纳德拉在Build：每家公司都该有自己的前沿智能

Sarah Guo

欢迎来到 No Priors 和 Latent Space 的交叉特辑，嘉宾是萨蒂亚·纳德拉。恭喜你们办了一场很棒的 Build。

Welcome to a crossover episode of No Priors and Latent Space with Satya Nadella. Congratulations on an amazing Build.

Satya Nadella

非常感谢。很高兴和你们在一起。我一直在听你们两边的播客，能上节目太好了。

Thank you so much, and it’s great to be with both of you. I listen to both of the podcasts all the time. It’s great to be on it.

Sarah Guo

你整个上午都在讲微软各条线的发布，大概讲了三个小时。对你来说，最重要的反思或带走的一点是什么？

So you’re just talking about these amazing announcements from across the Microsoft estate all morning for, I think, three hours. What is the most important reflection or takeaway you have?

Satya Nadella

也许最大的一点是：把它想成一场生态，而不是单一模型，甚至不是单一平台。至少对我来说，在微软长大、经历过大概四次重大平台迁移，我属于这一派：平台的定义，根本上是它在平台之外创造的价值，要多于平台自己捕获的价值。所以如果你看现在发生的事，今天上午的主题演讲就是：任何公司，无论是 AI 原生公司还是传统企业，怎样作为一等参与者进来，并能指着「这是他们自己创造的 AI」。他们当然也会用别人的 AI。但对我来说，路径是什么？配方是什么？栈长什么样？工具长什么样？什么东西有价值？怎么做？那就是我们的工作。

I’d say there are perhaps the biggest one for me is let’s sort of conceptualize this more as an ecosystem play as opposed to a single model or even a single platform. At least for me, having grown up at Microsoft, having seen four major platform shifts, I sort of fall into that camp where a platform is defined by fundamentally its ability to create more value outside the platform versus what’s captured in the platform. And so if you view what’s happening right now, I think this morning’s keynote was how can any company, whether it’s an AI native company or a traditional enterprise company, participate as a first-class participant where they can point to AI they created. It’s not that they don’t use other people’s AI. Of course they will. But to me, what’s the path? What’s the recipe? How do I do it? What does a stack look like? What does the tooling look like? What is valuable? How do you do that? That’s it. That’s sort of our job to do.

Sarah Guo

生态策略非常复杂，因为你最终会自建某些组件、为某些组件找伙伴、再支撑它们。你们刚发布了这一大套模型。能不能讲讲微软现在的训练策略？

Ecosystem strategy is very complicated, right? Because you end up building certain components, partnering for certain components, supporting them. You just announced this big suite of models. Like, tell us a little bit about the training strategy for Microsoft now.

Satya Nadella

我们想用 MAI 模型去做的，正如 Mustafa 讲的，首先是一条很好的血统：从预训练开始，数据质量要非常好，把消融都做完。某种意义上，做出干净血统的模型正在变得更难，因为外面的东西太多了，你必须真正消融掉，才能有一个出色的预训练模型。事实上，很多开源权重模型的挑战之一就是：它们在一两个基准上看起来很好，但在实践里不行。所以我们对这些 MAI 模型确实很兴奋——一个小小的 5B 模型到底怎么 hill climb？这又回到我认为最终的关键：去追那个认知核心。所以从干净血统开始，然后让公司不只把它当通才用，而是围绕它搭 hill-climbing 脚手架，做成自己的专才。不只是模型，而是周围有脚手架，然后你会开始建 RLE，开始收集 traces。最重要的是你会有私有评测，因为我们知道外面那些评测有意思，但已经不是真正关键——它们都能被刷满。重点是每家公司会有自己的私有评测。所以围绕我们模型的端到端平台故事，我觉得才有意思。

So the thing that we wanted to do with the MAI models was to build, and as Mustafa talked about, first of all, a great lineage, starting with pre-training, with very good data quality, doing all the ablations, making sure — because in some sense it’s becoming even harder to build a clean lineage model just because there’s so much stuff out there that you truly need to ablate out to be able to have a fantastic pre-trained model. In fact, that’s one of the challenges of a lot of the open weight models is they look great on one benchmark or two, but they’re not great in practice. So that’s why, in fact, we are pretty excited about these MAI models because how the heck can a small 5B model hill climb? And it goes back a little bit to what I think is ultimately the key thing to do, which is try to pursue finding that cognitive core. So to me, starting with a clean lineage, then creating that ability for companies to be able to use this not just as a generalist, but to create their own specialist by building this hill climbing scaffold around it. So it’s not just the model, but you have a hill climb scaffold around it, then you will start building your RLE. You will start collecting the traces. Most importantly, you’ll have private evals because we know all the evals out there are good, interesting, but they’re not really that critical at this point because they all can be maxed. And so the point is each company will have its own private eval. And so that end-to-end platform story around our models is sort of what I think is interesting.

另外一点，Sarah，既然你提到了：我觉得有一个新的前沿。人们谈前沿、你是否在前沿上运转。有意思的是，如果你加一点点时间性，你可以用——我们展示的那个演示就很酷。我们用了 GPT-55，收集了一堆 traces，然后拿一个 5B 推理模型，达到了更高。那是「在前沿运转」的另一层含义。

And then the one other thing, Sarah, since you brought that up, is I do feel there’s a new frontier. People talk about the frontier and are you operating at the frontier. Interestingly enough, if you add a little temporality to it, you can use — in fact, the demo we showed was pretty cool. We used GPT-55, then you collected a bunch of traces, and then you took a 5B reasoning model and achieved higher. So that is another aspect of what it means to operate at the frontier.

swyx

首先得祝贺你们，基本上在两年里于微软内部建起了一个前沿的 neo lab。你现在知道、而希望两三年前就能告诉自己的是什么？跟 Jensen 的合作大概三年，MAI 大概两年。

I first of all have to congratulate you on basically building a frontier neo lab inside of Microsoft in two years. I’m wondering, what do you know now that you wish you would tell yourself two years ago — or two or three years ago? Three years for the Jensen partnership, two years for MAI.

Satya Nadella

我经常回头想的是：我真正进入这些，是因为对 scaling laws 那篇论文感到兴奋；OpenAI 的合作也是因为那些人说，我们真的会把大量算力砸进 transformer。它们确实在爬升。一种很粗的说法是：智能是算力的对数，某种程度上成立。现在我觉得我们也许低估的，是把这些部署出去、让它们在真实世界真正交付价值的复杂度。任何基准测到的结果都有意思、也重要，但真正的评测是：外面的人能做成只有他们才珍视的独特事情，而且非常可衡量。我希望我们作为行业，更早把这一点放进意识里。因为现在人们说「我不想 token max」，其实是我们没把自己想成：我们是在用 token 一步步创造价值。我希望我们更早到那里，但很高兴我们现在到了。

I think the thing that I reflect quite a bit on is, I got into all this when I got excited by the scaling laws paper, and even the OpenAI partnership came about when those folks said, hey, we’re gonna really throw a lot of compute at transformers. The thing that I always look back and say is, wow, these things do have capability that they’re climbing up. This crude way of saying it is intelligence is log of compute kind of works. Now what I think we underestimated perhaps is the real-world complexity of deploying these so that they actually deliver the value in the real world. The outcomes as measured by any benchmark is interestingly important, but the true eval is when people out there are able to do unique things that they only can value, and it’s very measurable. That I wish we had even had more in our consciousness as an industry. Because right now I think when people say, wow, I don’t want to token max, it’s an artifact of us not having thought ourselves as an industry that we are using tokens to create value every step of the way. So I think that’s kind of what I wish we had gotten there, but I’m glad we are here.

Sarah Guo

你在客户那边看到、创造最多价值的用例有哪些？大家谈得很多的是代码，规模影响显然很大。还有没有你觉得客户普遍真正受益的其他领域？

What are some of the use cases that you’ve seen that have created the most value for your customers? Because I know that people talk a lot about code, and I think it’s pretty clear that that’s something that’s having very large scale impact. Are there other areas that you find in common that your customers are really benefiting from?

Satya Nadella

编码确实已经起来了。但有意思的是，即便谈编码：它已经好到我们现在必须重做 IDE。我们发布的东西简直疯了——我有这一百个智能体会话，它把认知负荷转回给我这个人，多到我需要一个新的 UI。另外，只把聊天当唯一工件也不行，所以我们需要画布。你以为软件或 UI 会在哪里被需要，即便在完全智能体化的写代码世界里，你还是需要它。

I think, yeah, to your point, obviously coding is now — but it’s interesting even to talk about the coding, which is coding has worked so well that we now have to rebuild the IDE. I mean, it’s kind of nuts to see what we launched is like, oh my God, I have these hundred agent sessions. The cognitive load it transfers back to me as a human is so excessive that now I need a new UI. Oh, by the way, chat as the only artifact was also impossible, so that’s why we need a canvas. So it’s kind of interesting for all the things about where is software needed or where is UI needed — you kind of need that even for code, in a fully agentic world.

不过我们开始看到的是，从 co-work 开始，再到我们展示的 autopilot，以及你在 claws 上看到的：大量人力资本在做胶水工作。如果你现在能用 token / 智能体去增强，而且它们是长程、可持久的，那么你放大判断和胶水工作的能力，就会像编码那样被放大。我很肯定，六个月后我们都会说：哇，整晚都有一堆 autopilot 在用我授权出去的权限替我干活。当然我需要新的界面来问：你做了什么？这工作是我做的吗？所以压缩工作流、完成任务，价值会大量在那里被创造出来。

But that said, one of the things that we are starting to see — we started seeing with co-work, but even some of the work we showed with autopilot, right, and what you see with claws is a good one — because if you think about a lot of human capital is doing the glue work. If you now can augment that with tokens/agents that are long-running, durable, then your ability to scale even what is still judgment and glue work gets amplified like coding does. I’m positive that six months from now we’ll all be saying, oh wow, all through the night there was a bunch of stuff that all these autopilots that I have working on my behalf with my delegated authority. Then of course I’ll need my new IDE to say, well, what did you do? Did I do this work? And so on. So I think that’s where compressing of workflows, completing of tasks, that’s where I think a lot of the value gets created.

Sarah Guo

你提了一个很有意思的点：一边是真正在写代码的智能体，一边是它周围的 harness——环境、上下文、开发者在编码智能体周围搭的一切。企业的 harness 是什么？更广的生产力工作有没有对等概念？

I think you raised a really interesting point, which is there’s the actual agent that’s doing the code, and then there’s a harness around it, and that’s the environment, that’s the context, that’s everything you’re setting up as a developer around actually a coding agent. What is the harness for the enterprise? Is there an equivalent concept for broader productivity work, or how do you think about that concept sort of generalized?

Satya Nadella

某种意义上，你希望 harness 定义模型、数据和工具，让这三者构成一个环。我们首先要确保自己建的每一款产品，无论是 GitHub Copilot、安全 Copilot、我们展示的 MDASH，还是科学发现，全都是多模型 harness，带工具访问，甚至可以逐步披露工具，好让它 token 高效。然后你要用非常丰富的上下文去喂它，因为过去两年另一个很难的教训是：天哪，你得做多少工作去准备上下文层，才能让计划以最高效的方式执行，魔法就在那里。我们有 GitHub harness，基本上用在所有产品里，在 Foundry 里也能用。我们是开放的：你可以用你的 Llama harness，或任何开源 harness、你自己的 harness，配上你的工具、多个模型和你的上下文。现在很多讨论是：如果我把 harness 加工具再加模型一起训，你就有评测。我们在证明的是——最好的例子是 MDASH：它上线时找到了 Mythos 没找到的 bug 或漏洞。所以我愿意说，有存在性证明：你可以有一个多模态 harness，在真实世界里实际上更强。

That’s right. So in some sense you kind of want the harness to define the models, the data, and the tools, so that you have a loop across those three. And so what we are trying to first of all make sure is each of our products that we build, whether it’s GitHub Copilot or the security Copilot, the stuff we showed with MDASH, or even the discovery for science, it doesn’t matter, all of them are multi-model harnesses, with tools access so that you can do this progressive disclosure of tools even so that they’re token efficient. And then you’re feeding it with very rich context because that’s sort of the other hard lesson we have learned in the last two years is, oh my God, the amount of work you need to do to prep the context layer, such that your plan can execute in the most efficient way, is where the magic is. So we have, in our case, the GitHub harness, which essentially we’re using across all our products. It’s available in Foundry, and we are open — you can use your Llama harness, whatever. Or you can use any open harness or any harness of yours and train with your tools and multiple models and your context. And so that’s the pitch. Because right now a lot of dialogue is, hey, if I train the harness plus tools and the model together, you get evals. And what we are proving out is — and the best example of that is what we did with MDASH, right? Because when it launched, it found bugs or vulnerabilities that were not found by Mythos. And so there is existence proof, I would claim, that you can have a multimodal harness that can in fact be more performant in the real world.

Sarah Guo

独立前沿实验室训练的前提，其实是：我们会有这些模型，会有 API 生意，会支持企业和创业公司。但第一方产品，无论是生产力、代码还是搜索，贡献了大部分收入。那是一种不同的价值方程，和你描述的微软生态不太一样。如果是这样，开发者在那个世界里，难的是什么、技能是什么、价值捕获在哪？

So a premise behind the training at the independent frontier labs is really, we’re gonna have these models, and we’ll have an API business, and we’ll support enterprises and startups. But a first-party product, be it productivity or code or search, drives the majority of revenue. That’s a different value equation than you’re describing, I think, with the Microsoft ecosystem. If that’s the case, tell me if it’s the case, because obviously you have first-party products and you have enablement products. What is the role of the developer? Like what is gonna be hard and the set of skills and the value capture the developer has in that world?

Satya Nadella

成功的平台构建者同时拥有第一方产品，这会一直存在。Windows 如此，SaaS 和云这边我们也是。但关键差别是：这不该成为别人取得同样成功的限制。这一次围绕智能的网络效应是这样的——它们从数据里学，而且还不是很多数据，往往是很少的样本，你就能看出某件事新在哪。所以游戏变成如何保护。我会说，每家公司拥有私有评测，可能是最大的 IP。想一想：那个你可以拿来让前沿模型去 hill climb、同时不把 traces 泄漏出去的私有评测，可能是 IP 最大的驱动之一。换句话说，一个试金石是：你有一个私有评测，你在用模型 A。你能不能换成模型 B，还能往上爬？如果能，你在掌控之中；如果不能，你就不在掌控之中。harness 的选择因此变得超级重要。所以有一个开放 harness，让所有模型进来，用你的评测、你的上下文、你的工具帮你 hill climb——这是 AI 原生创业公司、SaaS 公司、每家企业都需要的技能。

So I think that there’s always gonna be the case that someone who is super successful as a platform builder can also have first-party products. It was true with Windows. It is true with the SaaS side and the cloud side as well with us and others. But the thing is, it should not be a limiter to other people achieving that same success. That I think is the core difference, which is the network effects this time around around intelligence are such because they learn from data, and not really lots of data. It’s just a few samples that you have to see to understand what’s novel about something. So that’s why the game becomes how to protect. So that’s why I would say every company, having private evals may be the biggest IP. Think about it, like what’s that private eval that you can then use even a frontier model to hill climb on and not leak the traces may be one of the biggest drivers of IP. Like, so in other words, another acid test is you have an eval that’s private. You’re using a Model A. Can you switch it to Model B and climb up? If you can, then you’re in control. If you can’t, you’re not in control, and that’s where even the harness decision becomes super important. So therefore, having an open harness, letting all models come in, having your evals, your context, your tools help you hill climb, I think is the skills that an AI native startup needs, a SaaS company needs, or every enterprise needs.

swyx

某种很真实的意义上，微软历史上是操作系统公司，然后成为云公司。也许第三幕是你们是 harness 或评测公司。让每家公司拥有前沿智能，是使命，对吧？

Yeah, I think in a very real way you are — Microsoft historically is an operating systems company and then became a cloud company. Maybe like the third act is that you’re a harness or evals company. And I think enabling every company to have frontier intelligence is the mission, right?

Satya Nadella

就是这样。平台承诺就是：你跟我们一起建，你会为自己的数据得到你的智能。如果整个开发者大会只能有一句口号：每个人都能带着自己的前沿智能在前沿上运转。这太重要了，否则我不知道你怎么达到稳定均衡——我怎么能说，我的公司会有终局价值，因为我现在知道如何在一个会变得更好的平台上持续复利。Windows 出来时，Adobe 建了，Autodesk 建了；或者像 Jensen 说的，我们建了 DirectX，他在上面建了 CUDA。我总跟 Jensen 说，老天，那次我拿了较短的那一头，真希望我们当时认出来。但那个想法——你可以建一层平台，让别人延伸出去，在这个例子里建成他们自己的智能层——我认为就是一切。没有这个，为什么要开开发者大会？我大可以让你们来朝拜一个模型。但那就不是开发者大会。

That’s it. Like that is the platform promise, that you build with us, you will get your intelligence for your data. That’s it. To me, that is — if there was one tagline for this entire developer conference — can everybody operate at the frontier with their frontier intelligence? To me, that is so important because otherwise I don’t know how you achieve stable equilibrium, which is how do I then go and say, well, my company is gonna have a terminal value because I now know how to continuously compound on top of what’s a platform that gets better. When Windows obviously came out, Adobe built, Autodesk built, or even take what Jensen said. We built DX and he built CUDA on top of it. I always say to Jensen, God, I got the short end of that. I wish we had recognized it. But nevertheless, that idea that you can build a platform layer that someone else can then extend out and build their own intelligence layer in this case, I think is everything. Without it, why have a developer conference? I can just come and have you all sort of just worship at the altar of one model. But that’s not a developer conference.

swyx

后台我们讨论过什么是 IP、公司的价值在哪。以前是公司里人类经验的长度，现在是另一件事：评测，以及把智能体应用到公司上的经验。你能不能再展开一点，因为那很有洞察。

Backstage we had a discussion about what is IP or what is the value in a company. It used to be the length of human experience at a company, and now it’s this other thing which is the evals, the experience in sort of applying agents to the company. Can you flesh that out a bit more because it was very insightful.

Satya Nadella

这是很好的框法。说到底，每家公司都会同时拥有仍然极其宝贵的人力资本，因为人、以及他们找出始终存在的缺口的能力，会是我们创造价值的方式。我确实认为，即便 token 资本上升，这将是关于表达新形式的人类能动和野心。所以假设任何一家公司有大量 token 和大量人力资本，问题是你如何让两者复利。如果你在 Teams 里有一堆智能体在干活、一堆人在干活，那些 traces 之间的东西，才是这家企业如何创造价值的重要上下文。然后那会回到不是去训一个通才模型，而是训公司元老智能体。那再次非常有价值。一家公司如果说「这实际上应该上资产负债表」，我就是这么想的。人力资本从来没办法上资产负债表，因为你不知道如何捕获隐性知识。而现在我觉得可以，用那些随时间、通过所有 traces 学到东西的智能体。所以至少我们认为会发生这件事。

It’s a great way to frame it, right? Because at the end of the day, every company is gonna have both the human capital that is still gonna be super valuable, because humans and their ability to find the gaps that exist at all times is going to be the way we all will create value. I mean, so I’m definitely in the camp that this is going to be about expressing new forms of human agency and ambition even as token capital goes up. So let’s say any corporation has lots of tokens and lot of human capital. The question is how do you compound the two. So if you take in Teams I have a bunch of agents doing work and a bunch of humans doing work, and the traces between those, that is really important context of how that enterprise is creating value. Then that goes back to train not a generalist model, but to train the company veteran agent. That is super valuable again, which is when a company goes says, it should in fact go onto the balance sheet, is how I think about it. In fact, human capital was never possible to go put on a balance sheet, because you didn’t know how to capture the tacit knowledge. Whereas now I think you can with the agents that have learned through time, through all the traces. So that’s what at least we think will happen.

swyx

我觉得 SEC 将不得不为 token 专长制定会计准则。

I think the SEC is gonna have to have accounting standards for token expertise.

Sarah Guo

你在谈均衡、稳定均衡，公司能有这种复利、能看到自己的终局价值。对公认均衡的另一个挑战是：有些应用和工作流对某个垂直或水平是共通的，那是一代 SaaS 公司，微软自己也有很多 SaaS；另一些则对每家企业非常具体，是它们的差异化。

You’re talking about the equilibrium state, and a stable equilibrium where companies have this compounding value and can see terminal value for themselves. Another challenge to the considered equilibrium of, okay, there are applications and workflows that are sort of common to a vertical or a horizontal. And this was the generation of SaaS companies, and Microsoft has lots of SaaS properties as well. And then there are things that are very specific to every enterprise that they’re differentiated against.

Elad Gil

你肯定听过、也参与过很多关于「软件终结」的辩论，因为这些工作流现在生成起来很便宜。你觉得未来均衡里，企业里自建的智能体，和供应商那边建的智能体，看起来会不一样吗？

I’m sure you have heard much and participate in much of the debate about the end of software because all these workflows are cheap to generate now. Do you think the equilibrium looks different between what agents get built in enterprises versus in their vendors in the future?

Satya Nadella

我们曾经有一种把工作流捕获进应用的方式：我们建一个数据模型，把某部分业务流程图式化，再堆业务逻辑，再在上面放 UI。加上一点配置。大概二十年，计划就是这样。现在有意思的是，你要重新诉讼这一整条垂直堆叠。我仍然觉得，每个 SaaS 应用底下那个数据模型非常好。何必重造？我的总账最好还是总账。我不需要新的 schema 创造。那个实体关系其实相当稳，我想喂给它，你也希望它稳定。业务逻辑也一样。看 Power BI：人们造了铺天盖地的仪表盘，仪表盘底下的美是非常丰富的语义模型。有人吃了苦去造仪表盘、做所有度量，你想要那个。那就是业务逻辑，我希望它对我可用。所以 SaaS 商业模式的挑战是：我们曾经用一种方式打包。我们现在必须学会把这些东西拆开，再用新方式重新打包，并发现新的商业模式。

Yeah. So I think what’s happening there is, we had a particular way we captured workflow in apps. Because we built a data model. We schematized some part of some business process. We then built a bunch of business logic. And then we put a bunch of UI on top of it. And a little configuration. For like 20 years that was the plan. So interestingly enough, now you kind of get to re-litigate that vertical stacking. So I still think, for example, that data model that you built underneath every SaaS application is super good. Why reinvent it? My general ledger better be a general ledger. I don’t need new schema creation. In fact, that entity relationship is actually a pretty good, robust thing that I want to feed. And you want it to be stable. Then same thing with business logic. If you look at Power BI, it is like dashboards galore people created. The beauty underneath that dashboard is a very rich semantic model. Someone took the pain to create a dashboard and do all the measures, and you want that. That’s business logic. I want that to be available to me. So I think the challenge of the SaaS business model is we packaged one way. We now have to learn how to unbundle these things and rebundle in new ways and discover new business models.

看看今天微软 365 正在发生的事就是好例子。我们有一个叫 Work IQ 的东西。我们意识到：天哪，如果你看历史平行——我们先卖 Exchange 和 SharePoint，Teams 之前还有 Lync Server 之类，我们以为那都会迁到云上。但我们几乎没意识到，会在云里用服务器的人数是 10 倍、100 倍，因为人们不是在买服务器，只是在买订阅。同样的事现在发生在 M365 上，因为有了 Work IQ，我们暴露出也许是一家公司里最重要的数据库，而它从未被当成数据库用，因为它只被我们的应用俘虏。邮件在上面操作，Teams、Word、Excel、PowerPoint、SharePoint 都在上面操作。现在我能用 Work IQ 做的最酷的事之一：我走到一个 GitHub 仓库说，嘿，上周我参加了一些跟这个仓库相关的设计会。你能不能把那些都抓下来，告诉我该改什么？它真的能去看那些转录，回来带一份改代码库的计划。以前你根本不会想到拿 M365 做这种事。所以智能体世界里的价值创造机会其实是 10 倍以上，但确实需要我们去重构。比如 M365 周围会有用量，也许比终端用户还多；而我用来服务收件箱或邮箱的那套，不能用来服务智能体。我们正在做的就是这个。

I mean, if you look at it, what’s happening today with Microsoft 365 is a great example. We have this thing called Work IQ. In fact, what we are realizing is, oh my God, if you look at — in fact, there’s a historical parallel too. We sold first Exchange and SharePoint and, before Teams, we had a thing called Lync Server and what have you, and we thought, oh, that’s all gonna move to the cloud. But little did we realize that the number of people who will use servers in the cloud is 10X, 100X, because people were not buying servers, they were just buying a subscription. The same thing is now happening with M365 because with Work IQ, we have exposed what is perhaps the most important database in a company that never got used as a database because it was only captive to our apps. It was all email operated on it, Teams operated on it, Word, Excel, PowerPoint, SharePoint. But now, like this is one of the coolest things I get to do with Work IQ. I go to a GitHub repo and I say, hey, I attended a bunch of design meetings last week related to this repo. Can you capture all that and tell me what changes I should make? I mean, think about that. It literally can go look at all those transcripts, come back with a plan to change a code base. Previously, you could never have thought of using M365 for something like that. So the value creation opportunity now in the agent world is in fact 10X more, but it does require us to have — for example, there’s going to be usage around M365, which is going to be perhaps more than even the end users, and we have to even re-architect. Like, in fact, what I use to serve an inbox or a mailbox cannot be used to serve an agent. And so that’s sort of what we are doing.

Sarah Guo

我不相信这些领域有永久的商业模式，但近端你怎么看：按结果计价、按 token 计价、企业打包，你会押哪边？

I don’t believe in permanent business models for any of these domains, but in the near term, do you have a prediction between outcomes-based pricing, token-based pricing, enterprise bundles?

Satya Nadella

我一直这么想：就拿按人计价来说。按人计价其实是有人要做预算、需要确定性的产物。某人要预算，就需要按人。而按人只是一组用量权利。所以第一种打包会是：拿一些用量，打进按人档，再卖订阅。订阅会在，按人会在。下一件大事会是用量。人们会说，我要用量。也有可能人们说，我甚至不想为任何订阅或用量结果付钱。但记住：大多数人喜欢结果，直到他们真的有了一个结果，因为一旦有了结果，那就好像把特许权分出去。我跟喜欢按结果计价的客户谈过，我说我全押，直到他们说：天哪，你在分享我的结果？不，不，不。我要你回到按人计价，我要你按用量计价。所以这场辩论会继续。这些商业模式都有各自的时间和地点，没有一个统治全部。如果你是 SaaS 厂商或平台厂商，有这种灵活性——我们在 GitHub 上就面对这个。我们最近刚宣布 GitHub 的按人计价，因为 GitHub Copilot 当初是按人构建的，那时我们甚至还不理解智能体用量的强度。它是开发者做代码补全、也许做任务的交互方式，不是我启动一万个智能体跑一整天。调整的就是这个。所以会永远有按人，但也必须有用量计量表。

The way I think about this is always we’ve had — let’s even take the per-user pricing. The per-user pricing is really an artifact of someone creating a budget needing certainty, because it’s the most important thing. Somebody wants a budget, they need a per user. And per user is just a set of entitlements to usage. And so the way is, the first bundling will be take some usage, bundle it into per user stacks and then sell subscriptions. So subscriptions I think are gonna be there, per user is gonna be there. Then the next big thing will be consumption. So people will say, I want consumption. And it’s also possible that people will say, I don’t even want to pay for any of the subscriptions or the consumption’s outcome. But remember, most people love outcomes until they have an outcome, because once you have an outcome, it’s like giving away royalty. I’ve talked to customers who love outcome-based pricing, and I say, I’m all in, until they, oh my God, what are you talking about? You’re sharing in my outcome? No, no, no. I want you to go back to per-user pricing, and I want you to consumption price. So I think that debate will go on. But all of these business models have a particular time and a place versus one to rule them all. And if anything, if you’re a SaaS vendor or you’re a platform vendor, having that flexibility — and quite frankly, we face this with GitHub. We just recently announced a per-user pricing on GitHub because GitHub Copilot was constructed at a per-user level before we understood even the intensity of usage of agents. It was an interactive way for a developer to use code complete, maybe tasks. It was not like, oh, I launched 10,000 agents that are going on all day. So that is what the adjustment is about. So now that we really want, there will always be a per user, but there will have to be a consumption meter.

Sarah Guo

你怎么看 SaaS 更一般的耐久性？我观察到很多企业内部会有团队几乎有一种智能体亢奋，对能建的东西爆炸太兴奋，于是想重做大量应用，或去跟 SaaS 厂商说我们不再跟你合作了、我们在考虑内部项目。六到九个月后，也许有些人会回来说，其实我们没法把所有东西都重做。你觉得这个世界里什么耐久、什么不耐久？

How do you think about the durability of SaaS more generally? One thing I’ve observed is in a lot of enterprises internally, there will be teams that almost have agent euphoria. They’re so excited about the explosion of things they can build that they’re trying to rebuild a lot of applications or going to their SaaS vendors and saying, we’re not gonna work with you anymore, or we’re considering an internal project. And it seems like in six to nine months, maybe some of those people will come back and say, actually, we can’t rebuild everything. How do you think about what’s durable in this world and what isn’t?

Satya Nadella

我觉得我们必须走完一个完整的预算周期，才能真正看到均衡的出现。因为说到底，即便生成应用也有边际成本。一个简单说法：如果自建并维护某样东西的边际成本更高，你就该去买。这是可量化的。维护这部分很重要。你得记住，AI 现在会找出的所有安全问题，你也得修得很快。当然有编码智能体帮忙，但那会烧掉 token。那是谁的责任？这是一个你必须想清楚的循环。我们已经走过「我能生成很多软件」的兴奋。下一件事会是：我真正想生成什么软件？我想用别人的什么软件？我如何把这两者组成某种我拥有能动的智能体工作流？我觉得对任何在厂商层面不灵活的人，容忍度会非常低。但同时，谁有那种灵活性、出现、交付价值，会再回来。我们还在卖软件，只是商业模式不同。

Yeah. I think we have to go through one full budget cycle on this to really see the emergence of the equilibrium, because at the end of the day, there’s marginal cost to even generating the app. In fact, there can be even a simple way to say it, like you should always acquire something if the marginal cost of building and maintaining something on your own is higher. That should be a quantifiable thing. And the maintenance part is important. You got to remember, hey, all the security stuff that now AI will find, you better fix them too fast. Of course, there’s a coding agent to help you with, but then that burns tokens. So whose responsibility is it? It’s kind of like a cycle that you’ve got to think through. And I think we have gone through the excitement that I can generate a lot of software. I think the next thing would be what software do I really want to generate? What software do I want to use from others? How do I compose these two into some agentic workflow that I have agency over? Because I think there’ll be very little tolerance for anybody who’s inflexible at the vendor level. But at the same time, I think that anyone who has got that flexibility shows up, delivers the value, will be back at again. We’re selling software, but with just different business models, in fact.

swyx

说到建软件，我最喜欢的 Build 瞬间之一，大概是一两年前，有一节是你在建自己的软件。你现在还在建什么吗？

Speaking about building software, one of my favorite moments from, I think, a previous Build maybe one or two years ago was there was a section of you building your own software. I’m curious if you’re building anything now.

Satya Nadella

首先得承认：建软件已经让即便是我们这种公司的 CEO 的无能，也能把东西做出来，谢天谢地。但话说回来，GitHub Copilot，尤其是新的 Sessions 应用，让你对以前觉得碰不到的工件重新拥有能动。对我这个 CEO 来说，甚至走进一个代码库去了解它。我记得很久以前加入微软，人人都得去看 Cutler 的、Malik 的，才能学怎么写好的 C、C++。现在这种上下全栈看东西、学东西的能力太好了，但并不意味着我们每个人都该做同一件事。问题是：你如何有能力去检查、去学、去看。所以对我来说，我大量在建的是这些长程的 Foundry 智能体，autopilot。最容易的是，我上周刚建了一个：能不能有一个智能体持续监控，本质上是我自己的参谋长 autopilot。我们显然会在 Scout 里做，那是我们展示的。但建起来太容易、太琐碎了。我拿了 Work IQ，说：拿 Work IQ，去建一个 Foundry 长程智能体，把记忆存在后端即服务里。结果它不但建了，我还能说发布到 Teams，它就把那东西发到了 Teams。能把这样一个端到端项目做完，相当奇迹。

Yeah. So first of all, let’s face it, right? Building software has made it possible for even the incompetence of a CEO of a company like ours — you can build, so thank God. But that said, I do feel that something like GitHub Copilot to me, and especially the new Sessions app, has just made it so much more possible for you to have agency over artifacts that you felt you couldn’t touch before. So for me as a CEO, even to go to a code base, to be able to learn about it. I remember joining Microsoft long back, you know, first and then you say, man, everybody had to go in and look at Cutler’s, Malik, or what have you to learn how to do good C, C++ code. So now that ability to be more full stack up and down is so good, but that doesn’t mean every one of us should be doing the same thing. The question is: how do you then have the ability to inspect things, learn things, see things. And so to me, what I’m building a lot of is these long-running Foundry agents. Right? So there’s autopilots. So the easiest thing is, to me, I think I just built one even last week, where the idea was, hey, can I have an agent that is continuously monitoring essentially my own chief of staff autopilot. We’re gonna have that obviously in Scout. That’s what we showed. But it is so easy and trivial to build. I took Work IQ. I said, take Work IQ, go and build a Foundry long-running agent. Store all the memory using a backend as a service. And lo and behold, it built it, and not only built it, I could say publish to Teams, and it published the damn thing to Teams. So the ability to have some end-to-end project like this complete is just pretty miraculous.

Sarah Guo

你觉得这会如何影响未来不同类型的工程角色？现在大概有十几种工程师：QA、前端等等。有人争辩说四五年后我们基本上会剩下四种工程角色：管智能体的人、前线部署工程师或 FDE、安全工程师，以及为少数服务做大规模基础设施的人，其余都塌进智能体世界。你觉得这看法对吗？

How do you think that impacts the different types of engineering roles that exist in the future? Because right now I think there’s a dozen different types of engineers that you can be, from QA, front end, et cetera. I’ve heard some people argue that in four or five years we’ll basically end up with four engineering roles. It’ll be people who are managing agents, it’ll be four deployed engineers or FDEs, it’ll be security engineers, and then people working on large scale infrastructure for a small number of services, and then everything else just collapses into the agentic world. Do you think that’s a correct view of the world?

Satya Nadella

我觉得我们必须实验着走过去。但你说的，有些非常大规模的事。在 LinkedIn，他们结构性改了，建起一个叫 full stack builder 的新学科。他们把设计、产品管理、前端工程的人放在一起。但也有锋刃：设计的人仍然有设计锋刃，前端的人仍然有前端锋刃，只是你可以给自己更大的角色范围，不被锁在一个角色里。同样，基础设施变得非常关键。我们意识到，即便对 Excel 团队也是：建一个可以学习奖励的 RLE，实际上是最难的基础设施问题之一。所以你需要新人才：在曾经被认为是终端用户应用的团队里，也要分布式系统的人，因为技能集不同。是的，基础设施；科学是另一个。我们会看这些怎么演化。世界总会有一堆专家。通才角色会最令人兴奋，因为通才的杠杆，是我们会看到最大回报的地方。你问我在不在写代码，我现在是通才——我把知识工作翻译了出来，以前我做 Word 文档或表格，现在我能做一个应用，是同一句话。那种「我的通才技能杠杆变高了」的感觉，我们会全面看到。

Yeah, I think we’ll have to experiment our way through it. But what you said is what — there are some very at scale things. At LinkedIn, they did structurally change and basically built up a new discipline called full stack builder. So they went and said, hey, let’s bring people from design and product management, front end engineering, all put them together. But also have an edge. It’s not like the design person still doesn’t have the design edge, or the front end person doesn’t have the front end edge, but you can give yourself bigger scope in roles so that you’re not confined to one role. And then equally, infrastructure has become very critical. So in other words, RLEs — one thing we’ve realized is even for the Excel team, for example. Building the RLE in which a reward can be learned is actually one of the hardest sort of infrastructure problems. And so you kind of need even new talent. Distributed systems people even in what was considered an end user app team, because it’s a different skill set. So yes, infrastructure, science is the other one, obviously. So I think we’ll see how these evolve. Always the world will have a bunch of specialists. I think the generalist role is going to be the most exciting, because the leverage of a generalist is where we are going to see the maximum returns. When you said, hey, are you coding? I’m now a generalist. I’ve basically translated knowledge work, which I did, where I created a Word document or a spreadsheet, and now I can build an app. It’s in the same sentence. That idea that, oh wow, my generalist skills have gotten higher leverage, I think is what we’re gonna see across the board.

Sarah Guo

对 CEO 和 VC 来说，这话很好听。想法者的黄金时代，带着大量能动。

Music to the ears of CEOs and VCs. Golden age for idea people. With a lot of agency.

Sarah Guo

如果把个人能动放大到组织语境，我的合伙人写过一篇文章，一个大带走是：这是一个你可以更有野心、也必须更有野心的时代，因为环境和用户、公司接纳新技术的速度都很快。问一个管着万亿美元以上公司的人好像有点傻，但你怎么看微软现在如何能更有野心？

If you take that idea of personal agency and you just zoom it out to the organizational context, my partner, who actually started his career at Microsoft, just wrote an essay where one of the big takeaways is it’s an age where you can be much more ambitious, and you need to be, given the pace of the environment and how quickly, actually, users and companies are open to adopting new technologies. I feel silly asking this of somebody running a trillion-dollar-plus company already, but how do you think about how Microsoft can be more ambitious now?

Satya Nadella

这是好问题。这类迁移里，要有一个关于工作如何改变的概念模型，去追求你以前几乎无法想象的结果。Kevin Scott 有一句很好的话：当你在把难的事变容易，那是一种杠杆；但真正的野心是让不可能变成可能。我们所有组织里现在缺的一点，就是那个新的概念模型：我们能建什么？什么曾不可能、现在能建？我举一个例子。我从管 Azure 网络的人那里得到很大启发。这甚至是去年的事。我们在扩。你看到我讲过，过去 15 个月我们建的 Azure 产能，超过我们前 15 年建的。疯了。而且是同一支团队。他们看到这个，说：Bob，如果我们不重新概念化我们的工作，这就行不通。于是他们建了——本质上他们说，我们的工作不是去做 Azure 网络，我们的工作是建那个去做 Azure 网络的智能体系统。这些人管着五百多个光纤运营商、管着全球 WAN。光纤运营最终是物理运营：东西会被切断，必须修。我们有 DevOps 这种漂亮词，基本上是邮件进来，你得去响应、处理。所以他们建了这个智能体系统，甚至给它一个角色，叫 Miles，它就做这些事。他们开始喊要更多 token。他们说：听着，我们不需要编制。我们需要 token，才能管我们的运营。那种对工作的重新概念化——他们把工作做成了元工作，元工作现在是他们的新工作。

It’s a great question. I think the thing in these type of transitions is to have a conceptual model of how work can change to go after outcomes that you could hardly imagine previously. In fact, Kevin Scott has this nice line, which is, when you can make the impossible — like, when you’re making hard things easier, that’s sort of one point of leverage. But true ambition is about making the impossible possible. So now the thing that is missing a little bit in all of our organizations is what is that new conceptual model of what can we build? What was impossible and what can we build? And I’ll give you one example of this, which is I take great inspiration from sort of the people who were managing the Azure network. And they came to — this was from even last year. You know, we were scaling. You saw that I talked about sort of how we built in the last 15 months more Azure capacity than we built in the first 15 years. I mean, it’s crazy. And it’s the same team. So they saw that and they said, Bob, this just ain’t gonna work if we don’t reconceptualize our work. So they built — essentially they said, our job is not to do Azure networking. Our job is to build the agentic system that does Azure networking. These are the folks managing the 500-plus fiber operators managing the WAN, all over. And fiber operations ultimately is a physical operation. Things get cut, things have to be repaired. You know, we have fancy words called DevOps and so on. Basically, emails are coming in and you gotta go respond to them, take care of it. So they built this agentic system. They even have a character for it. It’s called Miles, and it sort of does all this stuff. They started sort of screaming for more tokens and so on. And so they were saying, look, we don’t need a headcount. We need tokens in order to be able to manage our operation. That reconceptualization of what their work is — they basically took their work and made it meta. That meta work is now their new work.

八十年代如果有人跟我们说，四十亿人早上起来开始打字，我的模型会是：我们需要四十亿打字员？但我们不是在打字，我们是在做知识工作。无论是微软还是任何组织，就是给自己许可，去做新类型的元认知、元工作，用这些新工具去改变真正重要的产出，然后真的让不可能变成可能。把这些点连起来，我认为大量企业价值会在那里被创造出来。

Right? In the ’80s, if somebody had come to us and said, 4 billion people are gonna get up in the morning and start typing, my model would’ve been, we need 4 billion typists? But we’re not doing typing, we’re doing knowledge work. So that, to me, I think is it, which is whether it’s Microsoft or whether it’s any organization, is to give ourselves permission to do new types of metacognition, meta work, using these new tools to change the outputs that matter, and then really make the impossible possible. So completing that dot or the connective tissue across those, I think, is where a lot of the enterprise value will get created.

Sarah Guo

我们该谈谈数据中心吗？微软的扩建规模，以及其他所有人的，正在重新定义什么叫超大规模。财务上、公司运转方式上，还有被影响的社区上，都是前所未有的规模。你在现场看到什么？

Should we talk about data centers? This leads nicely into the data center build-up. I’m just impressed at the sheer scale of the build-out from Microsoft, but also everyone else, that this is redefining what it means to be a hyperscaler. And I just feel like that is at unprecedented scale on finances, on the way you run the company, but also the communities that are impacted. Just talk a bit more about what you’re seeing on the ground, like when you visit.

Satya Nadella

我觉得有两面。扩建当然非凡。没发生过这样的事，能成为参与者之一很好。但你提到了另一面。到了这个节点已经很清楚：除非我们作为行业非常有原则，确保我们谈的所有这些好处，以真实的方式在社区层面被感受到。这不能只是一场宣传。必须是真的：人们会说，这没有在改我的能源价格。事实上如果有的话，它在把价格打下来，因为长期会有更好的电网、会有更多能源。水的消耗——事实上水正在被补给。你必须真正去教育人们到底在发生什么，我们在建的闭环系统。我们必须投资培训、工作、税基。最少被谈到的，是建设期间、建设之后创造的大量工作。社区里的税基是什么？这些都必须是真的。如果是那样，我们会有许可。如果不是，我们就不会有许可。就这么简单。我们必须作为行业非常认真地对待。社区保持怀疑、问硬问题，对我们做艰苦工作、去挣这个，是好事。

Yeah, I think there are two aspects of it. Obviously, the build-out is extraordinary. Nothing like this has happened, and it’s great to be one of the participants in it. But you brought up the other part. I think at this point it’s clear that unless we as an industry are very principled about ensuring that the benefits of all the stuff we’re talking about are felt in real ways at the community level. Because this is not just a campaign. It has to be real, where people are saying, look, this is not changing the prices on energy for me. In fact, if anything, it’s bringing down prices because long term there’s going to be a better grid, there is going to be more energy. Water consumption is, in fact, water is being replenished. You gotta really educate folks on truly what’s happening, the closed loop systems we are building. We have to invest in the training, the jobs, the tax base. In fact, the least talked about stuff is the amount of jobs that get created during construction, after construction. What’s the tax base that’s there in the community? And all this has to be real. And if that is the case, then we will have permission. If it is not, we won’t have permission. It’s as simple as that. I think we have to take it as an industry pretty seriously. I think it’s good for communities to be skeptical, ask the hard questions, for us to do the hard work, earn that.

但说到底，我一直觉得人类历史上，如果你用大量能源、也为社会创造大量价值，故事会非常好。如果你不那样做，就没那么好。这一次，我坚信，最终如果你真的有一个驱动生产力、驱动经济增长、驱动广泛参与、更好健康结果的 token 经济，我们就会处在一个很好的位置。那至少是我们都必须聚焦的。

But at the end of the day, I’ve always felt like in human history, if you use a lot of energy but also create a lot of value for society — the story has been fantastic. If you don’t do that, it’s not been that great. And this time around, I’m a firm believer that ultimately if you do have a token economy that drives productivity, that drives economic growth, that drives broad spread participation, better health outcomes, then I think we’ll be in a great place. And that’s at least what we all have to be focused on.

Elad Gil

你目前对社会影响最乐观的是什么，或者你个人模型更新最多的是什么？

I guess I wanna talk about what you’re most optimistic about currently, or what have you most updated your personal models on regarding societal impact of AI?

Satya Nadella

最关键的，其实还是我们一开始的那个问题：我们需要把故事讲出来、做成真的——每个人都有真正的机会，作为一等参与者进入这个新经济。未来 12 个月、18 个月，我们需要一种方式让人们说：哦，我懂了。会有巨大的能力、巨大的基础设施，但我能看见会发生什么，无论是健康结果这种好处，还是我创办一家创业公司的能力，还是更高效地经营我本地的店。它就在发生，我自己看见了那个好处。对我来说，以路径依赖的方式去挣那份许可，我们等不起。Eli，我现在学到的一件事是：世界会对那些说「相信我们，未来会很辉煌」的科技和科技公司非常怀疑。你必须交付可触摸的好处。坦率说，政客会因为主张那个而赢得选举。那至少是我的调整，因为这一次太重要了，占经济的比重太大，不可能不是这样。

I think the most critical thing is the first question we even started with, which is we need to tell the story and make it real that everybody has a real shot to participate as a first-class participant in this new economy. That’s kind of, I think in the next 12 months, 18 months, we need a way for people to say, oh wow, I get it. There’s going to be tremendous capability, tremendous amount of infrastructure, but I can see what is going to happen, whether it’s the benefits like health outcomes or my ability to create a startup or my ability to run my local store more efficiently. It’s just happening, and I see that benefit myself. That to me, earning that permission in a path-dependent way, we can’t wait. See, the one thing, Eli, that I’ve now learned is I think the world is gonna be very skeptical of tech and tech companies that say, trust us, we’ve got it. The future is gonna be glorious. You kind of have to deliver tangible benefits. And quite frankly, politicians winning elections because they have advocated for that. That will be at least my adjustment because without it — because it’s too important this time around. It’s too much of the economy for it not to be the case.

Sarah Guo

我对 AI 广泛好处的一个很简单框架，除了在科技里工作的社区：大量不同公司、创业公司和大型公司里会发生的财富创造。然后是医疗。你们今天有很棒的演示，也有像 Open Evidence 这样的公司。我觉得那正在发生。教育似乎是另一个明显的好事，但我们还没看到我预期的那么大的影响。你有没有假设，为什么会这样，或者它会不会来？

So one very simple framework I have for what is gonna be the broad benefit of AI, beyond the communities just working in technology, are wealth creation — it’s gonna happen in a ton of different companies, startups and large companies. Then you have healthcare. You had amazing demos today. There are companies like Open Evidence. I think that is happening. Education seems like another one that’s an obvious good where we haven’t seen as much impact as I’d expect. Do you have a hypothesis on why that might be, or if it’ll come?

Satya Nadella

这又回到我们如何思考教育。最近我见了 Alpha School 的创始人，听了很多他们怎么做，听他们如何重新思考教育到底长什么样，非常有意思。我觉得这其实非常重要。我不是说传统在做的更不重要。我还在看斯坦福的某门课，忘了是哪门 CS，亚洲的一些指导——因为你仍然需要人去学习。那是一门有意思的 AI 课，他们确保人们在学怎么恰当地用 softmax，而不是说「嘿，修我的训练」。学概念很重要，会是关键。但我们如何创造激励、什么是凭证、我们如何给这些凭证定价、这些凭证对应的就业机会是什么？所以必须有一场完整的改变，因为获取信息的方式、自学的方式、持续让自己更新的方式，已经变了这么多。有意思的是，也许下一个大的创业成功故事，会是有人建一所新大学，或者一种新的教学法，让人走完一套课程并找到非常有价值的经济机会。那很长时间以来都像不可能，但这是一个很好的结束点，也是可能变得可能的事。

Yeah, I mean, I think this is where, again, how we think about education. Recently I met with the founders of Alpha School and learnt a lot about what they were going about, and it’s fascinating to listen to how to even rethink what does education really look like. Because I think it’s actually very important. And I’m not saying anything traditionally being done is less important. I was even looking at — it’s fascinating to see. I forget which Stanford class it was, the Asian guidelines for CS something. Because you still need people to learn. It was an interesting AI class that they were making sure people were learning how to apply softmax appropriately versus saying, hey, fix my training run. So I think learning concepts is important. It’s going to be critical. But the way we create the incentives, what are the credentials, how we value those credentials, what is the employment opportunity for those credentials? So I think that there’s a complete change that has to happen, given the way to get to information, way to educate yourself, way to continuously keep yourself updated has changed so much. So I think interestingly enough, maybe the next big startup and success story could be someone who builds a new university, or a new pedagogy even of how to get someone to go through a curriculum and find economic opportunity that’s highly valuable. Well, that has felt perhaps impossible for a long time, but it’s a great note to end on and something that might be possible.

Sarah Guo

仍然可能。谢谢你，Satya。

It’s still possible. Yeah. Thank you, Satya.

Satya Nadella

非常感谢。谢谢你们。

Thank you so much. Thank you. Yeah. I appreciate it. Thank you all.