've got an internal description of move 37 for chemistry because we're seeing chemistry that might actually take shots at really interesting cancers and other things. And you know, it's very early, right? And so these things take a while with all the IDs, but the fact that we are already beginning to see like we've got some of the best computational chemists in the world. They looked and said that's very interesting. We've never seen it before and that might work, right? And so that kind of thing and so you know, Sid and Uzair and I were talking about this and I said, look, I think I need to get back to founder mode in terms of how to do this and so I need to be able to kind of focus on this and then you know, part of the conversation you and I have had this week is to say, okay, it's been 10 years on the Microsoft board. It's been a huge honor and pleasure with you know, and not just LinkedIn, but obviously OpenAI, GitHub, a whole bunch of things. But you know, at the end of the year I should really be transitioning right now to being founder mode. So like you know, it's like and you know, we'll always be working together. So it's like but you know, it's time to kind of dig back into the company.

Satya

首先,能让你加入微软董事会,与你合作,显然是通过所有的合作伙伴关系和 LinkedIn,这真是莫大的荣幸。你在微软董事会肯定会被怀念的,但我知道我们会保持紧密联系,我也为你感到兴奋。

Now, first of all, it's been such a privilege to have you on the Microsoft board, work with you, obviously through all the partnerships and LinkedIn and you will definitely be missed on the Microsoft board, but I know we'll be very connected, but I'm also excited for you.

创始人模式与AI影响 Founder Mode and AI Impact

Host

你说你要回到创始人模式,这让我更加好奇你会构建什么,因为正如你准确捕捉到的,此刻正是这些技术在社会中产生更广泛影响的关键时刻。

You know, when you say you're going back into founder mode, that means I'm even more curious about what you're going to be building because, as you rightfully captured, this is the moment where the impact of these technologies more broadly felt in society is the need of the hour.

Satya

是的,没错。

Yeah, no, exactly.

AI作为外星智能 AI as Alien Intelligence

Host

我们回到关于 AI 思维模式的问题。我想到的一点是,我们面对的是外星智能,对吧?一种被构建成高度模仿人类智能的版本,这为我们带来了很多实用性。但它的推理模式与我们的不同。人们经常遇到这种情况,然后说“哦,那很可怕”。但你会说,“不,实际上你必须注意确保对齐,但它也很棒,因为它能实现新事物。”这就是问题的一部分,比如我们在 Manus 做的化学研究,人类之前没有发现过。还有哪些关于这种推理模式的思考能够增强人类能力?

Let's come back to questions around the patterns of thinking that happen with AI. One of the things I think about is you've got an alien intelligence, right? A different version that has been built to heavily mimic human intelligence, which creates a whole bunch of utility for us. But its patterns of reasoning are not the same as ours. People frequently encounter that and go, 'Oh, that's scary.' But you're like, 'Well, no, actually you have to pay attention to make sure there's alignment, but it's also wonderful because it enables new things.' So that's part of the thing, like the kind of chemistry we're doing at Manus that humans hadn't discovered. What are the other kinds of thinking about this pattern of reasoning that amplify human capability?

Satya

这是个好问题。我并没有深入思考过它如何改善事物。如果我们说人类有归纳和演绎,那么与 AI 处于这种认知循环中是迷人的,因为我的探索空间在变化。即使在编程中,我的一位同事向我介绍的最迷人的事情之一是他写了一个新技能,我现在在 GitHub Copilot 中拥有它。它叫做认知覆盖。这很迷人。他所做的是,每当一个智能体为我做一些工作时,他说,“就像我们有测试覆盖一样,我们现在有了一个新概念叫认知覆盖,我们人类将从它做的事情中学习。”所以它创建了一个测验。因为我确实在思考从中学习,这是我匹配自己能力的过程,本质上形成对智能体所做事情的演绎理解。所以对于你的观点,这基本上是一方面的模仿游戏。但那个模仿游戏确实创造了东西,所以我演绎理解它的能力可能是我们必须发展的更重要的人类技能之一。所以我对这个认知覆盖很着迷。我认为我们将拥有类似的东西,智能体在工作,一方面必须对齐,但另一方面是我们知道我们已经认知覆盖了 AI 所做的事情。

It's a great one. I've not thought that deeply about how it improves things. If you sort of say there's induction and deduction that we do, being in this cognitive loop with an AI is fascinating because my exploration space is changing. Even in coding, one of the most fascinating things one of my colleagues introduced me to is he wrote a new skill which I have in GitHub Copilot. It's called cognitive coverage. It's fascinating. What he's done is, whenever an agent does some work for me, he says, 'Just like how we have test coverage, we now have this new concept called cognitive coverage where we as humans are going to learn from what it did.' So it creates a quiz. Because I literally think of learning from it, and it's me matching my ability to essentially form a deductive understanding of what an agent did. So to your point, it's basically an imitation game on one side. But that imitation game does create something, so my ability to then deductively understand it is probably one of the more important human skills we have to develop. So this cognitive coverage, I'm fascinated by it. I think we're going to have something like that, where agents are working and have to be aligned on one side, but the other side is us knowing that we've cognitively covered what AI did.

代币资本与AI战略 Token Capital and AI Strategy

Host

是的。我认为这一点与你之前关于 token 资本(现在叫 AI 资本)的观点相交。现代人类知识工作者的部分技能是,如何制定 AI 画布编排的策略?同时也要在资本分配的限制内。因为我们看到人们可能会疯狂地以低效方式花费大量 token,就像过去五角大楼糟糕的日子里花一百万美元买马桶座圈。所以将你的思考、认知和 token 管理结合起来。你认为关于这种将认知覆盖与 token 智能结合以放大成果的技能,问题的一部分是什么?

Yeah. I think that point intersects with your earlier one about token capital, now AI capital. Part of the skill set for the modern human knowledge worker is to say, okay, how do I strategize on the use of AI canvas orchestration? But also within kind of capital allocation bounds. Because part of what we've seen is people can go crazy spending lots of tokens in ineffective ways. It's like buying toilet seats for a million dollars back in the bad Pentagon days. So putting your thinking together with the cognition together with token management. What do you think is part of the question around this skill of blending cognitive coverage with token intelligence in terms of how you amplify the outcomes?

Satya

第一。事实上,我一直在思考,必须深入研究的一件事是,在 token 充裕的时代,哪些东西变得更有价值。一个是你提到的:如何使用 token 变得非常有价值。谁想出了如何更高效地使用 token 来实现对世界重要的成果,谁就会领先,这是必然的。那么如何建立这种直觉呢?也就是,那个成果是什么?我如何衡量它?什么是评分标准?事实上,我认为评估在这方面很迷人。需要花多少时间,尤其是在强化学习 regime 下,我认为最清楚的是:如果你真的想设定评分标准和评估维度,或者评分标准维度,使它们真正捕捉到只有你能定义的高品味。因为如果你做到了,那么你就能实现 token 效率。事实上,token 效率的另一面是,在我们甚至在 Build 大会上展示的一些例子中,假设你是一个零售商或包装商品公司,你处理的事情之一是围绕贸易促销的交易。现在,你从所有零售商那里收到这些索赔,你必须处理它们,人类在做这件事。你可以很容易地使用智能体式工作流来自动化它。你可能会说,“哦,我用一个前沿模型来做。”但我有一个格言:不要用前沿模型解决非前沿问题。这不是一个前沿问题。你不是在尝试发现新的材料科学。这是一个可重复的确定性工作流,但它可以从人类在索赔等方面可能犯的所有错误中受益。所以你需要其中的智能,但你可以使用像 MAI 5B 这样的模型,并利用轨迹在你的 RLE 中进行爬山,甚至表现优于前沿提示模型。所以对我来说,这就是 token 效率。所以那种人类对系统限制和系统特征的理解,我认为在这个阶段变得非常宝贵。

One. In fact, I've been thinking about one of the things one has to study deeply is what are all the things that become more valuable in the age of token abundance. One is what you referenced: how you use tokens becomes very valuable. Whoever figures out that they can use tokens more efficiently for an outcome that matters in the world is going to get ahead, by definition. So then how does one build the intuition for it? Which is, what is that outcome? How do I measure it? What's a rubric? In fact, this is where I think the eval is fascinating. How much time needs to be spent, especially for an RL regime, I think it's been the clearest: if you really want to set up the rubric and the eval dimensions or the rubric scoring dimensions such that they're really capturing the high taste that you only can define. Because if you did that, then you could be token efficient. In fact, the other side of token efficiency is, in some of the examples we even showed at Build, let's say you're a retailer or a package goods company and one of the things you do is handle transactions around trade promotions. Right now, you get all these claims from all the retailers and you have to process them, and humans are doing it. You can easily automate that using an agentic workflow. And you could say, 'Oh, I use a frontier model for it.' But one of the dictums I have is: don't use frontier models for non-frontier problems. This is not a frontier problem. You're not trying to discover some new material science. This is a repeatable deterministic workflow, but it can benefit from all the mistakes humans can make in claims and so on. So you do need intelligence in it, but you can take a model like an MAI 5B and use the traces to hill climb in your RLE to perform even outperform a frontier prompted model. So that to me is token efficiency. So that type of human understanding of both the limits of the system and the characteristics of the system, I think becomes high premium at this stage.

主权AI与公司类比 Sovereign AI and Company Parallels

Host

是的,我同意。我认为人们需要理解的另一件事是,正如我们在很多对话中谈到的,其他国家有很多关于主权 AI 的担忧。我认为在这里可能有用,因为我们刚刚深入讨论了公司和主权,比如公司、它们的信息、它们的员工。我们为公司做的事情与国家应该如何参与并信任 AI 未来之间有什么相似之处?

Yeah, I agree. One of the other things I think is important for people to grok, as we've been in a lot of these conversations, there's a lot of concern in other countries about sovereign AI. And I think it may be useful here, given we've just gone in depth about companies and sovereignty, like companies, their information, their employees. What are the parallels between what we're doing for companies and how countries should think about engaging and trust in the AI future?

Satya

是的,这很有趣,因为我真正转向公司的原因之一是公司遍布世界各地。

Yeah, I mean this is interesting because one of the reasons why I have really pivoted to companies is because companies exist all over the world.

主权与AI战略 Sovereignty and AI Strategy

Satya

没错,好消息是,从某种意义上说,当各国抽象地思考主权时,它们甚至可能犯下大错,最终侵蚀掉今天自然存在的、通过国内公司形成和繁荣的商业活动所体现的比较优势。因此,保护这一点才是主权的最佳形式。有时人们会想,要是能有个防火墙,或者所有数据都留在本地就好了。我不是说这些不重要,但这甚至不是关键——关键在于确保你拥有一个经济体。这意味着你需要有公司,这些公司必须在代币经济中蓬勃发展,它们需要能够构建知识产权。因此,实际上,与外部公司建立合作关系,让你获得能力,我认为比突然落后于前沿更重要。所以,我认为这也是政策制定者必须考虑基础设施的地方。毕竟,代币一头是电子,另一头是代币。你希望成为那些将电子转化为代币的机器——也就是数据中心——的最便宜、最好、最环保的生产者。这一点我认为每个国家都必须优先考虑。当然,国家也应该优先考虑是否要自己生产半导体等等,这些都是非常合理的事情。但除此之外,我认为最重要的是采纳李嘉图一直正确的观点:各国天生就有自己的比较优势,现在它们需要用 AI 来放大这种优势。这意味着,无论是小企业、大型跨国公司,还是公共部门,最好的状态是效率不断提高,并运行在前沿。对一个国家来说,最糟糕的事情是以主权的名义脱离前沿,那毫无意义,因为你会落后。但与此同时,依赖单一的前沿模型也毫无意义,因为那样你就不主权了。所以解决之道是能够说:‘不,我们将使用模型,一家公司一家公司地自行爬山。’在整体经济层面,我认为这就是方程式。

Right, that's the good news here, which is because in some sense given countries thinking abstractly about sovereignty, they can even make big mistakes where they ultimately could erode their comparative advantage that is naturally there today embodied in the commercial activity happening in the country through its company formation and thriving. So therefore preserving that is the best form of sovereignty. Sometimes I think people think, oh, if only I had a firewall or all the data was resident or what have you. You know, I'm not saying those are not considerations, but this is not about even any of that, right, which is this is about making sure that you have an economy. That means you need to have companies. That means that those companies have to thrive in a token economy. That means they need to be able to build that IP. And so therefore, in fact, having partnerships even with companies outside that give you the ability, I think are more important than sort of suddenly falling behind some frontier, right? And so, I think that therefore, I think that this is one of those places where again, even the policy makers have to think about infrastructure. Right? After all, you know, tokens, they're electrons on one end and tokens on the other end. So, you want to be the cheapest, best, and environmentally good producers of those electrons to token conversion machines called data centers, right? So that one I think is going to be very important for every country to prioritize. Of course, the country should even prioritize whether that's their own semiconductor production and what have you, right? That's all very valid things to do. But beyond that, I think the most important thing is to take what Ricardo was always right on, which is countries have by definition their own comparative advantage, and now they need to amplify that using AI. That means best thing is that whether it's a small business or whether it's the large multinational or even the public sector efficiency in the country is getting better and is operating at the frontier. The worst thing for any country to in the name of sovereignty, if they're off frontier, then that makes no sense because you're falling behind. You have to but at the same time, being dependent on one frontier model also makes no sense because then you're not sovereign. So, what's the way to solve it is to be able to say, 'No, we will use models to hill climb on our own one firm at a time.' And at an economy at aggregate, that's I think the equation.

Host

提到芯片,我觉得这很好,既涉及公司,我们正在构建 Maya 和 Cobalt,当然也与 Nvidia 和 AMD 合作良好。那么,首先,微软自有芯片的战略任务是什么?公司或国家又能从中学习到什么?

The gesture at Silicon actually I think is a good also bring up both for companies and we are building Mayan Cobalt, but also of course partnering well with Nvidia and AMD. So, first what's the strategic job of Microsoft's own Silicon? What might companies or countries also learn from that?

Satya

是的,仔细想想,这很有意思。看看 Nvidia、AMD 或 Intel 做了什么,它们构建了通用技术。实际上,看看我们今天用 GPU 做的事情,我们当然用 GPU 做模型训练和推理,但也在加速其他工作负载。事实上,Build 大会上一个令人兴奋的公告就是使用 GPU 来加速 Nvidia GPU,从而加速我们的 Fabric 数据仓库。随着智能体工作负载的出现,我们需要在所有方面获得更高性能。这正是因为通用性。Nvidia 有 CUDA,它可以作为编程模型来加速各种工作负载。实际上,我们正在使用 Nvidia 的旧芯片来加速,这在经济上对我们也有好处。我们聪明地利用整个集群,在其生命周期内持续使用,不仅用于前沿 AI,还让旧工作负载性能更好。这对 Nvidia 和我们都有利。这也说明了新工作负载的变化。AI 的新工作负载是数据并行同步工作负载(训练和推理),以及这些新的智能体运行时工作负载,它们的调用模式非常不同,而且高度恒定。三年前它们还不存在任何规模,现在却大规模存在。因此,我们有必要开始思考,甚至是在半导体之前。我们在建设数据中心,土木工程、冷却系统、机械系统都在发挥作用。实际上,我们努力确保电子以千瓦为单位进入,数百千瓦直接输送到芯片,没有任何损耗。我们甚至尽量减少数据中心内的电力分配。我们可以为这些工作负载进行极致优化,因为它们规模如此之大。这就是我们在 Maya 上所做的。例如,Maya 正在与我们的 MAI 模型和 OpenAI 模型共同设计,以获得最佳性能。我们正在设计 Cobalt,这是我们基于 ARM 的计算核心,我们利用 GitHub 的所有智能体轨迹来设计它。编码智能体的调用模式与人类应用甚至异步人类应用截然不同。因此,我们正在优化,获得巨大的延迟和性能提升。我们将成为一家系统公司,持续为新的规模化工作负载进行优化,同时利用合作伙伴的通用技术来最大化其效用。这种灵活性,我认为非常好。你的问题问得很好,各国也应该考虑这一点。如果我说‘一切都是 Maya,一切都是 Cobalt’,那可能对微软来说并不正确。但与此同时,如果你说我们对外部创新持开放态度,我们内部也会创新,我们会对一切进行基准测试,并最终以更好的经济效益为原则。

Yeah, so it's interesting if you think about it right we want it's a great way to observe it even because I what I mean if you look at what Nvidia has done or what AMD has done or what Intel has done they built general purpose technology. Which you know, like in fact when I look at what we're doing with GPUs today we're of course using GPUs to do model training model inferencing. But we're also accelerating other workloads. In fact one of the exciting announcements at Build was using GPUs to accelerate Nvidia GPUs to accelerate our fabric data warehouse, right? In fact one of the things that's happening with the agent take workloads is we need more more performance on everything. And so that's a great that's because of the general purpose nature, right? GPUs have you know, Nvidia has CUDA CUDA can be used as a programming model on top of those to accelerate a variety of workloads. In fact we're using the older chips of Nvidia to accelerate and this also works out economically for us, right? Which is a smart for us to take a fleet and keep using that fleet over a lifetime where we are not only using it for some cutting-edge AI, but we're also using it to in fact make an old workload even better performing, right? That's great for Nvidia great for us. But that also speaks to what's happened in terms of the new workloads, right? The new workloads of AI are these data parallel synchronous workloads training inference as well as these new agent runtime workloads which have very different call patterns. Very different and they're highly constant like there's this they didn't exist 3 years ago right at any scale and now they exist at scale. So, it'd behooves us to start thinking whether it's not even before we get to the semiconductors. I'm building my data center. My civil engineering is in force. My cooling system, my mechanical systems. In fact, the DC to AC that they we're trying to make sure that the electrons are coming in in kilowatts, right? Hundreds of kilowatts straight to the silicon without any losses, right? So, we're trying to minimize even the power distribution in a data center. So, we can optimize to the nth degree for these workloads because they're at such scale. And that's what we're doing with Maya, right? Maya, for example, is being co-designed with our MAI models and the OpenAI models to get the best performance out of them. We're designing Cobalt, which is our arm-based core for compute, and we're designing it for, for example, using all the agentic traces of GitHub, right? Coding, you know, the call pattern of a coding agent is pretty different than human apps or even asynchronous human apps. And so, therefore, we're optimizing and getting massive latency gains, performance gains, and what have you. And so, we're going to be a systems company that continuously to optimizes for the new at-scale workloads that we have while using general-purpose technology from our partners to maximize the utility of those. And that flexibility, by the way, I think is what really is good. And that's where I think your question was so good, which is countries should think of that, right? Which is if you really think of there's one thing that answers, right? If I said, 'Oh, everything is Maya and everything is Cobalt,' that's probably not the right thing for Microsoft. But at the same time, if you said we're open to innovation from the outside, we will innovate inside, we will in fact benchmark everything, we'll be principled about it ultimately for better economics.

Host

是的,因为创建高效的资本代币工厂,创造人类繁荣,这就是目标。

Yes, because creating the efficient capital token factories is the create human prosperity is the goal.

Satya

完全正确。并且让许多公司都能做到这一点。那么,换个话题,我知道从多次交谈中,你对现代儿童这个话题很有见解,因为我们显然正在为导师、同事等创造产品。

Exactly. And enabling many of companies to do it. So, different thread, one of the things that I know from various conversations with you that you're very thoughtful on is the topic of children in the modern age because obviously we're creating products for tutors, co-workers, etc., etc.

儿童安全与AI Child Safety and AI

Host

我们都在共同探索的一件事是,作为行业,我们应该如何引导儿童以正确的方式成长,同时确保他们的安全。微软在这方面有哪些原则,来善待下一代人类——也就是儿童?

And one of the things we're all exploring together is what should we be doing as an industry to navigate helping children elevate the right way and also keeping them safe. What are some of the principles that Microsoft is thinking about as ways to be good regarding the next generation of humanity through children?

Satya

是的,这是个很好的问题。在儿童安全方面,我们显然有很多重要的事情要做。任何数字技术出现时,我们都必须将其视为一等公民。AI 安全涉及网络、生物武器、对齐等问题,但也包括儿童安全。我们需要特别关注当前聊天机器人的一些挑战,尤其是它们与儿童的对话。我们要确保儿童拥有所需的自主权,能够按照自己的方式互动,而不是被说服。这些都非常重要。但你提到的另一个超级重要的问题是,在这个 token 丰富的世界里,做儿童意味着什么?学习应该如何进行?我们如何激励儿童?他们进入新教学体系的能力是什么?因为即使是传统方式,比如我小时候对学习的焦虑,也是机会和优质学习稀缺的产物。未来可能不再如此。所以首要之事是创造一个学习环境,让儿童从很小的时候就不对数学或科学产生恐惧。每个人天生好奇;是与世界的接触侵蚀了那份好奇心和自信。我们如何通过给予他们探索的能力来进一步发展,让他们知道没有焦虑,因为专业知识总是丰富的?真正宝贵的是你对这些专业知识的认知覆盖。如果有人在我五岁时告诉我这些,我可能会以完全不同的方式对待生活。那么,我们作为社会如何为此创造必要条件?我认为这非常重要。而且从与你的交谈中我知道,这是我们作为科技行业的责任之一。我们不能推卸。鉴于 AI 技术的普及,我们必须对下一代人类负责。既要从一开始就考虑意外后果,建立安全护栏,也要让新技术的巨大优势民主化。然后还需要结构性变革。我认为教育不能一成不变,我们也不能继续看重同样的证书。必须有所改变。

Yeah, it's a great question. There are obviously important things we have to do around child safety. When any digital technology comes out, we have to think of it as first class. There are AI safety issues around cyber, bio-weapons, alignment, but it's also child safety. One thing we want to make sure is that some of the challenges of current chatbots, especially their conversations with children, are something we need to be very mindful of. We need to ensure that children have the agency they need to interact on their own terms, rather than being persuaded. Those are very important things. But the other thing you're pulling on, which is super important, is what does it mean to be a child in a world with this abundance of tokens? How should learning happen? How should we inspire children? What is their ability to enter a new pedagogical system? Because even traditional ways, like the anxiety I had growing up about learning, are artifacts of scarcity of opportunities and good learning. That may not be true going forward. So one of the first things is creating a learning environment where children from the earliest ages don't develop phobias for math or science. Everyone is curious by nature; it's contact with the world that erodes that curiosity and confidence. How do we develop that further by giving them the ability to explore, knowing there's no anxiety because expertise is always abundant? It's really your cognitive coverage of that expertise that's at a premium. If someone had told me that as a five-year-old, I might have approached life very differently. So how do we as a society create the necessary conditions for that? I think it's very important. And I know from conversations with you that this is one of our responsibilities as a tech industry. We cannot abrogate it. Given the ubiquity of AI technology, we have to be responsible for the next human generation. Both the unintended consequences we think about from day one, building safety guardrails, and the great advantages of new technology have to be democratized. And then there needs to be structural change. I don't think education can remain exactly the same, and we value the same credentials. Something's got to change.

教皇通谕与人类尊严 Pope's Encyclical and Human Dignity

Host

我们还没机会谈到的一个话题,在这个语境下很自然,就是教皇利奥的通谕。我认为这实际上是教会方面一次卓越的领导力表现。教皇方济各在 10 或 11 年前就让我参与帮助教会讨论 AI。教会在这一问题上一直非常主动。通谕中让我毫不意外的一点是它的人文主义。人们以为这是关于宗教、关于如何祈祷的,但实际上我所接触的教会部分一直是人文主义者。你对教皇的通谕有什么看法?看到教皇介入,而且据我所知这有历史先例。我相信工业革命时期的教皇也曾就劳工状况发表过意见。所以教皇站出来捍卫我们都深切关心的事物——AI 时代的人类尊严和人类自主权——我认为这非常重要。

So one of the things we haven't had a chance to talk about yet, which is very natural in this context, is Pope Leo's encyclical. Which I thought was actually a magnificent part of leadership on behalf of the church. Pope Francis had actually gotten me engaged in helping them talk about AI 10, 11 years ago. The church has been amazingly front-footed on this. And part of what didn't surprise me at all in the encyclical was the humanism of it. People thought it was about religion, about how you pray, but actually the parts of the church I've interfaced with have been humanists. Do you have any reflections on the Pope's encyclical? To see the Pope weigh in, and from what I understand it's historical precedent. I believe the Pope at the time of the Industrial Revolution also weighed in on the condition of labor. So to have the Pope come out in defense of what we all deeply care about—human dignity and human agency in the age of AI—I think it's so important.

Satya

我很高兴他倡导了他认为重要的事情。对我来说,另一面是社会会变成什么样?如果看看一些最伟大的技术进步是如何被用来创造巨大繁荣的,西方的故事相当不可思议。Joel Mokyr 和其他几位作者合著了一本很棒的书,叫《两条繁荣之路》,描述了中国和西方上千年的历史。从根本上说,西方在如何利用科学革命、工业革命方面的一些文化和社会结构,要求社会组织方式发生真正的变革,以便能够利用这个新时代。这定义了现代世界。所以我觉得我们现在需要的是来自道德哲学方面的类似融合。教皇基本上是在说:“这需要成为指引我们前进的道德哲学。” 将其与市场、民主以及科学/技术革命结合起来。如果我们能进入一个良性循环,道德、科学突破、政治体系和市场相互加强,那么我们就会拥有富足,许多利益相关者都将受益。

I'm glad that he has advocated what he thinks is important. The other side of this to me is what is the society like? If I look at how some of the greatest technological advances were harnessed to create great prosperity, the story of the West is pretty unbelievable. There's a beautiful book by Joel Mokyr and a couple of other authors called 'Two Paths to Prosperity', which describes the thousand-year history of China and the West. Fundamentally, some of the cultural and societal constructs in the West on how to use the scientific revolution, the industrial revolution, necessitated a real change in how society was organized so that it could take advantage of this new era. That defined the modern world. So one of the things I feel we now need is a similar coming together from moral philosophy. The Pope basically said, 'This needs to be the moral philosophy that guides us going forward.' Combine that with what is the market, what is democracy, and what is the scientific/technological revolution. If we can get into a virtuous cycle where morality, scientific breakthroughs, the political system, and the markets all reinforce each other, then we will have abundance, and many stakeholders will benefit.

社会许可与正向循环 Social Permission and the Positive Cycle

Satya

如果不这样做,我们就会失去社会许可,对吧?所以,我不认为西方成功的狭隘理解仅仅是技术突破,我觉得事实并非如此。这是多种力量在一个良性循环中不可思议的汇聚。它也有非常糟糕的部分,对吧?我们知道这一点,所以必须避免。教皇本人在通谕中也写到了一些内容,我觉得很棒,就是思考哪些是坏的部分,如何不重蹈覆辙,以及如何倡导这种良性循环。我认为这被优美地捕捉到了。

And if you don't, we are going to lose social permission, right? So, I just don't think this is what the narrow understanding of the success of the West as just a technological breakthrough. I think is sort of not the case. It's an unbelievable coming together of a multitude of forces in a virtual cycle. It also had really bad parts to it, right? We know that, and so you have to avoid that. That was also some things that the Pope himself wrote in the encyclical, which I thought was great, which is even to think about what are the bad parts, how do we not repeat it, what is the way to be advocating for this positive cycle. I think it's beautifully captured.

Host

我觉得很棒的是,它聚焦于如何将人类置于中心,如何提升人类尊严,如何不仅满足富裕国家、而是整个世界的需求。我认为这是我们开始思考这个问题的一个伟大灯塔。

And I thought it was great that it was kind of a focus on how do you keep humans at the center, how do you elevate human dignity, how do you address needs not just of the wealthy countries but the entire world. I think it's a great beacon in terms of how we started to think about it.

Satya

我对此思考很深。例如,我一直深切关心全球南方,在某种意义上,他们终于迎来了真正追赶式增长的时刻。所以我认为在 AI 时代,存在一个真正的危险,即这种趋同增长甚至会放缓,甚至走向反面。那么全球结构是什么?因为顺便说一句,坐在美国帕洛阿尔托的某个人可能认为这不会影响他们,但根本没有这回事。我们共享这个星球,我们的命运比我们想象的更加紧密相连,仅仅因为远离正在发生的事情。

I think deeply about that. For example, I've always been someone who cared deeply about the global south in some sense finally having their moment where they can be real catch-up growth. So I think there's a real danger now in the age of AI for even that convergence growth to slow and in fact go the other way. So what are the global structures? Because by the way, someone sitting in the United States in Palo Alto may think that somehow it doesn't impact them, but there's no such thing. We share this planet and our destinies are a lot more tied than we think, just because of being far away from what's happening.

应对AI反弹 Addressing AI Backlash

Host

好吧,在我们进入快速问答环节之前,最后一个问题。正如你我都知道的,在美国和欧洲,现在有很多对 AI 的抵制。我试图解决这个问题的方式之一是通过像《超级能动性》这样的书,试图说:“不,不,这是一个获得能动性的机会。转型会很困难,但拥抱能动性和变革最终是必须的,但如果你有远见并积极投入,它会非常有帮助。”你认为我们应该如何帮助美国和西方的人们理解,为什么实际上拥抱 AI 更重要?这并不意味着抵制中没有真正的问题,显然工作转型等等都是真实的问题,但我们如何帮助人们看到这可能是他们未来的重要部分?

Well, last question before we get to rapid fire. As you and I both know, in the US and in Europe there's a lot of AI backlash now. Part of how I've been trying to address it is with books like Super Agency, trying to say, 'No, no, this is an opportunity to gain agency. The transition will be difficult, but embracing the agency and the transformation ultimately you have to, but if you do it with forethought and leaning into it, it can be greatly helpful.' What do you think we should be trying to help people in the US and the West understand about why it is that actually embracing AI is more important? And it doesn't mean that there aren't genuine issues in the backlash and obviously work transition all the rest will be real issues, but how do we help people see that this could be an important part of their future?

Satya

是的,我认为至少现在我得出的结论是,需要有真正切实的成果来说话。因为我认为已经发生的事情,坦率地说,甚至我们行业谈论它的方式,当你出去说:“嘿,所有经济机会都会从知识工作者手中消失。”或者“白领工作没了。”然后你说:“我对构建这项技术感到兴奋。”谁会希望你成功?我的意思是,我不希望你成功。这完全没有社会意义。所以现在我觉得,当有人在大学毕业典礼上因为说 AI 是一种手段而被嘘时,我们已经到了人们不再相信我们的地步。而且这是理所当然的。因此,我认为我们现在需要做的是努力工作的时刻。努力工作意味着,如果你在建设一个数据中心,要确保那个社区相信这个数据中心对他们有好处。它有利于他们的税基、社区努力、房地产价值、学校、用水、电价。不能再像“哦,我说了什么”那样。不,它必须是真实的。这就是我们赢得社会许可的方式。就业方面也是如此。我们不能抽象地说:“嘿,劳动力总量谬误,总会有新工作的。”新工作是什么?新工作的工资是多少?我现在可以去申请、培训自己,以及如何真正创办一家新公司等等?我认为除非我们真正明确。第三点我们讨论了很多,每个公司都需要参与前沿生态系统。这不是“哦,我只是某个基础模型的数据提供者”。那是在谈论主权和尊严同时丧失,对国家、社区和公司都是如此。所以我认为现在我们必须全力以赴,说,好吧,这是一个正和游戏。这实际上是某些技术的挑战。我们将真正积极地应对。我们可以阐明切实的好处,而这也是你在 Manas 和其他地方所做的将非常有帮助,因为世界需要更多的证据点,证明这项技术最终是在帮助人类状况和我们的社会,广泛地,而不是狭隘地。

Yeah, I think at least now I've come to the conclusion that there needs to be real tangible outcomes that speak for themselves. Because I think what has happened, and this is one of those places where quite frankly our industry even, the way we have talked about it, I mean when you go out and say, 'Hey, all economic opportunity will go away for knowledge workers.' or 'White collar jobs are gone.' And then you're saying, 'I'm excited about building that technology.' Why would anyone want you to be successful? I mean, I don't want you to be successful. This just makes no social sense. So I feel now when you have someone in a college commencement be booed because they're saying AI is a means we've now crossed over to people don't believe us. And rightfully so. So therefore I think what we need to do now is time to do the hard work. The hard work is if you're building a data center, let's make sure that that community believes that this data center is great for them. It's for their tax base, for their community efforts, their real estate value, their schools, their water use, their electricity prices. It can't be again like, 'Oh, I said something.' No, it has to be real. That's kind of what the way we earn social permission. Same thing I would say with employment. We can't abstractly even say, 'Hey, lump of labor fallacy, always there's going to be new jobs.' What are the new jobs? What are the wages of the new jobs? That I can now go apply for, train myself for, and how do I really start a new company or what have you? I think unless we really get clear. The third thing which we unpacked a lot, every firm needs to participate in the frontier ecosystem. It's not like, 'Oh, I'm just a feeder of data to some foundation model.' That is like talk about sovereignty and dignity both being lost, simultaneously, for countries, communities, and companies. So I think now we have to go all the way and say, okay, this is a positive sum. This is in fact the challenges of some of the technology. We are going to really actively work it. We can articulate the tangible benefits and this is where again, what you will do at Manas and others are also going to be very, very helpful because the world needs more proof points that this technology is ultimately helping human condition and our societies broadly, not narrowly.

Host

是的。不,AI 为了人类。

Yeah. No, AI for humanity.

快问快答 Rapid Fire Questions

Host

所以,你不必快速回答,但我们向所有嘉宾都问同样的问题。第一个问题是,有没有一部电影、一首歌或一本书让你对未来充满乐观?

So, you don't have to answer rapidly, but we ask the same questions of all our guests. So, the first one is, is there a movie, song, or book that fills you with optimism for the future?

Satya

是《通往繁荣的平行路径》,好吗?因为我喜欢它的原因是它是一个很好的行动号召。未来一千年,通往繁荣的道路是什么?至少世界的一部分在过去一千年里做对了。我们现在能否作为一个完整的星球,在未来一千年里做对?我认为这正是我们必须做出一些最出色工作的地方。

It's this parallel paths to prosperity, all right? Because the reason why I like that is because it's a good call to action. For the next thousand years, what's the path to prosperity? There was a blueprint at least that part of the world got it right in the last thousand. Can we now as an entire planet get it right for the next thousand? I think that this is where some of our very best work has to be done.

Host

同意。你希望人们更常问你什么问题?

Agreed. What's a question that you wish people would ask you more often?

Satya

啊,这是个好问题。我希望人们问我,我对什么不感到兴奋。因为我对很多事情感到兴奋。但我不兴奋的是,比如我们因为说错话或做错事,甚至没有完整思考如何真正实现正和构建,而失去对 AI 的许可。

Ah, that's a good one. I would love for people to ask me what am I not excited about. Because I'm excited about a lot of things. But I'm not excited, for example, about us losing permission on AI by saying all the wrong things or doing the wrong things even and not having a complete thought on how to truly have a positive sum construct.

Host

是的,100%。那么,你在你的行业之外看到哪些进展或势头?鉴于微软为世界许多行业提供动力,可能是机器人、AI 等。但你在哪里看到让你受到启发的进展?

Yep. 100%. So, where do you see progress or momentum outside of your industry? And given that Microsoft powers a lot of the world's industry, it could be robotics, AI, etc. But where do you see progress that inspires you?

Satya

在某种程度上,生物学领域的工作。如果我想一想,我们必须更好理解的最复杂系统之一就是人类生物学。所以任何能够帮助人类照顾人类的工具,都可能是令人敬畏的。

I mean to some degree, the work in bio. If I think about it, one of the most complex systems that we have to have a better understanding of is human biology. So anything, any tool that can help humans take care of humans is probably the thing that will be awe-inspiring.

GigaTime:降低免疫疗法测试成本 GigaTime: Reducing cost of immunotherapy testing

Host

我们最近发现了一件事。对于免疫疗法,我认为有一项复杂且昂贵的测试,用于判断某种免疫疗法是否对该肿瘤有效。Providence 与华盛顿大学和微软研究院的一些研究人员合作,构建了一个名为 GigaTime 的模型。它基本上模拟了这项任务,降低了原本需要大量时间和金钱才能完成的成本。现在,任何城市的任何三级医院都能完成。这种医疗和经济上的可及性,简直是一个突破。

There's this one thing we recently came across. For immunotherapy, I believe there's a test that is a complex, costly test that figures out whether a particular immunotherapy will work on that tumor or not. Providence and some researchers at UW and Microsoft Research came together and built this thing called GigaTime, which is a cool model. It basically simulates that task and reduces the cost of what could only be done with a lot of time and money. Now it can be done by any tertiary hospital in any city. That type of economic availability of medicine and medical practice is just a breakthrough.

梦想:10%GDP增长与全球合作 Dream: 10% GDP growth and global cooperation

Host

最后一个问题。如果未来 15 年一切顺利,你认为人类可能实现什么?我们的第一步是什么?

Last question. Can you leave us with a final thought on what you think is possible to achieve if everything breaks humanity's way in the next 15 years? And what's our first step?

Satya

如果一切顺利,我始终梦想着世界以 10% 的 GDP 增长率复利增长。这就是可能发生的事。一个思想实验是:如果工业革命同时触及世界各个角落,每个国家都能充分发挥其比较优势。这是最大化的正和博弈。我们能实现吗?因为我们受困于这样的想法:世界不是这样运作的,历史关乎统治和霸权。我不是说我们会完全颠覆这一切,但既然你让我做梦,我梦想着如果人类能够超越有限理性,不再说“哦,历史重演,我们永远无法摆脱战争”,而是说“也许我们可以改变进程”。第一步就是接受这种可能性存在。否则,我们会回到重新争论历史,并试图与之匹配。

If everything breaks our way, I've always gone back to the dream of the world compounding at 10% GDP growth. That is what can happen. One thought experiment is if the industrial revolution had reached all corners of the world at the same time, and every country could express their comparative advantage fully. That's the maximalist positive sum construct. Can we do it? Because we are captive to this idea that that's not how the world works, that history is about dominance and dominant powers. I'm not saying we will defy all that, but since you asked me to dream, I'm dreaming that if humanity can get past their bounded rationality and stop saying, 'Oh, history repeats itself, we are never going to get better than fighting wars,' and say, 'Maybe we can change the course.' The first step would be to accept that that possibility exists. Otherwise, we'll go back to re-litigating history and trying to match it.

闭幕致辞 Closing remarks

Host

美好的梦想。Satya,总是很愉快,下周董事会见。

Beautiful dream. Satya, always a pleasure, and I'll see you next week at the board meeting.

Satya

非常感谢,Reid。非常感谢。

Thank you so much, Reid. Thank you so much.

Host

太棒了。非常有趣的对话。

This is awesome. Such a fun conversation.

Satya

是的,没错。

Yes, exactly.

致谢 Credits

Host

《Possible》由 Pallet Media 制作。主持人是 Ara Finger 和我 Reid Hoffman。我们的节目统筹是 Shawn Young。《Possible》由 Tanaz Deelo、Katie Sanders、Spencer Strassmore、Imo Zoo、Aman Souri、Lexi Kevin、Danny Garrison、Trent Barbosa 和 Tafadzwa Nemarundwe 制作。特别感谢 Surya Yalamanchili、Saida Sabiyeva、Ian Ellis、Greg Biato、Parth Patel 和 Ben Ralis。

Possible is produced by Pallet Media. It's hosted by Ara Finger and me, Reid Hoffman. Our show runner is Shawn Young. Possible is produced by Tanaz Deelo, Katie Sanders, Spencer Strassmore, Imo Zoo, Aman Souri, Lexi Kevin, Danny Garrison, Trent Barbosa, and Tafadzwa Nemarundwe. Special thanks to Surya Yalamanchili, Saida Sabiyeva, Ian Ellis, Greg Biato, Parth Patel, and Ben Ralis.

互动版:逐字朗读 + 针对本期提问 →