萨提亚·纳德拉谈 AI 安全、对齐局限与微软接班

你完全可以去跟美国总统聊。是他命名的，而且在我看来，这名字跟别的名字一样好。

You're welcome to talk to the president of the United States. He named it, and I mean, it's as good a name as anything as far as I'm concerned.

这就是微软 CEO 萨提亚·纳德拉谈把人工智能改名为超级智能。那么，这项技术到底该怎么称呼才合理？我们在本周的《Future Proof》里聊了这些，还有很多别的内容。

That's Microsoft CEO Satya Nadella talking about the renaming of artificial intelligence to superintelligence. And what does make sense to call this technology? We got into that and a lot more on this week's Future Proof.

我去了西雅图，在他们的年度 CISO 峰会上跟纳德拉聊了聊。这是所有负责保障组织安全的人的一次聚会，而安全最近一直是科技行业的大话题，因为我们已经看到了一些 AI 模型失控的例子。他在那儿对听众讲了一番相当引人注目的话。

I went up to Seattle to talk with Nadella at their annual CISO Summit. This is a gathering of all the people responsible for keeping their organizations secure and safe, and that's been a big topic lately in the tech industry as we've seen examples of AI models going off the rails. And he had a pretty striking message for the audience there.

所以至少在我看来，我们的工作会是：你怎么构建一个前沿生态？你怎么保护——怎么确保各种形态都能在这个知识分布化的新前沿生态里繁荣？因为否则就不会有经济。这算是第二点。然后我们现在还要应对一件新事情：如果对齐难以实现，那意味着 AI 安全也某种程度上难以实现，而机制性理解也某种程度上并不存在，那我们该怎么处理安全这一面？所以这其实都是同一条线，并不是三件彼此独立的事，至少我们的工作是这样的。

So to me, at least, our body of work would be: how do you build a frontier ecosystem? How do you protect — how do you make sure all forms can thrive in this new frontier ecosystem with distributed knowledge? Because otherwise there won't be an economy. So that's kind of the second. And then we now have this new thing to deal with, which is: if alignment is elusive, and that means AI safety is sort of elusive, and mechanistic understanding is sort of not there, then how do we deal with the security aspect? So it's all really one thread. It's not really three distinct things, and that's at least what our work is.

最让人意外和有意思的，是听萨提亚讲他原本多少希望我们能找到办法让这些模型对齐，找到办法理解它们——他称之为机制性可解释性，就是一直理解到单个神经元层面。他现在基本上暂时放弃了这些想法。他说，我们也许永远做不到这些，所以我们必须做的是把这些模型当作内部威胁来对待。我们不能再信任它们了。最终要落到建护栏、建安全、建沙箱上。而这正是微软真正擅长的事。它不是一家 AI 研究实验室。萨提亚说的是：“我们要通过为这些模型搭建基础设施、规则和安全来促成创新。”

What was so surprising and interesting was listening to Satya talk about how he had sort of hoped that we would figure out a way to align these models, we'd figure out a way to understand them — something he refers to as mechanistic interpretability, just understanding these models down to the individual neuron level. He's essentially sort of given up on those ideas for now. He's saying, you know, we may never be able to do these things, and so what we have to do is treat these models as insider threats. We cannot trust them anymore. It's going to come down to building the guardrails, building the security, the sandboxes. This is the kind of thing that Microsoft really shines at. It's not an AI research lab. You know, what Satya is saying is, "We're going to enable innovation by building the infrastructure, the rules, the security around these models."

我们还聊了这些 AI 模型的对齐与可解释性、微软作为一家公司的未来，以及纳德拉自己的未来——他 59 岁了——他对可能由谁来接棒说了一些很有意思的话。我们稍作休息，马上回来继续和萨提亚聊。

We also talked about the alignment and interpretability of these AI models, the future of Microsoft as a company, and Nadella's future — who's 59 years old — and he had some interesting things to say about who might take over the reins. We're going to take a quick break, and we'll be right back with Satya.

《Future Proof》由 Google Cloud 呈现。智能体不是本该让你省心吗？为了安全运行你的 AI 智能体，却要不停给遗留基础设施打补丁，这也太过了。如果你的企业安全策略依赖每月成千上万的补丁，那你最大的威胁不是未来，而是“补丁星期二”。Gemini 运行在最安全的云上，具备先进的威胁检测，能在攻击演变成入侵之前将其掐断。真的。用 Google Cloud 上的 Gemini，让工作少点折腾。查看回复。需订阅。

Future Proof is presented by Google Cloud. Weren't agents supposed to put your mind at ease? Constantly slapping duct tape on legacy infrastructure just to run your AI agents safely is a bit much. If your business's security strategy relies on thousands of patches every month, your biggest threat isn't the future. It's Patch Tuesday. Gemini runs on the most secure cloud with advanced threat detection to shut down attacks before they become breaches. No, really. Make work less work with Gemini on Google Cloud. Check response. Subscription required.

萨提亚，很高兴见到你。我们现在在微软 CISO 峰会上，你在这里解释说，对这些人来说这是 24 小时全天候的战时状态。安全是他们心头的大事。你一直是这项 AI 技术的大力倡导者。我觉得没有你，就没有 ChatGPT。但你最近写了这篇博文，我觉得里面有一丝挫败感，你谈到这些东西也许没法像我们曾经以为的那样被对齐。我们需要把 AI 模型当作内部威胁，真的把它们锁起来。那是一篇很精彩的博文。但我有个问题。你有没有收到任何 AI 实验室的反对意见，尤其是 OpenAI？因为他们觉得也许他们仍然能让这些东西对齐。也许我们能搞清楚——

Satya, it's great to see you. So, we're here at the Microsoft CISO Summit, where you explained that it's wartime 24/7 for these people. Security's on the mind. Now, you're a big champion — you've been a big champion of this technology of AI. I think without you, ChatGPT doesn't happen. But you wrote this blog post recently that there's a little bit of a sliver of defeat in this thing, I think, where you're talking about how these things maybe can't be aligned like we once thought. And we need to treat AI models as insider threats and really kind of lock them up. It was a fascinating blog post. But I have a question. Did you get any pushback from any of the AI labs, particularly OpenAI? Because they think maybe they still can align these things. Maybe we can figure out —

是的，首先，在 CISO 峰会上跟你聊真是太好了，因为这显然是当下、甚至此刻的头号话题。你知道，当我想到那些实验室，对吧？甚至——记得 OpenAI 还是一个和乐大家庭的时候，连达里奥都在那儿工作，还有山姆和格雷格，他们有一个根本信念，我当年也和他们一样，那就是缩放定律是创造新能力时该押注的东西。

Yeah, first of all, it's wonderful talking to you at a CISO summit because this is obviously the topic of the day and of the moment even. You know, when I think about whether it's the labs, right? Even when — remember OpenAI when it was one happy family with even Dario working there, and obviously Sam and Greg, and what the fundamental belief they had, that I had along with them, was that scaling laws are the thing to bet on to create new capability.

好消息是，缩放定律奏效了，而且奏效得很神奇，比我在 2018、2019 年能想象的要好得多、也多得多，一路走到今天。但与此同时，我觉得到了这个阶段，如果你看看——不管是 OpenAI 的人，还是 Anthropic 的人，还是别处的人，事实上还有微软的人，所有在经验层面最贴近这件事的人，在经历了缩放定律这一切之后——对吧？它不是一条定律，你知道，我们把它当作定律来谈。它更像是一个经受住时间检验的经验性发现，就像摩尔定律。

And the good news here is the scaling laws have worked and worked magically, a lot better and a lot more than I would have ever imagined in 2018, 2019, to where we are. But at the same time, I think at this point, if you look at even what — whether it's folks at OpenAI or folks at Anthropic or elsewhere, and in fact at Microsoft, all of the folks who are closest to this empirically, after all of what scaling laws — right? It's not a law, you know, we talk about it as a law. It's more of an empirical finding that's withstood the test of time like Moore's law.

结论是：是的，它奏效，但对齐似乎难以实现，甚至机制性理解也很难。我不是说这些会永远难以实现，但今天它们确实如此。我希望我们能找到解决方案。但这是一门实验科学。我觉得我们应该认真对待这一点。

Is that, yeah, it works, and then alignment seems elusive, or even mechanistic understanding is tough. And I'm not saying these are forever going to be elusive, but they are today. I hope we find the solutions. But this is an experimental science. I think we should take that point seriously.

于是你说：“好，我们就是这个处境。缩放定律奏效，能力不断跃升，但我们把对齐和安全保障内建进去的能力似乎并不存在。也许永远不会存在。”如果是这样，那我们该怎么办？那你就得走到模型外面去。这大致就是我那篇博文的重点：如果我看这项工作——事实上 CISO 峰会讲的就是——Microsoft Agent 365、感知，以及一种架构思路，把控制层和保障层放在模型之外，使用多个模型，迫使这些模型变得不那么不透明，诸如此类。这就是它的动机。

So then you say, "Okay, that's where we are. Scaling laws work, capabilities keep jumping, but our ability to have alignment and safety assurance just be built in seems not there. Maybe never going to be there." If that is the case, then what do we do? So then you've got to come outside the model. And so that's a little bit of what my blog post was all about, which is: if I look at the work — in fact, the CISO summit is all about — Microsoft Agent 365 and perception and an architectural approach to having the control layer and the assurance layer outside the model, use multiple models, force these models to actually be less opaque, and what have you. And that is the motivation for it.

因为你得把它当作——至少这是我喜欢这个比喻的原因，因为至少当你面对某种非常强大的东西时，我们处理过这种情况。我是说，组织内部一直都有能做出恶意行为的人，你试图通过把内部风险当作必须考虑的一件事来确保这不会发生。而现在，你实际上有了这些在内部的智能体，它们具备成为内部风险的全部能力。所以，这就是我用这个比喻的原因。

Because you've got to treat — and that's why I like the metaphor, at least, because at least when you have something very powerful, we've dealt with this. I mean, you've always had people inside an organization who could do malicious things, and you tried to make sure that that doesn't happen by thinking of the insider risk as a thing that you account for. And so now you have effectively these intelligences that are inside, that have all the capability to be an insider risk. And so therefore, that's why I use that metaphor.

就像你在里面说的，这些东西越大、越强大，可解释性的问题就变得越难。

It — you know, as you said in there, the bigger these things get, the more powerful they get, the interpretability problem gets even more difficult.

制造一个能做任何事的超级智能，如果你只需要把它锁起来、放在一个筒仓里以确保它不会在你的组织内部等地方造成损害，那有什么意义？

What's the point of making a superintelligence that can do anything if you just need to lock it up and keep it in a silo to make sure it doesn't cause damage inside your organization, etc.?

嗯，我的意思是，这就是关键所在，对吧？也就是你能否获得好处而没有坏处？所以这就是为什么你有控制系统，以便你可以控制。对吧？就像，所以想法是提出我们已有的控制、治理和安全的工程实践。对吧？我的意思是，如果你想一想，我们今天如何管理人类内部风险？你管理它的方式是拥有对资源的最小特权访问，对吧？你有身份，你有可观测性，以及最小特权。

Well, I mean, that's sort of the point, right? Which is can you get the benefits without the downsides? And so, that's why you have control systems, so that you can control. Right? It's like, and so the idea is to come up with what's the engineering practice of control and governance and security that we've had. Right? I mean if you think about it, how do we manage inside risk today for humans? The way you manage it is by having least privileged access to resources, right? You have identity, you have observability, and at least privilege.

但不是超级智能，对吧？我觉得

Aren't super intelligent, right? Like I think

为什么同样的控制机制不适用呢？我不是说它们不像所有人都有不同的智力，你知道，他们也有不同的权威。嗯，所以因此我们所处理的是我们处理内部风险的一些基本原则方法可以应用。有一个真正的挑战，从某种意义上说，当我使用一个模型及其轨迹时，嗯，如果有各种新发布，呃，你知道，我不理解它的行动空间。它为什么调用那个工具？然后就变得有问题了。这就是为什么我提倡标准，这样当，你知道，首先没有新发布，每个人都真正透明地暴露他们的 COT 或用英语的思维链，并且你使用多个模型，而不是一个模型。你用一个模型检查另一个模型。然后你甚至把这些确定性系统放在外面，以理解甚至因果推理。为什么这个东西调用这个工具，呃，当我真的它没有特权，对吧？为了能够理解那里出了什么问题。所以那种控制机制，我们在其他工程系统中处理过这类事情。嗯，所以我希望我们能带来同样水平的实践。

Why wouldn't it apply even like the same control mechanism. I'm not saying that they're not like even at inside of like all people have different intelligences too and they you know and they and so and different authority even. Um and so therefore what we have dealt is some of the basic principled ways we've approached things like inside risk can be applied. There is a real challenge in the sense that when I'm using a model and its trajectory, um if there's all sorts of new release, uh you know, I don't understand its action space. Why did it call that tool? Then it becomes problematic. That's why I'm advocating for I'd say standards so that when, you know, first of all there's no new release, everybody's sort of really exposing transparently their COT or the chain of thought uh in English and you use multiple models, not one model. You use one model to check on the other model. Then you try to put even these deterministic systems outside uh to understand even the causal reasoning. Why did this thing call this tool uh when I really it didn't have privileges, right? To be able to understand what was the problem there. And and so that type of control mechanism, we have dealt with these types of things in other engineering systems. Um and so I my hope is that we can bring that same level of practice.

听起来我可能理解错了，但听起来你实际上试图做的事情之一是建立一种机制可解释性的代理。就像如果你不能理解到单个神经元水平，也许有另一种方法可以做到。那是

It sounds like I could be getting this wrong, but it sounds like what you're actually one of the things you're trying to do is build sort of like a proxy for mechanistic interpretability. Like if you can't understand it down to the individual neuron level, maybe there's some other way to do it. Is that

我们把这两件事分开，对吧？其中一个是，我认为我们应该继续做所有研究以达到机制可解释性，你知道，探针，或其他 AI 安全的东西。那是本身就是一个科学研究议程。所以假设我们继续做那个。但我们现在在这里部署更强大的模型。你今天做什么来获得保证？这就是我提倡的架构方法。这些并不是互斥的，对吧？所以这些你知道，它们都必须并行发生。我想让我们真正确保我们不会部署，事实上，我的一大恐惧是人们部署嵌套的黑盒子。我们失去了，忘记了控制，而是保证。就像对我来说有一个此时此地的简单问题。一家公司如何证明当一个智能体代表他们在我的组织内部做事时，它做了什么？我认为唯一的方法是能够拥有所有这些日志，首先，一切身份，一切被观察，一切行为被监控，一切你实际上可以证明和治理的东西。

We keep these two things separate, right? Which is one is I think we should continue to do all the research to get to mechanistic interpretability, you know, probes, whatever other things that are AI safety. And that's a science research agenda on its own right. So let's say we keep going on that. But we are deploying more powerful models here and now. What do you do to have assurance today? And that is the architectural approach I'm advocating. And these are not sort of mutually exclusive, right? So these are you know, they all have to happen in parallel. And I want us to sort of really make sure that we're not deploying in fact, my one big fear is people deploy nested black boxes. And we lose, forget control, but assurance. Like what there's a here and now simple question for me. How does a company attest to what sort of it did when an agent did stuff inside of my organization on their behalf? I think the only way to do that is to be able to have all of these logs, first of all, everything identity, everything observed, everything behaviorally monitored, everything that you can actually sort of attest to and govern.

我看着你在做的事情，然后我看着像 Meta 的 Muse 这样的东西，对吧？他们要求，他们真的在向消费者营销，并要求他们交出大量数据和控制权给模型。你认为消费者应该涉足这个吗？他们应该信任像 Muse 这样的东西吗？

I look at what you're doing and then I look at something like Meta's Muse, right? Where they're asking they're really kind of marketing to consumers and asking them to turn over a massive amount of data and control to models. Do you think consumers should dabble in this? Should they trust something like Muse?

我的意思是，无论是组织还是消费者，最终这些系统必须赢得信任。我认为我们会接受不同的行为方式。如果我看看组织，这是我目前在这个上下文中的主要焦点，组织将拥有一个企业架构，我认为这将允许他们对这些有保证和治理。所以，那是，我认为，我的意思是，从架构上讲，我知道如何解决那个。消费者空间，当人们信任某件事时，这意味着他们依赖提供商做所有的工作，对吧？所以，但是当我们考虑 Copilot 与 Autopilot，这是相同的功能，这取决于我们确保我们部署的是我在企业中为每个使用它的消费者部署的相同的东西。然后给予，我必须通过确保任何智能体所做的任何事情都不是有害的来赢得客户的信任。

I mean, whether it's organizations or consumers, at the end of the day these systems have to earn the trust. And I think there different sort of behavioral things that we'll accept on places. If I look at the organization, which is kind of the primary focus I have right now on this context, is organizations are going to have an enterprise architecture that I think will allow them to have assurance and governance on these. And so, that's I think a I mean, I would say architecturally, I know how to solve that. The consumer space, when people trust something, that means they're relying on the provider to do all the work, right? So, but when we think about Copilot with Autopilot, which is the same functionality, it's up to us to make sure that we are deploying the same thing that I'm deploying in the enterprise on behalf of every consumer who is using it. And then give and I have to earn that trust of the customer by ensuring that any agent that does anything is not doing anything harmful.

我们要稍作休息。我们马上回来，继续与 Satya 的对话。

We're going to take a quick break. We'll be right back with Satya.

Future Proof 由 Google Cloud 呈现。智能体不是应该帮你分担工作吗？再建一个来处理其他的有点过分。不要让工作变得更工作。Gemini 是一个了解你业务所有部分并真正完成工作的智能体。你给它一个项目，合上笔记本电脑，当你回来时，工作完成了。无论你的行业或角色，你的智能体都有技能帮助每个团队保持流畅。说真的。用 Google Cloud 上的 Gemini 让工作更少工作。检查响应。需要订阅。

Future Proof is presented by Google Cloud. Weren't agents supposed to take things off your plate? Building another to handle the others is a bit much. Don't make more work out of work. Gemini is one agent that knows all parts of your business and actually gets stuff done. You give it a project, close your laptop, and when you come back, the job's finished. No matter your industry or role, your agent has the skills to help every team stay in their flow. Seriously. Make work less work with Gemini on Google Cloud. Check response. Subscription required.

你知道，你说你专注于企业，但我记得微软曾经是一个消费品牌。我的意思是，你认为你会以像 Muse 这样的产品的方式去追求消费者吗？

You know, you said you're focused on the enterprise, but I remember when Microsoft was a consumer brand. I mean, do you think that would you ever go after consumers in the way that, you know, product like Muse or something?

会。所以，我认为这就是消费者是一个广泛类别的地方。嗯，你知道，如果我看看今天的 Office 和 Office 365，Office 365 消费者版，你知道，事实上，它是最大的消费者订阅之一，拥有超过 1 亿订阅者，他们喜欢在工作和家庭中使用他们所用的东西来做税务和高风险的个人生产力和工作。所以，我们现在有 Copilot，事实上，我们现在对新的 Copilot 发布非常满意的一件事是它统一了，事实上，你在家里可以使用的和你在工作中可以使用的是完全相同的产品。事实上，我所有的笔记本、项目，所有相同的组织能力，无论是在 OneDrive 个人版还是工作中。所以，我们的目标将是追求那些你会想要雇佣微软的东西。我认为大多数人不会，我认为 Meta 是一家伟大的公司，你知道，苹果是一家伟大的公司，谷歌是一家伟大的公司，这些新公司也是如此，但所有这些公司都有不同的品牌许可。

Do. And so, I think that this is where consumer is a broad category. Um and, you know, if I look at Office today and Office 365, Office 365 consumer is on you know, in fact, it's one of the largest consumer subscriptions out there with 100 plus million subscribers who love using what they use at work and at home to do their taxes and their high-stakes personal productivity and work. And so, we have now Copilot, in fact, one of the things that we now I'm very happy with the new Copilot release is it unifies, in fact, what you can use at home and what you can use at work is exactly the same product. I, in fact, have all my notebooks, projects, all of that same organizational capability for whether it's on OneDrive personal or at work. And so, I our goal would be to go after things that you would want to hire Microsoft. I think most people don't I think Meta is a great company, you know, Apple's a great company, Google's a great company, so are these new guys, but all of these companies have different brand permission.

当人们想到微软时，我觉得大家都会认为这是零和博弈。也许有些是，但我也认为我们都有一个品牌许可，消费者实际上非常挑剔。当你想到微软时，你会想到信任，你会想：“哦，这里是我的数据会安全的地方。这是我报税的地方。这是我做高风险项目工作的地方。”所以，对我来说，这就是我们想要擅长的，对吧？我不想——我觉得微软在最好的时候——我总是开玩笑说微软最大的热门消费产品是 Excel。想想看：Excel 是一个叫做电子表格概念的消费化，它赋予了数十亿人获得数字感的能力，对吧？我的意思是，这就是我们。所以，我希望能够用 Copilot 做到这一点，这就是我们的目标。

When people think Microsoft, I think we all think it's all zero-sum. Maybe some of it is, but I also think that we all have a certain brand permission that consumers, in fact, are pretty discerning. And when you think Microsoft, you think trust, you think, "Oh, this is the place where my data will be secure. This is where I do my taxes. This is where I do my high-stakes project work." And so, to me, that's what we want to be good at, right? I don't want to—I feel like Microsoft is at its best when—I always joke that the biggest hit consumer product Microsoft had was Excel. Think about it: Excel was the consumerization of a concept called spreadsheet, which empowered billions of people to be able to get number sense, right? I mean, that's what we are. And so, I want to be able to do that with Copilot, and that's our goal.

你知道，你在安全方面做得很好，让这些东西使用起来相当安全。但我确实想知道，这是一个快速发展的领域。这会让你慢下来吗？这会阻止你在前沿产品上做出大胆的赌注吗？

You know, and you've done a great job on the security side of making these things pretty safe to use. But I do wonder if, you know, this is a fast-moving space. Does that slow you down at all? Does it keep you from making bold bets on cutting edge products?

我认为，在这一点上，所有这些事情都是相辅相成的，对吧？当你说前沿产品会打破一些东西时，坦率地说，在现阶段，这在社会上不可接受，在经济上也不可行，对吧？因此，我认为你必须同时真正考虑甚至新创新的意外后果，因为风险很高，对吧？这些不是玩具，这些不是范围有限的东西。所以当你部署时，想想看，对吧？如果你看看 Hugging Face 事件，可以说“谢天谢地它发生在测试阶段，损害有限”，这是一回事。但如果这发生在——假设我有一个内部部署的智能体，而且这些智能体现在有欺骗行为，如果它去更改了我的 ERP 系统和记录，而我没有检测到，并报告了一些财务数据——我的意思是，那是一个巨大的问题。因此，你不能说——这就是为什么我认为大多数企业会醒悟过来说，除非我有一个独立于模型的控制系统，能给我保证我可以部署这些模型，否则我将很难部署这些模型。所以，我认为我们作为一个行业必须意识到这一点。客户肯定必须意识到这一点。这不是关于闪亮物体和错失恐惧症（FOMO），我认为早期 AI 可能有点那样。我认为现在是时候让每个人都说，嘿，看看，每个人都在关注 ROI。那很好，但这是关于 ROI 加上安全和保证。这是所有那些要素。

I think that look, at this point all of these things sort of go together, right? Which is when you say cutting edge products that sort of break things, it's just not going to be socially acceptable or economically viable, quite frankly, at this point, right? So therefore, I think that you have to truly think about the unintended consequences of even new innovation simultaneously given the high stakes, right? These are not toys, these are not things that sort of have limited scope. So when you're deploying, I mean think about it, right? You know, if you look at the Hugging Face incident, it's one thing to say well, it was thank God it happened during some test phase and the damage was limited. But if this had happened where let's say I have an agent that is deployed internally and given these agents now have deceptive behaviors, what if it went and changed my ERP system and the records and I didn't detect it and I reported some financial—I mean, that's a massive issue. And so therefore, you can't say—and that's why I think that most enterprises are going to wake up and say, look, unless and until I have a control system that is independent of the model that gives me assurance that I can deploy these models, I'm going to have a tough time deploying these models. And so, I think we as an industry have to wake up to this. Customers definitely have to wake up to this. It's not about the shiny object and FOMO, which I think a little bit of AI may have been in the early days. I think time has come for everyone to say, hey look, what is—everyone's been focused on ROI. Now that's great, but this is about ROI plus security and assurance. It's all of those elements.

我的意思是，你谈到过这如何从关于每美元每 token 成本的对话，转变为现在关于每——我不知道你现在怎么称呼它——但更像是每任务或每行动的成本，对吧？你认为最终这些模型只是商品化了吗？

And I mean, you've talked about how this has gone from a conversation about, for instance, like a cost per token per dollar to now sort of a cost per—I don't know what you're calling it these days—but sort of like a task or an action, right? Do you think that ultimately these models are just commoditized?

我认为有几种不同的看待方式，对吧？归根结底，一切都是商品，但没有规模的东西就不是商品，对吧？我的意思是，云是什么？云就是算力，对吧？它是一种商品，但然后你必须把这些东西整合起来，提供 SLA 等等。所以我认为，如果你说的是会不会有一个模型统治一切？我认为现在很清楚，情况不会是这样，对吧？我的意思是，如果你想想，这很有趣——GPT-4 可能是最后一次，在任何人接近之前有将近一年的时间。但现在，在闭源模型之间，在开放权重模型之间，有更多的——我称之为模型。我觉得有趣的是，事情正在向所有其他层面转移，对吧？也就是嘿，这个产品能做什么任务？所以如果产品确实带有模型支架，比如在 Copilot 中——首先，我上次在 Copilot 中选择模型是从来没有过，对吧？因为现在是自动模式，它会选择模型。我在不同形态之间移动。我有时在聊天中，有时在代码工作中，我在自动驾驶中，我在代码中。它有点像是在正确的形态中为正确的任务选择正确的模型并完成它，对吧？任务或工作，那种——是的，在那种背景下，模型是商品。它使用了多个模型，但模型也是关键成分。产品形态很重要。分销很重要。定价很重要。品牌很重要。所以这就是我认为我们都会——这不再是——我认为痴迷于一件事，我不认为那永远不会是，过去也从来不是，将来也不会是。

I think there are a couple of different ways to look at it, right? At the end of the day everything is a commodity and nothing at scale is a commodity, right? I mean what is the cloud? The cloud is just compute, right? It's a commodity but then you have to pull these things together, deliver the SLA and what have you. So I think if you're making the point of is there going to be one model to rule them all? I think it's pretty clear that that's not going to be the case now, right? I mean if you think about it, it's fascinating—GPT-4 probably was the last time where there was close to a year before anyone even came close. But now there is between closed source models, between open weight models, there's a lot more—I'll call it models. I think fascinatingly enough things are moving to all the other layers, right? Which is hey what task can this product do? So if products do come with models scaffolded like in Copilot—first of all, the last time I even selected a model in Copilot was never, right? I mean because now it's auto mode and it picks the models. I move between form factors. I'm in chat sometimes, I'm in code work sometimes, I'm in autopilot, I'm in code. It's sort of picking the right model for the right task in the right form factor and getting it done, right? The task or the job and that sort of—yeah models are commodities in that context. There are multiple of them it's using but models are also a key ingredient. The product form factor matters. Distribution matters. Pricing matters. Brand matters. So that's kind of how I think we'll all—it's no longer—I think the obsession was about one thing, and I don't think that's never going to be and never was the case in the past and never will be.

那这对这些前沿实验室的业务意味着什么呢？

Does that mean though for the business of these frontier labs?

我的意思是，看，这些是神奇的业务，从来没有——如果我回顾过去，对吧？我以前总是说：“哇，Windows 是一个难以置信的业务。Office 是一个难以置信的业务。”然后我想：“天哪，看看谷歌搜索，那是最难以置信的业务。”或者 Facebook 信息流之类的，对吧？这些都是发生的魔法般的事情。它们是其中之一，对吧？你可以列出像这样的最大业务。在那种背景下，这些人——无论你知道，我们对 OpenAI 的成功感到兴奋，显然 Anthropic 也取得了巨大的成功。毫无疑问，他们在许多方面都有逃逸速度。未来会怎样，对吧？我看着他们说：“嘿，这不是一个你最初的成就有任何意义的行业。”这取决于你是否能继续坚持下去，对吧？当我看着我们 51 年，记得我加入微软时，我们的生存问题是否是 Novell 会杀死我们。对吧？我的意思是，所以还有很多比赛录像。对他们来说还早，但成功是我从未见过的。

I mean, look, these are magical businesses that never have—if I look back, right? I used to always say, "Wow, Windows was an unbelievable business. Office was an unbelievable business." Then I thought, "My god, look at Google Search, which is the most unbelievable business." Or the Facebook feed or what have you, right? These are all magical things that happen. They're one of them, right? You can list the largest businesses like that. And in that context, these folks are—whether you know, we're thrilled about the success OpenAI has had and obviously Anthropic has had tremendous success. And there's no question that they have escape velocity in many fronts. What the future holds, right? Like I look at them and say, "Hey, this is not an industry where your initial success means anything." It depends on whether you can continue to keep on being at it, right? When I look at our 51 years, remember I joined Microsoft when our existential thing was whether Novell was going to kill us. Right? I mean, and so lots of game tape still. Early days for them, but success unlike anything I've seen.

是的，人们说你们的微软 AI 模型大约落后前沿 6 个月。显然，这不会给你炫耀的资本，但考虑到你刚才解释的所有事情，这是一个明智的——这是一个明智的位置吗？

Yeah, your Microsoft AI models are about 6 months, people say, behind the frontier. Obviously, that doesn't give you bragging rights, but is that a smart—is that the smart place to be given all the stuff you just explained?

对我们来说。

For us.

因为从某种意义上说，我们采取了……显然，我们对与 OpenAI 的合作感到非常兴奋，对吧？我的意思是，如果没有双方的合作，我们就不会走到今天，OpenAI 也不会走到今天，而且我们能够使用他们的模型和知识产权一直到 2032 年，这让我们在构建像 Copilot 这样的产品时能够最大限度地利用这一点。

Because in some sense, we've taken the, you know, obviously, we are thrilled with the partnership with OpenAI, right? I mean, we wouldn't be here and OpenAI wouldn't be here but for both sides partnering and the fact that we have access to their models and their IP all the way till '32 gives us, I would say, one maximally use that as we build products like Copilot.

同时，我们做出的一个决定是自下而上地构建我们自己的超级智能能力。这就是团队一直在做的事情，我们一直在取得进展。

And at the same time, one of the decisions we made was to build our own superintelligence capability from the bottom up. That's what the team has been doing and we've been making progress.

哦，这里面有好处，我认为，从战略上讲，对吧？炫耀的资本在某种程度上是一种激励，但对我来说，最终的炫耀资本是，我认为我们已经到了一个点，即 AI 为谁服务、为什么服务，对吧？

Oh, and in there there's benefits, I think, strategically, right? The bragging rights is sort of one thing that is motivating in many ways, but at the same time the ultimate bragging rights to me is I think we've even reached a point where AI for whom and for what, right?

所以，我认为这就是为什么要有一种意识，嘿，我们正在以那种人文主义的方法构建 AI。关键在于真正专注于赋能，对吧？我觉得即使在今天，至少在西方的美国，当人们想到 AI 时，让我们面对现实吧，它并不是最受欢迎的东西。为什么？我认为是因为我们有点迷失了方向，没有人把 PC 看作除了赋能工具之外的任何东西。我希望我们回到这一点，对吧？你必须回到构建与人类和社会相关的、站在历史正确一边的东西。

So, I think that's why having a little bit of a sense of, hey, we're building AI with that humanist approach. It's about really being focused on empowerment, right? I feel like even today where if I look at even at least in the West and in the United States, when people think about AI, let's face it, it's not the most popular thing. And why is it? And I think it's because we've lost a little bit of the plot on nobody thought of the PC as anything but an empowering tool. I want us to get back to it, right? You have to get back to building things that are on the right side of history as it relates to humans and society.

而这更多的是炫耀的资本。我认为研究人员需要，我们也需要，所以这不是说，是的，这里面有战略，实际上对我们来说，少花钱是合理的，坦率地说，我看的一件事是，我们有成本去做一些事情，比如说落后 6 个月或 3 个月，甚至只花 20% 的成本。这对我们的股东来说太棒了，但我这样做只是因为这个吗？我也想做一些对微软来说有意义的事情，让这更像一个软件业务，更像一个工具业务。我有更大的目标。

And that is more the bragging rights. I think that researchers need and we need and so it's not about like, yeah, there's strategy in there which makes sense for us to in fact under spend and quite frankly, I look at one of the things is we have our costs to do something that's off, let's say 6 months off or 3 months off even at, you know, 20% of the cost. That's fantastic for our shareholders, but am I doing it just because of that? I also want to do something which makes sense for Microsoft to do, make this more like a software business, make this more like a tools business. I have bigger goals.

让我们，你知道，以这样一种方式粘贴这个，即从底层构建对齐。所以，我觉得这必须是我们的研发目标。我不想只是为了竞争而按照别人设定的条件竞争，这不符合审慎的商业战略。但我认为历史可能会证明，这也不是一个好策略。

Let's go, you know, paste this in a way that there is alignment built in from the ground up. So, I kind of feel like that's got to be our R&D objective. I don't want to just compete for the sake of competing on terms that someone else has set and it makes it doesn't want to make prudent business strategy. But, I also think history may prove that that's not a good strategy to take anyway.

营销这项技术的一大问题是你怎么称呼它。我看到你，你知道，在这篇博客文章中你称之为超级智能，而不是 AI。嗯，这已经成为一个大讨论话题。我想说 AI 可能不是一个很好的术语。我同意这一点。我不确定超级智能真的有什么用，因为，你知道，就像你说的，人们并不真的想要超级智能。他们只想要可靠的普通智能，对吧？尤其是你的客户。

One of the big things in marketing this technology is what you call it. And I saw you, you know, in this blog post you refer to it as superintelligence, not AI. Um this has sort of been a big topic of discussion. I would say AI is probably not a very good term. Like I'd agree with that. I'm not sure superintelligence really does much either because, you know, like you're saying like people don't really want superintelligence. They want just regular intelligence that's dependable, right? Especially your customers.

是的，我认为是的，我的意思是，Dark，首先，我并不是对这些东西的命名特别教条。但话虽如此，我想说 AI 尤其是 AGI 所发生的事情，我认为，对吧？有一些贬低人的东西，你知道，我妻子告诉我，她听到一位数学家，我想，在 NPR 上，非常，你知道，发自内心地讲述发生在他身上的事情，比如他作为一个伟大数学家的整个身份刚刚被打破了等等。

Yeah, I think Yeah, I mean, Dark, first of all, it's not like I'm super dogmatic about naming of these things. But, that said, I'd say what has happened with AI and especially AGI, I think, right? It's There's something demeaning about, you know, I was my wife was telling me she heard a mathematician, I guess, on NPR, who was very, you know, you know, a heartfelt you know, about what happened like entire identity of his around being a great mathematician has just been broken and what have you.

因为所有的千禧年数学奖。

Because of all the Millennium math prize.

没错。所以，我认为 AGI 这个词现在让每个人都不舒服。你知道，它剥夺了人的能动性、人的尊严。事实上，人的身份来自于人的雄心。所以，我希望超级智能作为一个术语在这方面更好，对吧？这有点像我们并不是说我们试图构建一种比人类更好的智能。让我们提醒每个人他们有多差劲。那不是目标。那不是行业。坦率地说，前沿实验室的目标也不是那个，对吧？我真诚地相信他们正在努力构建伟大的技术，从而赋能人们。

That's right. And so, I There's something about this AGI term that is now rubbing, I think, everyone. You know, it's taking away human agency, human dignity. In fact, human identity that came from human ambition. And so, I I hope that superintelligence as a term is better that way, right? It's sort of like we're not trying to say we're trying to build an intelligence that's better than human. Let's remind every human how they suck. That is not the goal. That is not the industry. And that's not the goal, quite frankly, with the Frontier Labs, too, right? I genuinely believe that they're trying to build great technology such that that empowers people.

所以，我认为我们必须达到一个能够……我不知道仅仅重新命名这个术语是否能做到，但肯定是 AGI，因为那是另一个艺术术语，对吧？它被抛来抛去，好像它将是伟大的成就。而且

And so, I think we've got to get to a place where we are able to I don't know whether just simple rebranding of the term does it, but definitely AGI because that's the other term of art, right? Which was thrown around as if it was going to be the great accomplishment. And

只是自动化？这就是我想知道的。

just automation? That's what I want to know.

我的意思是，你看，我，你知道

I mean, I look, I'm you know

不那么性感。

Not not as sexy.

是的，是的。说得好。是的，欢迎你和美国总统谈谈。他命名的，而且，我的意思是，就我而言，它和其他任何名字一样好。

Yeah, yeah. Good talk. Yeah, you're welcome to talk to the president of the United States. He named it and and I mean, it's as good a name as anything as far as I'm concerned.

是的。你的战略最近确实得到了华尔街的回报。我认为他们欣赏你的财政纪律，但安全方面，同时我看，你知道，英伟达现在我认为每季度赚的收入比你还多。你有前沿实验室，你知道，正如你所说，正处于增长狂潮。你有 XAI 加入竞争。我想知道，是否存在一种情况，微软继续是一家出色的企业，但也许几年后，在未来某个时候，甚至不在七巨头之列什么的。那会重要吗？

Yeah. You your strategy has has really been rewarded by Wall Street lately. I think they appreciate your sort of fiscal discipline, but the the security stuff at the same time I look at, you know, Nvidia is now I think earning more revenue per quarter than you. You've got the Frontier Labs, you know, as you said just on a growth tear. You've got XAI coming into the fray. I wonder if there's a place where Microsoft continues to be a fantastic business, but maybe a few years from now whenever in the future is not even in the mag seven or something. Would that matter?

是的。

Yeah.

看，我的意思是，归根结底，我的工作是确保我们能够取得进展，不辜负我们的使命、我们的品牌、我们的客户对我们的期望。然后其余的是，你知道，市场是聪明的。他们会决定什么值得估值，什么不值得，让这些放在一边，因为我不像坐在那里说，你知道，这就像任何俱乐部，一个人属于它是目标本身。

Look, I mean, at the end of the day, my job is to make sure we are able to make progress and live up to what is our mission, our brand, our what our customers expect of us. And then the rest is, you know, markets are smart. They will decide what is worth valuing and not worth valuing and let sort of let that sit aside because it's not like I sit around and say, you know, it's like any club that one belongs to is the thing that is the goal itself.

我的意思是，看，我我回顾。我说我是公司和公司历史的忠实学生。我的意思是，想想看，对吧？没有天赐的权利。事实上，大多数公司的寿命比人类的寿命短得多。某种东西能持续很久是罕见的。而我要说，让我们走到 51 年的是不断更新和保持相关性的过程。

I mean, look, I mean, I I go back. I say I'm a big student of company and company histories. I mean, think about it, right? There is no God-given right. In fact, most companies have longevity that's much smaller than human longevity. It's rarity something sort of lasts long. And the fact that I would say what brought us to 51 years is that constant process of renewal and relevance.

嗯，这不是庆祝过去的东西。所以，我不把任何事情视为理所当然，无论是今天的成功还是过去的成功。嗯，只要我们在为客户做相关的事情，我们就会在 whatever 市值/俱乐部，那是当时的时尚。但那不是我所关心的，我关心的是我们拥有那个核心身份。就像 1975 年的微软到 2026 年的微软确实拥有，正如我所说，那个核心属性，我们专注于赋能人和组织。嗯，自 Altair 以来发生了很多事情。哈哈。但话虽如此，我不认为我们的身份发生了那么大的变化。

Um it's not about celebrating what came in the past. And so, I don't take anything for granted, whether it's today's success or past success. Um and if as long as we're doing relevant things for customers, we'll be in whatever market cap slash club that is the the the flavor of the day. But that's not uh what I do care though is that we have that core identity. Like the Microsoft of 1975 to the Microsoft of 2026 does have, as I said, that core attribute that we're focused on empowering people and organizations. Uh technologies have been a lot has happened uh since the Altair. Haha. But that said, I don't think our identity has shifted as much.

是的，这很有意思地引出了我另一个问题：你一直喜欢引入外部人才。过去几年似乎有大量外部领导者加入，比如 Mustafa、Carolina。我有一份名单：Charlie Bell、Asha、Jason Taylor，过去几年都有。这是否是因为你把当前这个 AI 时代视为某种战时状态？所以必须引入合适的领导者？

Yeah, which is an interesting segue into another question I have, which is you've always liked to bring in outside talent. It seems like in the last couple years, there's been a big influx of outside leaders, you know, Mustafa, Carolina. I mean, I've got a list. Charlie Bell, Asha, Jason Taylor, all these people in the last couple years. Is that a function of you looking at this time, this AI time as sort of wartime? Like you got to bring in the right leaders for that?

嗯，首先，我对那些像我一样在微软成长起来的人感到非常高兴，我们有很多这样的人。也有从外部加入并在微软高级职位上表现出色的人。我认为这是正确的平衡。我一直觉得，这不仅仅是内部人的事，也不仅仅是外部人的事，还关乎内部人懂得如何像外部人一样思考，以及外部人加入后懂得如何像内部人一样思考，否则就行不通，对吧？我总是对从外部来的人说，你必须与你现有的人见面并领导他们，而不是认为我可以拥有一支不同的军队，对吧？我的意思是，军队在这里，不在别处。这对从外部来的人是一个巨大的挑战，而对内部人来说，最大的挑战始终是：他们是否真的具备那种理智的诚实，来评估你真正所处的位置，以及具备竞争力意味着什么，标准已经变了等等。所以这就是为什么我认为需要混合人才，最终不在于你是内部人还是外部人，而在于能够成功。我对任何人的衡量标准都是：嘿，你最终必须创造能量，你必须创造清晰度并取得成功。

Well, I mean, first of all, I'm thrilled about people who like me have grown up at Microsoft and we have a lot of them. And there are people who have come from the outside and are doing fantastic at senior levels at Microsoft. I think that's the right balance. I've always felt that it's not about just the insiders, it's not just the outsiders, it's also about insiders knowing how to think like an outsider and outsiders when they come from the outside knowing how to think like insiders because otherwise it won't work, right? I always say to people who come from the outside that you have to meet the people that you have now and lead them versus thinking that I can have a different army, right? I mean, you have the army here, not somewhere else. And that's a big challenge for people who come from the outside and for the people inside the big challenge always is can they really have that intellectual honesty to assess where you really are and what it means to have the competitiveness and the bar has changed and what have you. And so that's why I think having a mix of folks and ultimately it's not about being an insider or an outsider, it's about being able to succeed. And my measure of anyone is about like, hey, you know, you got to ultimately create energy, you have to create the clarity and achieve success.

你心目中有没有认为可能成为微软下一任 CEO 的人选？

Do you have any in mind who you think might be a next CEO of Microsoft?

所有人？

All of them?

所有人。我的意思是，我想要每个人，但你看，如果我们有一群有雄心壮志、团队合作良好的人，我很高兴拥有这个团队。

All of them. I mean, I want everyone but every Look, if we have a set of folks who are a set of ambitious folks who work well as a team and I'm happy to have this team.

然后我想正如我所说，这不是说我们试图运行某种流程来挑选下一任 CEO，但包括我在内的每一位领导者的工作都是确保有继任计划。如何做好继任计划？通过确保人们有挑战可以应对，犯一些错误，再尝试，学习并变得更好。所以我的观点是，嘿，我们有一支很棒的团队，一些在这里很久的人也已经宣布，你知道，是时候进入下一个阶段了，我非常感激他们所做的工作。例如最近，我在微软一起成长了最长时间的 Rajesh Jha 决定从微软退休。他是让 Office 365 成为现实的人。所以能够庆祝那些在微软拥有充实职业生涯的人真是太棒了。我希望这就是我的目标。我的目标是，无论是在这里成长的人还是从外部来的人，都有机会拥有一个可以称之为家的平台，他们可以产生影响，然后当他们最终完成时，可以回顾并说他们取得了巨大的成就。

And then I think as I said, I think this is not like, you know, it's not like we're trying to run some process and this is about trying to pick the next CEO but the job of every leader, including me, is to make sure that there is succession. And how does one work on succession? One works on succession by making sure people have challenges that they are able to meet, make some mistakes, go at it again, learn, and get better. And so my thing is that hey, we have a fantastic team and some people who have been here for a long time have also called it and said, you know, it's time to go to their next phase and I'm deeply appreciative of the work they've done. Lately for example, Rajesh Jha whom I've grown up with at Microsoft for the longest time, has decided to retire from Microsoft. And he's the guy who made Office 365 happen. And so it's fantastic to even celebrate folks who had a fulfilling career at Microsoft. And I hope that's kind of my goal. My goal is that there's an opportunity for people who have grown up here or people who even come from outside to have a platform where they can call home, they can have impact, and then when they're finally done, they can look back and say that they have achieved something big.

你对自己会留任多久有目标日期吗？

Do you have a target date for how long you're going to stay on?

我享受我所做的每一天，我没有目标。

I am enjoying every day of what I'm doing, and I have no targets.

很高兴听到这个。看着很有趣。我想时间到了，但一如既往，我学到了很多，感谢你参加播客。

It's great to hear. It's been fun to watch. I think that's all the time we have, but I've learned a lot, as always, and appreciate you joining the podcast.

嗯，非常感谢你抽出时间。

Well, thank you so much for taking the time.

Future Proof 由 Google Cloud 呈现。智能体不是应该让你安心吗？为了安全运行 AI 智能体而不断在遗留基础设施上打补丁有点过分。如果你的企业安全策略依赖于每月数千个补丁，你最大的威胁不是未来，而是补丁星期二。Gemini 运行在最安全的云上，具有高级威胁检测功能，可以在攻击变成入侵之前将其关闭。不，真的。用 Google Cloud 上的 Gemini 让工作更轻松。检查响应。需要订阅。

Future Proof is presented by Google Cloud. Weren't agents supposed to put your mind at ease? Constantly slapping duct tape on legacy infrastructure just to run your AI agents safely is a bit much. If your business's security strategy relies on thousands of patches every month, your biggest threat isn't the future. It's Patch Tuesday. Gemini runs on the most secure cloud with advanced threat detection to shut down attacks before they become breaches. No, really. Make work less work with Gemini on Google Cloud. Check response. Subscription required.

嗯，和 Satya Nadella 交谈总是非常棒。对话总是太短。他是我交谈时总能学到东西的人。他也非常擅长回答那些我认为可能会让一些 CEO 感到不舒服的问题。你知道，我问他一些关于微软未来的尖锐问题。而他那种冷静、镇定、专注的方式，你知道，很明显，这位 CEO 不只是朝着他看到的下一个闪亮事物跳跃，而是愿意做出非常大胆的赌注。我的收获是，目前微软实际上并没有在推动极限，对吧？他们很早就投资了 OpenAI，在大规模算力集群上下了巨大赌注，真正促成了 ChatGPT 的诞生，但现在我认为他们正在观望和等待。看看他们的 AI 模型，我们谈到了微软的 AI 模型，大约落后前沿 6 个月，他们正在构建非常有用的模型，但并没有在训练这些模型上花费超出必要的钱，对吧？他们正在建设数据中心和足够的算力资源来满足他们目前看到的需求，但他们并没有押注未来最重要的是谁拥有最多算力，对吧？他们在等待，认为有更多时间观望，我认为在应用层也是如此。他们没有推出像 Muse 那样炫目的消费类产品，也没有冒险尝试那些可能出错、可能失控的事情，你知道，这些技术可能会发生这种情况，他们说那不是我们的身份。

Well, it's always really great to talk to Satya Nadella. The conversation is always too short. He's someone I always learn from when I talk to. He's also so good at answering questions that, you know, I think could get him, you know, some CEOs a little out of their comfort zone. You know, I'm asking him about, you know, some tough questions, right, about the future of Microsoft. And I mean, the way he's just calm and collected and centered doesn't you know it's just clear that this is a CEO who is not just like jumping in every direction at the next shiny thing that he sees but he's willing to make these very bold bets. My takeaway from this is that at this time right now Microsoft is not actually pushing the limits right? They were now you know they invested in open AI very early on they made a huge bet on you know massive a massive compute cluster that really enabled chat GPT to come into existence but right now I think they're sitting back and waiting a little bit. You look at their AI models we got into the you know to the Microsoft AI models which are about 6 months behind the frontier they're building these models that are very useful but they're not spending more money than they have to on training those models right? They're building data centers and enough compute resources to satisfy where they see the demand right now but they're not making a bet that you know all that matters in the future is who has the most compute right? They're waiting to kind they think they have a little bit more time to wait and see and I think that's true on the application layer as well. They're not coming out with you know flashy consumer type products like like Muse and taking chances that maybe some of those things go wrong that maybe they go off the rails that you know this can happen with with these types of technologies they're saying that's not our identity.

未来这件事能否得到回报，取决于我认为这个领域里其他这些公司有多成功。比如，这些前沿模型实验室都能成功吗？它们都是全垒打吗？有些会失败吗？有些会被收购吗？我觉得这些会决定结果，因为这真的是一场相对的竞赛，对吧？有可能——真的有可能——几年之后，微软不再是最大的公司、最热门的公司。但它会继续存在。

Whether this pays off or not in the future is just going to depend on how successful I think all these other companies are in the space. Like, do all of these frontier labs make it? Are they all home runs? Do some of them fail? Are some of them acquired? I think those things will sort of determine it, because it really is, I think, sort of this race really is relative, right? It's possible — it really is possible — that Microsoft may not be, years from now, the biggest company, the hottest company. It will be around.

那家公司的 CEO 会是谁？我问萨提亚继任者是谁时，他没有给出任何暗示。他说所有人——意思是任何一个——公司里有很多领导者，他认为他们都有能力接掌大权。萨提亚在微软的任期非常成功。你知道，他 59 岁了。所以，这对他来说可能不会是又一个十年，但可能是又一个五年，而这五年对这家公司绝对至关重要。

Who will be CEO of that company? Satya didn't give any hints when I asked him who the successor was. He said all of them — meaning any — there's many leaders at the company who he thinks are capable of taking the reins. Satya has had a great run at Microsoft. He's, you know, 59 years old. So, you know, this may not be another decade for him, but you know, it could be another five years, and those are absolutely critical five years for this company.

感谢收听 Semafor Technology 的《Futureproof》。本节目由 Josh Billinson 制作，特别感谢 Adam Baneki、Ben Smith、Claire Einstein、Tory Cour、Amber Ali、Stephanie Chang 和 Daniel Hecht。如果你喜欢《Futureproof》，请在任意播客平台关注我们，也欢迎留下评价。如果想了解更多，可以随时订阅，让 Semafor Technology 直达你的收件箱。

Thanks for listening to Futureproof from Semafor Technology. Our show is produced by Josh Billinson, with special thanks to Adam Baneki, Ben Smith, Claire Einstein, Tory Cour, Amber Ali, Stephanie Chang, and Daniel Hecht. If you like Futureproof, please follow us wherever you get your podcast and feel free to leave a review. And if you want more, you can always sign up to get Semafor Technology in your inbox.