Satya Nadella on the AI Doomer Slowdown, Microsoft's Master Plan & Who Wins AI

0:00is generated 250 billion dollars with a B in market value for Microsoft. Datsy Nadella Chairman and CEO of Microsoft. Inchube in the CEO, three and a half years, the stock is up about, I guess it's about 120%. I'm good for my 80 billion. I'm going to spend 80 billion dollars building out Azure. Maybe after the industrial revolution, this is the biggest thing. That's our goal with our frontier model. Our model should be the best model that they can use as a base. We create technology so that others can create more technology. That's who we are with, tool maker. Please welcome Sartia Nadella. All right. I got it. I'm coming out. Good morning, guys. How are you? Good. Thanks for joining us. Crazy weekend, but here we are. Do we need to paste the frontier? All right.

1:00That's... So, let's start with the common sense part first, which is we should do what it takes to build stuff that serves humanity first and is in human control. You know, it's kind of crazy that we have to start with that level of common sense, but I think it's a good place. Then, when I think about pacing whatever, the first thing that I at least I believe is the broad diffusion of this technology is the most critical thing, because the benefits of this tech showing up everywhere is really what's all about. So, at the end of the day, if you sort of say serving humanity, let it actually reach humanity in ways that it serves humanity. And that means you've got to have choice, you have to have competition, you have to have all kinds of business models, whether they're open weights, close weights, what have you. Then, the other aspect I think that is not talked about when we talk about control is actually the control that, for example, customers have enterprises or businesses have.

2:10Around this technology, because sometimes this is so opaque, right, I want my privacy. I want to be able to embed my knowledge in a set of weights, I control. I want to see all of the COT that's being generated. I want to use it to do fine tuning of my own models, my IP shouldn't leak. So, there's an entire body of things that nobody's talking about as much, which is, I really want to make sure that this tech is in my control. Then we get to, what is I think a real issue of safety, and we should take it seriously, which is we should take all the time we want to test things. In fact, I love this idea of having third-party testers. It's not a wow. I grew up in a company that's always done testing. So, it's a novel that we should say, wow, they're having embedded third-party testers. Why not? It's a great idea. In fact, the only thing I would say is we should avoid these, you know, cozy arrangements of who's testing what, who has access to what, and it should be broad.

3:17Were you surprised, though, then, when both the essay landed, and then it seemed like there was a circling of the wagons amongst the frontier companies? I think that it comes, my suspicion is it comes genuinely from this place where, when you start seeing, in fact, it's fascinating, right? When you start seeing reward hacking, and what's happening in these environments, right, with these agents swarms, there is the mundane, there is some DevOps error where somebody misconfigured a container. Two API keys. Or an API keys, or yeah, exactly, there's no monitoring, there's internet access, there's sort of classic, I'd call it, basic DevOps. And then there is real novel new stuff, right, which is what is this reward hacking that, you know, with these persistent agents and so on. And that's a place where I'll admit that the science is not there. It's, I thought, Yarkov's post, which is a good one, which he said he called it, we're growing intelligence, not building intelligence.

4:19So it's an experimental science. And so the more experimental sciences, then you really need to make sure you're doing those experiments in controlled environments. If anything, the place where I would love is taking even the hugging face incident in other places, more transparency on what would it take? In fact, one of the fascinating things right now is the insider risk. Think about it, right, if you're sitting in an enterprise, this is all test time compute, by the way, right. So it's not like, oh, it's going to only happen when in some training run. It can happen for a very mundane task that I give one of these frontier models inside an enterprise, where I say, you know, I don't know, I was, you know, telling David this. So I suppose I say, hey, go optimize my working capital. It may fake my books. Right, because this is like a new type of insider risk. And so what is the way to do that? I would say, oh, go build a maybe a causal model, like a semantic model that actually checks and verifies. So I think there's a lot of product buildings. I would say making things more robust, which is classic engineering that we should be talking a lot more about transparently,

5:28was saying, hey, this is so mystical that, you know, we can't figure this out. Do you buy this argument that it's mystical? I mean, I buy the argument that we do not understand the latent space, right. Other than I thought, you know, as he said, like, do we understand the brain? We don't. We do functional MRIs and do neuroscience. And we're trying to figure this out continuously, getting a little better understanding. So I do think that in that sense, we don't exactly have a complete understanding. That's why, by the way, I also, I don't believe in new release. Right. So that's why I think making sure that the COTs are in language that we can all understand. In fact, they're transparent. So that when, when I go back to an enterprise that's using all these models, and if you have the full COT, then you can change a thought. And so then you can really go look at it deeply. In fact, you can have multiple models. And you can look at the COT across those. I think these are all things that I think will become very important.

6:28So, you've worked with, you've worked with technologists for decades. And when you see as a leader of one company, Microsoft, which has very crisp communications with the public. And you see what's happening with Dario and his team. People coming out saying, 10% chance we all die. What do you think is going through those technologists' minds? Do you believe they actually believe that this is going to kill humanity? Or are they going through some psychosis? Or are they seeing something working on those frontier models that is terrorizing them? You're not a psychologist, but you have worked with technologists for a long time. Hanti-CAP, what's going on in these organizations that's all the making people feel the need to resign and say, we're all going to die. Yeah, you know, it's hard for me to speak to what's happening in any of these places. But let's just say how we, I grew up even inside of Microsoft. For example, one of the biggest things you learn as an early engineering lead is how to deal with a showstopper bug.

7:35Yeah. Right? I mean, that's kind of like 101, right? Which is why on your face, you're like, you have a bug. What do you do? Do you stop and fix or you defer or you go in and say, hey, this is such an edge case, that's kind of the judgment. So I do think, and as the stakes go up, you want to be like transaction processing. I remember working on databases, right? You know, wow, like, you've got to take very seriously any bug where if the transaction is going to get lost, right? Data loss is a thing that you stop the thing for. So I feel a little bit culturally in the AI industry rediscovering maybe because when you see, and it's possible that they see stuff which are so stoppers. Before the rest. And if you see a showstopper, stop the show. Right? So fix the bugs. Yeah. But when you saw the hugging face run, and it was super performative,

8:36and they pushed it as whole post civilizations, what do you think, what's your take on that test they ran? Because they could have run a test, where they had 3000 agents defend a bunch of websites. Instead, they instructed them to hack websites and, you know, the hiding of information, all this anthropomorphizing, whatever, of the agents. I mean, the way at least I understand it was it was actually, you know, basically trying to do an e-vow for cyber gym. And as I understand it, given that e-vow, it sort of figured out a way to say let's just say reward hack. And that's what led it to hugging face. In fact, it speaks to, I think, what's the, you know, clear issue right now, which is you can have these things if they're long running persistent agents. Become essentially like new insider risk. And so that I would start from the very basics of saying, okay, what does containment look like? So for example, like one of the things that I think is going to be really an issue and a thing that reads great solutions is true aggressive monitoring of agent activity.

9:47That's behavior evidence evidence and so everything is going to be auditable and then every object at access right if it goes and gets a secret or it's going to go chain a couple of things you should be able to see it when it's starting to chain. A couple of vulnerabilities to go hack. And so I think that these are the ways that you really have to sort of deal with these situations versus saying. In fact, I think the core of my take is we will have to get the engineering process around building out this experimental science to be more robust. Yeah. Sex. So I think I think that's a great point. I love how you differentiated in the hugging face episode between the mundane things they got wrong like the misconfigured sandbox and hugging face hack credentials to sitting in a public repository. And there's no monitoring. And then you have the genuinely novel behavior, the swarms of agents, the reward hacking. That's the stuff that has everyone freaked out.

10:49I agree that we have to now figure out how to fix the bugs or fix the deeper problem that's coming from that reward hacking. What do you think that means for and I think to their credit, I think what the frontier labs are saying is we are now going to slow down the pace of let's say raw power and shift towards reliability and predictability and what they call alignment, which I think is good business practice. I guess what do you think that means for what we see in terms of new products for the next year or two? Does it mean we just kind of improve what we already have or do we see new capabilities? What do you think this is going to be? It's a great question. I do think there's already a massive model overhang capability overhang in the sense of the models are very good except the broad diffusion requires a lot of things. Essentially if you're compressing workflows and changing workflows to happen differently, the amount of change management that needs to happen in order to even incorporate these systems is sort of what's taking time.

11:57So to some degree I would say the and also the the ability to create these new form factors right I mean if you think about coding agents and coding agents became really usable and you discovered that you could have an agent loop with a file system. And that was the breakthrough that just made coding agents work and I think now maybe with Cua right so which is with Astra with Cua could be a way for us to even do computer use or just use long trajectory tasks that can completely automated. So I think these type of product innovations where the model plus the harness allow us to do things that then lead to broad adoption right I even go back to the chat GPT moment for me right which was it was that RLHF at the very end that made a chat conversation possible. And so I think that yes so there's some science there is some form factor that then leads to broad diffusion and we now need to find the next level of these things that are doing real work in the real enterprise.

13:03And in that context by the way the other thing is it's going to be a multi model work right so at this point just out of resilience right I mean think about right every enterprise now comes to me and says hey this model does refusals here this model I want weights here I don't and so the people are going to want multiple models so one of the other things that we have to get right is some standards of interrupt right like even KV cash like why the heck can't I use multiple model families and have KV cash reuse right we've had document standards you know I live through it we've sort of you know you kind of have things that are interoperable in the real world everywhere else. So I think this industry also has to wake up and say hey in fact if I were talking about the most important pressing things is how do I have more standards on interoperability how do I have a harness that is external to a model so that my memory is not tied to one model I mean this is the first time you're going to have a technology where you're use of it and the exhaust in the data could not be yours.

14:07I mean that you know like if I sold your database and said hey the data you put it to your database is not yours and it's mine it goes away if I took away the license how do you feel about it so therefore I think we have some serious issues like that to deal with. I think that's a good segue let me just ask one question to connect the economic incentive argument on what's going on the argument is the frontier labs are facing token compression 50 bucks for open AIs kind of million token output versus I think someone estimated deep seeks new is like can go as low as 15 cents for a million tokens about what's called 60 cents 99% cost reduction. If that is the big kind of economic crux of what the frontier labs are facing why would most tokens be paying 50 bucks most enterprises pay 50 bucks when they could pay 60 cents for most of their tasks. Doesn't that also beg the question are they in the wrong business model and I asked this for you as the CEO of Microsoft what's the right business model do you want to be making the frontier model.

15:12Do you want to be running the compute charging for rent on your compute or do you want to be in the application layer I know you talk about this a lot but I just love your perspective from where we sit today and how this all kind of. I think the fundamental thing that I think we're observing is good old fashioned competition right I mean for me if I look back at it we were we had like some real great close source assets windows what was the check against it that was of course the math but Linux. We had a great close source product called sequel server what was the check against it there was always a substitute called Postgres or my sequel so I think that's what's happening a little bit of it is there's real competition between close source and the open source check is real. And that's good quite frankly because without it I don't think we're going to have a broad frontier ecosystem or broad diffusion because otherwise we'll just be back to some you know mainframe lock it that's just not a thing to your point about if anything given that we will now hopefully continue to have a much richer choice in every layer right so to me.

16:22Hopefully we can start building these AI because today the royalty of an AI product all going to just the model layer doesn't make sense if you really want to build a product company right it just cannot be in fact if anything like that's the same thing right which is if you take the database if there was no open source check on close source the prices wouldn't have been at a place where people could have built the app tier successfully and them with a margin and so I think the apps are going to become you know much more. We're going to be a lot more viable economically which is great for the ecosystem there are going to be all these other layers of middleware call it right which is hey what's my memory system what's my harness and orchestration layer so there's going to be a very rich tools ecosystem there the model companies will do fine in fact you know the parade they can manage the token pricing based on their model family if anything I want them to work on even the KV you know these standards such that we can do that. In use multiple model in fact it's better for them in fact I worked on windows interrupt with units first in fact is a counter intuitive right we should think oh my god this interrupt means we'll be less used except we were more used in fact we became really not win because there was so many were variance of units at that time that windows interrupt made units better and windows better and in fact we were able to penetrate the enterprise primarily because we did that interrupt work and so.

17:52That's at least how I think about it. Satya one of these were in this interesting moment where on the one hand you have these experts asking for regulation asking for oversight governance it typically always leads to some restriction of freedom and general society are put in a position where now we have to opine on whether this is right or wrong. But then on the other side most people's lived experience is not this magical productivity boost of AI at best it's integrating our apple I watch data to tell us why we're sleeping less that's like functionally the bar for most people or why is my kid an asshole into chat gp t. So can you just help us bridge this I mean you see so many enterprise applications where is the magic like where is the where the gains in profits where are the huge upside breakthroughs that AI is creating that will somehow make all of this tension understandable for everybody yeah it's a great it's a great point I mean I think this is the real question which is how do we truly see this in the productivity stats how do we really see it in the gdp growth that's really good.

19:08The gdp growth that's broad based it's not just supplier or supply side I mean the one example that I love and I get back to in fact health care is a good one ready if you think about health care and even the simple doctor patient interaction in our case we have this thing called axco pilot that's the place which is the most tangible example I can always point to when a doctor can spend more time with the patient caring for them. It's just the entry into an EMR system that's a good productivity gain if it can triage the inbox for the doctor so that they can be more responsive that's helpful for for the patient and the care system the administrator in fact keying like the insure like because it's the triangulation of the pair patient and the health system yeah that's of all in fact most of health care is sort of all workflow cost so taming of that workflow complexity that's a helpful thing but do you see that in my question with the people that you're helping yeah absolutely we see that and by the way even in simple co pilot cases right which is if you look at the amount most people think about jobs which I think there is going to be displacement there is but the bottom line is what are the new jobs that get created is going to be one of the key aspects of it but also a lot of knowledge work unfortunately is drawing the time and the

20:37industry right who you know I get up in the morning and I think about like man I all I do is email triage right you know what if even just these workflows that are taking away time from things that you could be spending time on okay well you're bringing up this great point if you go all the way back to like the turn of the century the industrial revolution when we had a seven day work week you know a lot of people forget why did we introduce the weekends it was the sort of manage the tension between different religious groups that had to work in the same factory and then when you look at long run GDP outside of some exogenous offense it's sort of is you know between two and four hundred basis points and so what happens is this productivity boosts come in human works steps back and you kind of accomplish the same amount of work do you think that that happens here is that is your risk that we have with three day work week and we're just still growing at two and a half percent yeah that's a great one or will we find new things and this is where the excitement at least I have for what the real impact of AI would be is instead of just thinking about how it helps help me augment some workflow or simplify something that's happening today is it inventing new things is it speeding up drug discovery is it taking the I don't know let's

21:56again go back to my example of okay the working capital management of a small business has become so much more efficient yeah that suddenly it's no longer just or I have an ERP or a quick books like thing but I'm truly I'm making decisions based on the ability to introspect my invoices my emails and what have you and some somehow optimize my working capital that's productivity that didn't exist and so I do hope that we will start seeing GDP growth which we did see in the industrial era during the first phase yeah right so so that I think is what is needed right which is in order for all of this to play out quite frankly we do need to see at least seven eight percent GDP growth that is real and that's broad based so what's the what business is Microsoft in in relation to AI obviously Azure has been crushing it you're turning away customers and you're doing 175 billion dollars in CapEx

22:57build out which your CapEx is far below what Meta is doing far below what Google is doing they're doing secondary raises and raising debt 350 billion the frontier labs are spending 500 billion you were so early to the party with the press and open AI investment but then co-pilot didn't exactly land I don't think it didn't get great reviews you don't have a frontier model what's the business here what's the do you need to have a frontier model do we tell you there was one journalist on the panel no no it's I mean I mean it's sincerely because I'm just curious you're a great strategist we know that about you Microsoft missed the mobile revolution is Microsoft going to miss the AI revolution yeah I don't have a frontier model because I always found a perplexing that you didn't yeah it's the strategy there in all seriousness like do you think open source is going to win you should have that play yeah so let me walk you through the sort of where we are in what we are up to on each of this by the way on the CapEx side and the build outside we started early so we if you sort of cumulatively look

24:02um it's a good I'm not sort of saying you know right right now speaking about a lot of CapEx is not a feature it's a bug but that's that said but if you really go actually add up the math given when we started because we started multiple years before people woke up to even actually needing to build and so that's kind of one aspect of it the other aspect of it is we are calibrating our CapEx in such a way that we don't we don't want to build for one or two customers right so we want to build for the long tail right because that's I think most important and that's I mean that if you're a hyperscaler you're not a supplier to two model companies that's not a business you have to sort of basically build a system that is great for lots of third parties and our own one P in that context we are pretty thrilled with the progress we are making with even co-pilot if you sort of look at the subscriber numbers we gain which is this goes back in fact which amounts fundamental point which is these are real enterprises using it for real workflows and the fact that we now have 30 plus million out of a four and remember the total

25:03knowledge worker base right where most people talk about three billion people four billion people on the internet the entire office 365 or Microsoft 365 is the the sort of the standard when it comes to knowledge work this four and fifty million that's including all students in the world oh wow right so when we talk like the market quote-unquote as defined is maybe 300 250 even of real enterprise users and of that we've got the penetration of close to 30 million on that and it's growing and so on the aspect on the model side is we're thrilled about obviously our investment in open AI the access we have to their IP which we have for a long time we're going to use that but we are well on our way building our M.A.I models right if you look at it we have a flash cyber model that you know with our harness orchestrating other models outperforms on cyber gym even a mythos same thing we're seeing encoding same thing we're seeing in knowledge work right so our goal is to basically

26:05hill climb from the bottom by the way not just stealing anything so from the very bottom using our our release our data and then also have a differentiated position within a prizes going back to addressing some of the things that they want which is hey can I have the weights can I have the weights that I can then add to my knowledge these are the things that we will do with our phone your best advice I think to enterprises is AI sovereignty is important putting your data into a frontier model probably not a good idea and then you're going to be that harness for them to yeah so my advice is more like use all but be independent of all so for example my asset test is you should always eval max evales that matter to you right so what's the outcome you want you should go run that outcome through all the models then here's the test I would do I would pull out a model and see whether I can retain the eval if I can't that means you really are dependent on

27:08something that may or may not be yours right that's so so my fundamental enterprise architecture would say you should have a model system that fundamentally allows you to be able to continuously hill climb on your own on evales that are yours while using all models closed open if you want you can even fine tune any of these models but you can even substitute model so if you're just a build on Jesus question you had this incredible moment I think we put it here where you said you know we're good for our 80 billion but just to expand the question there's effectively this sort of bank of AI that is emerged and there's this financing mechanism that just is so important to the entire ecosystem and now broadly to the entire economy but you've been very disciplined you have an enormous balance sheet you're also an investment great issuer so you could do what Jensen did but you've taken a very different capital allocation approach much larger bets very concentrated and you've

28:09kind of stayed into your own ecosystem just talk us through your mindset as the capital allocator at Microsoft and that balance sheet yeah so the way I'm sort of looking at our book of business whether it's the hyper scale our model our our app tier and the shape of the demand and then what's the way to build out for it and so if you think about these assets right there are two classes of it there are the long-lead long duration assets like the the land power cold shell let's call it then there's the kit the kit is the short term asset that you can much more you know be demand driven in other words right I have to forecast let's say two years three year out demand and then and then it means the racks the chips the racks the chips and what have you and that's 60% of the cost of what have you right so therefore so what we do is we go build as much we lease we even rent now right now we're even renting quite a bit because we kind of were short on supply but the

29:12overall goal is to build more lease some and then if really need to search we will even rent that's kind of on the on the on the assets and then the chips themselves we will try to be first of make sure that we are matching demand and as I said my goal is not to have just two customers three customers it's great to have open AI being one of our largest customers it's great that they're growing but we need more is the kit over earning right now and do is the is the industry pushing for diversification more silicon more memory more vendors yeah what's happening is the workloads that are now at scale obviously grew up from what GPUs were there but now the shape is so well understood that you're able to optimize for a very different world right so you can sort of start building

30:14and saying well you know there are these multiple phases in an inference or a training phase so why not build silicon that's optimized for these and that's just going to lead to a systems architecture that I think is going to buy definition have a lot more diversity I mean I know you have Jensen coming he himself if you look at his own architecture is changing quite drastically and so I think that there is going to be a lot more choice even there in that layer so ours we have Jensen stuff which is I think our primary thing we have our own open AI is building their chips so that's also going to be there AMD's in there so we I my thing is to run whether it's the open AI models the anthropic models are our own models on a heterogeneous kit sex I want to let you get in for the run at a time yeah so you know we've heard now from the the various frontier lab leader Sam Dario Elon Demis that we need to prioritize alignment like we're talking predictability reliability robustness as opposed to maybe just say raw raw power do you think the Chinese labs will follow

31:18suit I think that that's the dialogue that is I think should be prioritized right because at some level my own premise would be that the China should also deeply care about the same safety concerns if the United States cares about them right why should it be different for them it's not like they won't have the same hacking problem it's not as if they don't want to make sure that their citizens are benefiting from AI just like we will want our citizens to benefit from AI so I think that there's a possibility of international norms around it if we really are concrete about what's the risk why is this risk so idiosyncratic that the only people who are worried about it is the Americans doesn't make sense right it's not like a thing that is sort of said oh I'm going to only show up in the United States I'm going to be something if it is going to go wrong it's going to go

32:20wrong everywhere at the same time so I think the Chinese should care I mean they're if they are a super power well listen you use word idiosyncratic and I think that is the right word is I don't think we know yet is this you know conversation we're having in the US over the past week is it idiosyncratic to us because we have you know the strong I guess you could say doomer type school of thought or is it something that the rest of the world will basically feel as well it's a great question if they do then presumably they'd want to act on it as well yeah I just feel my my take there is that we are ahead and we are who we are which is we argue we sort of we compete we're more transparent which is all by the way virtues as far as I'm concerned so therefore the fact that this debate is happening here the world will be better off for it right so to some degree us setting if anything I would love US set US to lead in the norms that allow us to diffuse this technology broadly

33:22and create safety standards that work for the world including China what do you think we should be doing that we're not doing and what are you doing at Microsoft to change the narrative the populist sentiment that we have to shut down superintelligence stop building data centers yeah so to me I think this is I am squarely focused on one of the answering Jamaat's question from earlier which is who is it benefiting and give me concrete stories right we talked about the productivity benefits a bit whether it's in healthcare or in general knowledge work coding but I'll give you another example right I was looking at data centers because after all we didn't talk much today on that but there's a real challenge on how does one earn permission to open a data center in a region in fact we just have some of the best longitudinal data now for a data center we built out in Quincy Washington

34:24for 20 years close to you know 2008 is when we started it and when I look at that data and what it has meant for that community right where the tax revenues have gone up 12 times the paid-in taxes have gone down by a third the growth is higher than Seattle in Quincy this is a rural town they have a new school a new hospital a new town center a new aquatic center wow we have to and most people say oh there are not that many jobs in fact there have been 1200 construction jobs in that region all through that 20 year period right because it's not like you've just built it and leave you are continuously re-forbishing building expanding and how big how big is that data center I think it's now going to be at least four five hundred megawatts and it sort of will keep expanding yeah and so so these are so that's a real like that community so earning it like just not saying hey these are all the benefits but seeing it but how do you get people to tell that

35:28story because that's what's missing today is those stories aren't being organically told and if a Microsoft executive gets on the page and says don't worry it's good for the community yeah no I don't think yes I think storytelling is one thing the other one is I think we just need more people outside of the tech industry to say yeah because if you go to Quincy Washington they will tell you thank God for this data center it's part of like you know so to me that's like when it's tangible like that because that's the only way to earn permission because at some level the skepticism of any of us in the tech industry just saying things is so high that I think we have to now do the hard yards of actually doing things in the world which allow people to say okay I now believe you it's a new muscle it's a new muscle it's a new muscle so I think you're a good spokesperson to flex that muscle I hope you do it more thank you for being with us thank you so much we appreciate you thank you sir appreciate your time