Satya Nadella on AI Agents, Rebuilding the Web, the Future of Work, and more

101.79k views4071 WordsCopy TextShare
Rowan Cheung
Microsoft just revealed its next big AI moves at Build 2025 - from autonomous coding agents, Copilot...
Video Transcript:
I think we as a society celebrate tech companies far too much versus the impact of technology. think about the decades all of us have spent saying, can we have one tech intervention that makes a damn difference in education, we are definitely trying to build a scaffolding for the AI age. I'm curious what you think the world will look like when 90 or 95% of all code is generated by AI?
Ultimately the human is in the loop, I think we overstate the autonomy here. Remember at Microsoft we are not like there's one trick company. the hardest thing for us is being adjusting.
The reality is, case studies don't help. You have to do it yourself technology is powerful enough to disappear. This is something that recently went viral.
you mentioned that AGI is just some nonsensical benchmark hacking, and the true value of AI is. . .
All right. Satya, thanks so much for being here. So we're just minutes after you just hopped off the stage at Build.
Can you explain how all these new developments really are coming together to build that new agenetic web? Yeah. First of all, thanks so much for being at the developer conference.
And you know, for us, the it's it's interesting, you know, when I think about this, we're at that moment, right? Whatever. 2 or 3 years into a platform shift, and you start talking about the platform, not just about the one app or the few apps and how they got built.
Right. Because you're about to sort of start scaling, things in a much more. I'll call a generalized way where as a developer, you can start seeing how to build it.
I mean, one of the best things, that we talked about was that Stanford demo, right? Stanford Medicine, think about it. Right.
Something high stakes, like a tumor board meeting. How do you use AI to have better tumor board meetings? That's a real thing that one can go to work on.
In order to make that the case, though, you really have to get data from pathology from, you know, multiple labs from PubMed and all getting orchestrated by multiple agents. Right. And then have it show up where users are.
In that case, it was teams. If you're a teacher who goes to a tumor board meeting and then wants to go to a, you know, class as a teaching, doctor, you want to be able to turn that into PowerPoints. And that type of orchestration is what you want to be able to build.
And I feel to do that, you got to build up a real stack right where every layer is open. It composes. There are standards that protocols, and I feel like they're finally there.
And so what is exciting to me is to have the stack from sort of, Microsoft 365 Copilot to Foundry, all compose, with some of these things like NL web and MCP to create this agentic web. In some sense, I feel like even the original ethos of the web, maybe we can rediscover which is the true openness. Yeah.
And it sort of seems like Microsoft's trying to build this all in one UI, where you can manage fleets of agents on one spot. Is the goal a world where every knowledge worker effectively becomes this agent manager instead of just a knowledge worker? Yeah, it's it's a metaphor.
I think it that makes sense, right? When we say we are definitely trying to build a scaffolding for the AI age, right? Just like, say, back in the day, we built teams, right?
You remember? Or even back in the day. We built outlook right before outlook came together.
Calendaring was a separate app, contacts was a separate app and email was a separate app. And so then we said, oh, maybe we should build it all together. And outlook was born.
And then teams or meetings was somewhere else, channels or somewhere else. And then chat was somewhere else. And we said, hey, let's bring that together.
So some, you know, I mean always there's going to be a need for a new scaffolding. And in our case, with what I describe as a UI for AI is you say, well, let's bring chat, let's even bring search. And then our agents and things like notebooks and so on, all together in one interface.
And that's what we're doing with the M365 Copilot and teams is essentially the multiplayer version of it. But it's not the only I'll call it UI for AI. There will be many other places and many other developers will build them.
Right. You could say the UI for AI for developers is get up. UI for AI for some scientist will be, you know, what we talked about in our discovery or some third party application.
So I think there's going to be a lot more richness on the UI layer where people will build different modalities for different workflows and different needs. But the interesting thing is the underlying capability, right where you have, hey, you have data, you have multiple models, you have these agent orchestration layers, you have these reasoning models now that are capable of taking intents and decomposing it into multiple calls to multiple models. That's I think, the exciting thing.
You're the world of knowledge. Work is just changing so fast and surely with any tech revolution there will be some job displacement. What advice would you give to knowledge workers.
So that they effectively become these agent managers instead of agent replacement? Yeah, I think it's a good way to phrase it, because I think it's sort of always helpful for us to separate out, the knowledge work go from the knowledge work of today. Because if you think about it, right, if you're an alien intelligence had come to, and seen the world, you know, at, at, let's say, at work, even early 80s, and they said, oh, well, there's a typist pool.
There is a slide pool or a slide making pool. And then if they came back, today, they'll say, oh, wow, all of humanity is a typist pool. Because everyone gets up in the morning and with the two thumbs or, with multiple fingers are all typing.
But we are doing knowledge work. It's being abstracted. Right.
So therefore, I think the levels of abstraction, whether it's managing agents, if I look at even the simple, simple workflow from when I joined Microsoft in 92 to now, if I had if, say, a somebody had to get prepped for a, let's say, a customer visit, you know, you would sort of go to the account team. The account team would write up a report, the report would show up in the email that would get loaded into one note, and I would read it. In fact, that was pretty much the sort of, you know, I was preparing for most of my career.
Those reports and then eventually, you know, reading those reports a lot in the last even 45 months, thanks to these things like reasoning models like researcher, it's completely inverted. I just go prompt myself, hey, I'm meeting with the CEO of XYZ Corporation. Pull all the stuff I need to know.
It pulls from the web. It pulls from my emails, my documents. Most importantly my CRM system, my supply chain system.
Because most of these folks are both customers and partners of ours. Gives me one comprehensive report and then I share it with the accounting. Right.
So think about it. The workflow is inverted. So you could say yeah, I as a CEO quote unquote, I'm doing more knowledge work than I was doing, quite frankly.
I am more employable today because I feel more empowered. I feel that I can get to information faster, collaborate with my colleagues in the company. So I do feel the best thing that can happen is diffusion in this phase.
I think any knowledge worker, whether you're in software, whether you're in just horizontal knowledge work, in finance and sales or in science, use the tools, change the work. Have you have the agency to change your work artifacts, the workflow around you? And of course, you know, I'm clear eyed that there is going to be displacement.
And so therefore, the best defense against that is skilling reskilling. And it starts by using tools versus, not using them. And developers arguably have seen the biggest shift to their workflows with the rise of AI tools.
And now agents and GitHub new Copilot coding engine is just another example of another tool that's, you know, effectively changing the developer work workflow so much. You've actually seen recently said that Microsoft is sharing 30% of all new code with AI. I'm curious what what you think the world will look like when 90 or 95% of all code is generated?
AI yeah, it's a couple of different things, you know. One is you got to remember, like when I look at the amount of deficit we have, right. If I sort of looked at, let's call what people say is tech debt or it debt, right.
Which is if you look around the world and say the number of projects that are still unfinished, the reality is we need a lot more software development, in order to be able to really reach, the goal of being able to work down that deficit. So you start there. We have a problem.
That is, we don't have all of the software development capability and supply in order to meet the demand. Then you say in that context, that's why the all these form factors, right. Even code completions super helpful.
We always had intellisense, but we now just have better intelligence that works with code completions. Or I can highlight a bunch of code and have it explain it to me and it'll draw a diagram. You and yesterday I like took some code and sort of just highlighted and said, just give it to me as a flowchart, like just that simple thing I just because I do better with visual stuff.
Then you know agents right. Where I can assign synchronously some tasks. So multi file edits, the full repo changes that I want done.
Right. So that's all as an individual developer I can stay in the flow. I can use these tools then to be able to today the launch of the coding agent is pretty cool because you know like as I said, like even for me and it's no longer sufficient to for me to just file bugs.
I got to get involved in even fixing bugs, right? To the, I mean, a programmer that I can work with. And of course, remember, you know, ultimately the human is in the loop.
I think we overstate the autonomy here, right? Because even before it does any CI CD thing, it comes back for a human review. Right.
And it can be, you know, something that you can even automate with another agent. But ultimately, there is a workflow here where people inside of a development organization are going to work with these AI agents to essentially work down the deficit we have. That's the world I sort of see, more so than anything else.
Copilot fine tuning was another really interesting announcement, giving enterprise the ability to unlock, you know, their, their own data and fine tune their own code parts with it is a major unlock. I'm curious where you think these domain specific agents will really excel, and is there any types of proprietary data that will give companies a real advantage over just like generic Copilot? Yeah, it's a it's a super important question, right?
I mean, in some sense, what is the form, going forward and what is that edge? The knowledge edge of a form is the real question. And that's why I think I'm really excited about this.
Copilot. Fine tuning, because the idea is to be able to take what you have is a form, which is your knowledge, your data, and use it to essentially tune, the copilot system. But here's the interesting thing, right?
You asked, which is what's the sustainable advantage here? And the sustainable advantage is to get a new sample to then use these reasoning models with your data to then be able to do RL in the real world. Right?
So which is the reward function from the market, reinforcing the application of your knowledge to fine tuning the new sample. Right. That's the virtuous cycle.
So the world keeps getting better in terms of its model capability. Forms are about taking that. That's ultimately a commodity that you bring into the firm.
You then because of all the knowledge work you do internally and the data you have, fine tune it, put out the output into the world, get the signal. It could be a customer thumbs up. It could be marketplace thumbs up, whatever.
That's the reinforcement. Reward. And then you go back.
And that I think, is sort of the new theory of the form in my sort of, you know, sense of how one needs to go and you really are going to have to perfect that loop. You've led Microsoft through multiple tech transformations. What advice would you give to companies trying to restructure around this agent era?
you know, with in in tech companies in particular, in Microsoft in particular, because remember at Microsoft we are not like there's one trick company. In other words, at some level the hardest thing for us is being adjusting. But when I came to Microsoft, you know, at that time, it is unclear, like our existential competitor was Novell.
And so over the years, Microsoft has had to learn in the hard way. Sometimes we've gotten it right, sometimes we've gotten it wrong. How do I fundamentally change three things how we work, what we work on, and how we go to market.
Right. And that muscle of fundamentally reinventing the production function simultaneously with the product and the innovation, as well as the go to market in the business model is a hard thing. It's a harsh thing, because in some sense, I wish sometimes I worked on one product that lasts, you know, multiple decades.
For us, it's not been the case. Even something as big a hit as windows, you know, stopped growing at some point. And so therefore, it means it meant that we had to build other businesses.
But I think ultimately it comes down to culture and capability building. That allows you to take more shots on goal. And that, I think, is what one needs to do as any one of us faced with these shifts, if you have a stable business.
Fantastic. Then this is tailwind. You can you can take and get more leverage out of it.
If you have a business that's in decline, you can then use this as an opportunity to reinvent yourself. but ultimately it comes down to this culture capability and then hunt for new concepts that I think, is what you need to practice. You can't like.
One of the things I think a lot about is, you be all like, I do this as well. We all sort of look at the, you know, the shining company of the era. The reality is, case studies don't help.
You have to do it yourself right now. Somebody said to me once, which is you don't get fit by watching others go to the gym. You have to go to the gym.
And that, I think, is what this change is all about, right? It's not about admiring someone else. But it's about really doing the hard yards yourself.
I want to touch on upskilling kind of from the top down. So from my perspective, my company, we work with a lot of different enterprises, kind of consulting them on their strategy and what we've seen is that there's a lack of personalization in education. For example, like the tools and workflows a marketer might use are different than a developer or someone in Customer Success.
So I'm curious how Microsoft is kind of educating their employees from the top down on, on this sort of personalized education. It's a great point. So one of the things that we're emphasizing is getting the tools in.
Right. Because if you look back even like one of the things I'm heavily influenced by is how PCs became standard issue. Right?
Is it fascinating to study even how PCs penetrated the enterprise? First, you know, someone in legal said, wow, this is great to write contracts up because I need word, you know, with all of its footnotes and what have you finance obviously loved Excel, right? Because, wow, it can help me do models and so on.
But then it turned out that people don't work in their silos. They work in teams. And so then at some point, like even the way we did forecast back, you know, pre PC, I you know basically faxes went around.
Somebody did an interoffice memo. And then eventually, you know, you got to a forecast where, you know, once the PC became standard issue, you just took an Excel spreadsheet, entered some numbers, sent an email, all your area leads entered numbers, and you had a forecast. It changed the work.
The work artifact in the workflow. And it happened not by going to a training class. It actually happened by diffusion of the general purpose tool.
In that case, PC in an office. So a little bit of sort of what I'm seeing even inside of Microsoft, whether it's with GitHub Copilot or whether it's with the M365 copilot or anywhere else, it's just the diffusion of the tools. I'll just give you, you know, one of the best examples at Microsoft was I was talking to an engineer in our networking.
I went right. I mean, I was talking about, man, the AI when is growing like crazy. The number of fiber operators we work with is like some, you know, insane number.
And by the way, all of the DevOps on it, right. You know, if there's a fiber cut somewhere, it's all a manual, right? Literally it is emails and calls and so on.
So she just said, okay, I can't deal with this. I have to really automate this because otherwise, you know, all of Microsoft will be just working on DevOps of fiber. And so she just built an agent and in fact, a multi-agent orchestrator.
Right. And by the way, she just did this using some of the low-code, no code and the foundry tools. Right.
That ability to empower people, with tools and agency to go and look for the workflow closest to them, that I think is what needs to happen, just like how we were able to use, you know, I think I always say Excel is the greatest programing, most ubiquitous programing tool on the world. And all of us learned how to program in Excel, essentially. And that's the thing that I think is going to happen again.
I know we don't have much time here, but I want to quickly touch on proactive agents. So behind the scenes, we actually got a demo of like on device proactive agent working on the Copilot plus PCs where, this agent essentially opened up, outlook and summarize email without an internet. But that's just one example.
Are there any examples that you have that you're really excited for? Yeah. So I think that if I look at the continuum, right, so I, I'm really looking for in some sense I'm a big fan of this Mark Weiser quote about ubiquitous computing, right, where technology is powerful enough to disappear.
Right. So therefore, to some degree, the march of natural user interface and so on is about doing all the things that I intend to be done, with the least amount of, I would say friction. Right.
So that's why I think when I want, like, interpret my intent and plan and get it done. Right. So that's where I think some of these proactive agent stuff comes in.
And I want to be in control though, like, for example, that's why I think the, the cool agents, for example, the computer use agents have to sort of get to a point, one that can't be fast and better, and more reliable. But the more importantly, they have to be changed as part of some intent, of mine. Right.
It's not like I'm going to go give task. Task, task tasks. I'm going to have a high level intent, and then someone's going off and doing a bunch of work and I can inspect.
So that's why I think like this GitHub, you know, coding agent, the session log, it's pretty cool. Like I go back after having a sign. It's like I'm watching the session log with all the draft commits.
Right? So that I think is the balance. Right.
So in an interesting way, we talk about the proactive agent and a session log. And that way then it's sort of more transparency of it. And then when you invoke multiple of them you are able to still track effectively what's happening.
One last question for you. This is something that recently went viral. About two months ago, you said you mentioned that AGI is just some nonsensical benchmark hacking, and the true value of AI is global economic growth.
Where do you see agents having the most value in that regard? First, it's a great one. I mean, to me, take health care.
What is it, 19-20% of our, GDP in the United States is in health care. A lot of that cost the inefficiency. If you talk to any of our providers, is all in the workflow, right?
They're working so hard to provide unbelievable care, except sort of, you know, and they're using all these digital systems, but they're really not able to, tame the complexity and reduce the cost. And so if there is going to be one place where I would love to love to see that productivity gain, would be like that, you know, that multi-agent orchestrator for health care that we just launched, used by Stanford become ubiquitous. And there are stories that people are saying, oh wow, this is been the biggest thing that, you know, every provider is able to provide better health care at lower cost.
Right? So something like that has to happen because my only comment there was not about any sort of shot on any great AI research. It is more, I think we as a society to celebrate tech companies far too much versus the impact of technology.
Quite honestly, if there was a more balanced way to talk about sort of not the tech industry, but the use of technology, right. That's why even the, you know, copilot being an end to end, think about the decades all of us have spent saying, give me a one tech intervention that makes a damn difference in education, right? Finally, you have real statistical evidence, thanks to that world Bank study in Nigeria that, you know, you give people copilot their or any other agent.
It doesn't matter. Something changes. That's the story to me, right?
That's why I joined the tech industry. I mean, there is a very different time. And suddenly it became like, you know, the place we were celebrating ourselves.
And I just hate it. I just want to get to a place where we are talking about the technology being used and when the rest of the industry across the globe is being celebrated because they use technology to do something magical for all of us, that would be the day. Okay, well, I think that's all the time we have.
Thank you so much again for doing this and congrats on the great keynote. thank you so much. It's such a pleasure.
Great. Thank you.
Copyright © 2025. Made with ♥ in London by YTScribe.com