🚀 Add to Chrome – It’s Free - YouTube Summarizer
Category: AI and Technology Leadership
Tags: AI StrategyLeadership AdviceMicrosoftQuantum ComputingWorkforce Transformation
Entities: AIGitHub CopilotLinkedInMajoranaMicrosoftPalantirSatya NadellaSQLWorld BankYC
00:00
What are the tools that we can put in the hands of people that will give them that sense of empowerment? That's what I would love to work on.
I'm not into this anthropomorphizing AI at all. I come at it as it's a tool.
There
00:15
is going to be a job called a software engineer. It's going to be different.
When I look at it, you are really taking a software engineer and saying you're now a software architect. It's my pleasure to welcome the chairman
00:31
and CEO of Microsoft, Satcha Nadella. [Applause] [Music] Thank you.
[Music]
00:49
This is the home crowd. All right, man.
San Francisco, you should move to Seattle. I started my career in Seattle.
Very fantastic place. Anyone who's successful starts at Microsoft.
That's right.
01:06
So Satcha, you've uh emphasized before that AI is uh going to shape all that we do. What does this look like in practice?
Um you know at Microsoft, how does this actually drive your strategy? And particularly thinking about how AI will influence ideas behind beyond, you
01:23
know, the immediate incredible product suite, you know, like the broader economy at Microsoft, I I feel we are a platform company, uh, a product company and, uh, a partner company. So I I think of those three
01:39
dimensions and I've kind of in my 35 years I've lived through uh client client server uh web internet mobile cloud this is the fourth so that's just at least how I pattern match so the first thing that I think about is the
01:58
platform opportunity uh when I sort of look at all the folks here the interesting thing is the compounding effects of all these platforms right So this AI piece the reason why I think the rate of diffusion is so fast so well um
02:14
you know and so wide is because it builds on the previous generation. I think about like if the cloud was not there uh we wouldn't have been able to build the AI supercomputers which then led to the models which then led to the products right so that
02:30
compounding effect is the interesting thing to me so that's why you always sort of take the previous platform and build the next platform and you want to be able to get that right and then you got to build the next generation products on top of it right with each
02:46
one of these platform shifts there's a workload right I mean when I first remember looking at the large scale training job I mean it's kind of a very different workload to what we built for example the cloud with right it's a data parallel synchronous workload which is
03:02
so different than let's say a Hadoop job or what have you um and so the platform itself then completely gets re you know completely relitigated and changed uh so to me that's I think the exciting thing on the platform side it's golden age of system software uh quite But frankly um
03:18
you know today if I had to think about anybody who's building at the infrastructure layer not just the hyperscalers but even the startups I think that's a tremendous opportunity. Obviously there's a tremendous opportunity in the model side and then the products on top of it.
So yeah we think of these and then ultimately
03:35
what's it for? It's for one thing and one thing alone which is to drive ultimately economic growth and GDP growth.
So if I had to ask me my benchmark for AI is is it creating surplus in the world around us one
03:52
community one country one industry one company at a time I mean it seems like the app level you know you guys have built you know sort of the defining apps at the app layer for so many decades. It feels like we're at this weird lumpy moment where um you know maybe the
04:08
models have popped up and we're sort of astonished by what's happening but then you know sort of the compute and the apps need to actually catch up and you know the hope here is actually the people in this room will be the the people who build those apps. Yeah, it's
04:23
a it's a good question, right? One of the questions is is the model like uh SQL uh or is it the SAS app itself and the model, right?
I mean, I think the place where where does the model end and where
04:39
does the product begin? Um because if you sort of say model with some scaffolding and tool calling uh in some infinite loop is the product u if that is what it is then I think that that's where it gets a little confusing but
04:55
that's like saying a bunch of SQL business logic is with SQL is what is an app. So I think it's still possible for anyone to build uh an app tier on top of a model and you have to sort of abstract
05:12
yourself and say yeah the model is just like SQL uh was to me and so I think that that I mean I always dreamt of a moment when AI/Machine learning will have a SQL moment because if you think about it right we never had a stable platform layer in the past because
05:28
everything was vertically built and integrated for the first time in this model layer. Now we have something like a SQL engine uh that we can then use to build pretty sophisticated products and these techniques also right I mean just
05:44
the inference time compute plus tool calling is giving us I think a pretty robust harness to be able to build pretty sophisticated products. Yeah, it's kind of wild how much uh it's the integration piece that is also the app layer now.
Yeah, it's uh just you know
05:59
the the models sitting on their own are incredibly smart but uh right now they feel you know sort of there's just just giant gulf between that and uh the data that really matters to you for business users. I think that's a good observation because I think at least my read of the
06:16
situation is the model is an important piece. Um the model scaffolding and all this tool calling this uh you know sort of there's a real app server that you kind of need in order to be able to build sophisticated applications.
But
06:31
the interesting thing is the feedback loop the data path inside the product that then is used in order to post train in order to be able to do the right tool uh you know selection. Um that seems to be the place where product creation uh
06:47
is all going to happen. Um AI scaling laws are continuing to hold and the demand for intelligence appears to be potentially infinite.
Uh yesterday Elon was mentioning that there will be uh you know 99 hyper intelligent beings to one
07:04
human which is kind of a a wild prognostication but seems possible given this where does the the building for the future of AI truly demand for global uh compute infrastructure you know how do you anticipate these demands evolving as models don't just become larger but more
07:21
intelligent and capable of complex multi-agent interactions yeah I mean look If you sort of really step back and say um you know first if you sort of go with the uh you know compute or intelligence is what of a log of compute
07:37
um and then you ask the question um how much energy does compute consume let's take in the United States maybe 2% today 3% tops let's say doubles it's 6%. That's like massive because then uh the
07:55
amount of extra energy that needs to be produced um in order for AI uh to use it is pretty high. I think that's why we all have to sort of keep in mind that if there's one lesson history has taught us
08:10
is that if you're going to use energy, you better have social permission to use energy. Uh so that means you've got to make sure that the output of this AI is socially useful.
Uh in other words, if we really are not creating social
08:27
surplus, economic surplus as measured by countries and communities, uh then we will we just can't consume energy. And so that to me is the bigger thing like everybody's today hot and bothered about okay, what do I do about energy production?
I think the real question in
08:43
the next 5 years is we've got to produce enough products that are creating great value which I'm very confident of right by by the way in healthcare and education in uh in productivity. So there's many many domains but that's the real challenge for us as a tech industry
08:58
is to prove unequivocally that what we have created is showing up in real stats uh that is not just an AGI or AI benchmark. I mean the hope is that um this will show up in sort of the real
09:15
things that you sort of interact with on a daily basis that 100% you know you go use you you get a mortgage loan and instead of you know three months or two months of waiting around and you don't know if you're going to get approved or you know there's just so many things
09:31
that are important parts of your life that you know get drowned in paperwork or bureaucracy that those things could potentially go away. 100%.
So I think yeah if you take even take some of the the the public services right I mean if
09:46
you take any country you know its GDP or take health care like in the United States what is it 18 19% of our cost is uh healthcare and a lot of it like everybody talks about the magical drug blah blah blah except all of the cost is in workflow
10:02
um and so if you really take something like a simple thing like discharge uh the amount like you take the back end of an EMR system with a just a an LLM and a prompt. Uh that itself is going to save
10:18
so much time and money and energy that it would sort of pay for itself. I mean, it's kind of very direct, right?
You spend an incredible amount of GDP on health care and rightly so. But uh every dollar that's spent on clerical work could have been spent towards some sort
10:34
of treatment that would have saved someone's life or or or the simple time allocation of a physician away from paperwork to the patient uh is right there to be had. What do you see as being the biggest rate limiter for AI
10:50
deployment today? See, here's the interesting thing, right?
And this audience is so young that and none of my metaphors would sort of work. But nevertheless, if you sort of were came in um in the early part of uh you know let's say you were a multinational
11:06
company preps how the heck did we do forecasts uh like a simple sales forecast? The way one would do sales forecast was you would send faxes.
People would then take those faxes and send inter office memos and
11:23
those inter office memos would be annotated and a forecast would come hopefully before the quarter end. And then suddenly people said with email and PCs and Excel they said let's print your like let me just send an Excel spreadsheet in email people enter number and you have a forecast.
So what
11:39
happened was the work the work artifact and the work flow changed. That is what needs to happen with AI.
when someone says I I I I'm going to now do my job but with whatever 99 agents that I am
11:56
directing on my behalf the workflow is not going to be constant uh right I mean you now are really going to have to change even the scope of your job is going to change so that change management is a real rate limiter right because you're now taking the means of
12:13
production in an insurance company in a financial services company in a healthcare company in a software company and saying we are going to change everything uh in the way we work in fact we're going to change what jobs they are like you know at LinkedIn I think they
12:29
took multiple of these functions uh the design function the front-end engineer function the product function put them all together and said we're going to have full stack builders uh that's a change in scope of even a job and so how do you then rebuild the product team
12:45
with new roles new scopes and what have you. That's to me more the social rate limiter.
Not that there's lots of other things around deployment of this technology, getting it out to the world, power is one, there are other issues, but I would say change management when I
13:01
look at even a lot of the AI startups when I talk to them, everyone has now you know you worked at Palunteer so you know this uh everyone has forward deployment engineers. That's like the exciting thing is the palunteer model which I think is a fantastic model and why is that?
That is because of that
13:17
change management. Uh because I think you really need to help customers partners really understand the benefits of any product you're creating but not just the technology but even how to use the technology in a workflow.
At YC, we have this uh funny saying that we tell a
13:34
lot of people here to do, which is um you know, these are some of the smartest uh AI researchers, computer scientists who are just starting out in their careers. We tell them uh go undercover.
So go work as a medical biller and see to what degree how many you know quote
13:51
unquote knowledge work jobs are actually copying pasting from a browser into a spreadsheet into an email and then clicking send and do that for a while and realize like actually these are not necessarily you know using your
14:06
prefrontal cortex and your highest mind kind of jobs like these are not you know can you imagine so many people like their lives are basically like you We used to we used to, you know, coming up uh at our age, we would call it paper pushing, but you know, they're not paper pushing anymore, but they're sending
14:23
emails. You know, they're not sending faxes anymore, but they're trying to get business done by like attaching files to things.
You know, that seems like a pretty big shift actually that they can Yeah. I mean I I mean I think one of the most understated things as an opportunity for anyone creating products
14:39
or fundamental breakthroughs um even at the model layer is the amount of drudgery there is in knowledge work right I mean I mean in software engineering we saw that I mean the amount of you know we're taking the joy out of software engineering because you know you were out of your flow uh to be
14:55
able to stay in the flow to be able to complete a task uh that itself is a great example of what I think is going to happen to all knowledge work. You're absolutely right.
The amount of cycles you spend out of band uh collecting information because if you think about the prefrontal cortex and the synthesis
15:11
part uh you know the amount of time you spend there is pretty low now like having a sophisticated reasoning model and your prefrontal cortex work together. Uh whereas a lot of the mundane stuff is getting done by even some ka agent or what have you.
That I
15:27
think is definitely the frontier. So beyond simply adopting AI tools, uh what are the biggest transformational shifts you're seeing in the field today?
I think to me um even like I mean look this field is changing so rapidly right
15:45
I had not even imagined last year even this time that we would get this far with RL um and with basically test time compute. uh and it seems sort of pretty limitless.
So the way I think about it
16:02
is pre-training worked. Uh all the post-training techniques on top of it were fantastic.
Then this inference time compute seems to have really added uh in under a massive scaling law. So now I'm
16:18
interested in whether there is some new algorithmic breakthrough. uh because I always say this entire regime could be changed by one person here who comes up and says I have a more efficient thing to do or a way to do this stuff right so that's you have to be open-minded that
16:34
the last big breakthrough algorithmically has not yet been found uh so that's one um I'm always sort of interested in that the other one is what is the next step up uh right because what is the pre-training to RL
16:52
um the end toend training loop that's the next you know big sample uh that I think is also what I think will happen in the next year uh so I would say what if that is another scaling law breakthrough because we will be like if you sort of take any lab now all of us I
17:08
think will be working on saying what's a more integrated response reasoning model that we can build u and that I think is going to be the interesting piece there's something very interesting here I think in that uh if you think about like an LLM instance as a consciousness
17:25
which I think some people are starting to say uh you know it's sort of instantiated you do a bunch of work with it and then it sort of goes away and you open a new chat box and it's you know I guess I'm curious like do you think that that might be one of the things that
17:41
needs the loop needs to be completed right that like yeah I mean so I'm not sort of I don't to me this artificial intelligence is unfortunately the worst name we could have ever picked. And so I'm not into this anthropomorphizing AI
17:56
at all. I mean I think of it more I come at it as it's a tool.
It's not trying to replicate how we think. Uh it is it it's definitely showing signs of intelligence but it's not uh intelligence that I have and I think of human agency still will matter will be there and we will sort of
18:13
use these as tools. So that's kind of my position.
That said, let's just say oh yeah, a memory system uh is a good thing. Uh these things do need if I look at the next frontier, I would say there are three things, right?
One is memory. Uh the other one is tools used.
And then
18:31
the third which I think is perhaps the most important thing is entitlements. Uh which is basically if I'm going to take action, what entitlements do I have to take action?
Right? So these three systems have to be built as first class
18:46
around the model in order for us to build more sophisticated applications. One of the uh arguments people are starting to make around the future of software is well we have the database and then you're going to have uh basically middleware that is you know I think you know what you call
19:02
entitlements you it's kind of like access control list like you know what's the business logic who gets to do what and then you know you you basically put the agent on top there um you is that sort of that's right so that that's why I think about like know people say when
19:18
you think about the scaffolding layer right you have a model plus scaffold scaffolding the scaffolding now really gets first class by thinking of these three things uh tools use is one memory is one and then entitlements and you put that stuff together then you can create an agent an agent has an ID uh agent has
19:35
management and provision control on it it has you know so that's the way I think to think about it do you worry with codeg genen like would do you think users at some point will just prefer to uh make software just in time instead of
19:51
using package software. I mean that's something that uh you know we having lots of conversations in the hallway about that because a lot of us in this room you know YC will actually fund a ton of SAS and will continue to do so but you know in the background we're starting to have that worry.
There's
20:07
some venture capitalist friends of mine who are in the audience. They're actually like I actually don't know if I can continue to fund B2B SAS.
How do you think about that? Yeah, it's a great question.
I mean you know it's it's interesting. Um at the same time uh I look at the number of people who are
20:23
forking VS code and I say man we must have done something right. Um and u so therefore there is something to be said about building a great um uh IDE um in fact when I think about Excel I
20:40
think of it as an IDE. Uh so the fact that there's a great canvas uh you can then then bring let's call it the best analyst model to this IDE uh and then create a loop between the canvas and the
20:55
uh model. Uh so I think yes you can generate applications just in time uh you could have a prefabbed application that is really helping with the feedback loop to the model and I think both of these things will exist together.
Do you
21:10
think there's a role for design in all of this? I mean, basically a h, you know, a human being sitting in front of VS Code uh is sort of like the translator between, you know, the software and what the end user really wants.
And then I think some of this idea that software goes away presupposes
21:28
that just, you know, normal people walking around are going to want to create software. And, you know, I don't know if that's going to work.
I think that's a good point. So I think the way I sort of say um because one of the the basic question you're asking is what
21:44
happens to software engineering right I mean that's the let's take the following uh thought experiment right if you sort of said some Martian intelligence came um in uh the 1980s and watched how we all worked oh wow these humans kind of
22:00
work in the offices and they have a typist pool they have a slide pool and people then work with paper And then if they came back today they'll say god man all eight billion people are typists now right I mean that's what they'll uh sort of you know surmise and so I think what
22:17
I think will happen is all of us are going to be creating software but there is going to be a job called a software engineer it's going to be different but I look at it right you you you are really taking a software engineer and saying you're now a software architect right see the I I still the
22:33
metacognition of your I mean one of my biggest things is man Wipe coding is fantastic until it does stuff that I don't know what the heck happened. Uh so that means I have to have the meta model of my repo and exactly what happened and
22:49
I'm looking at the change logs right. So when I look at the my favorite feature of GitHub now is to really look at the complete change logs of all the agents that are working on my repo.
uh and I think that is where a a lot of the software engineering will be like a good
23:04
dev manager right I don't know who which dev manager you worked at Microsoft but I really looked and and a dev manager's job was to make sure builds don't break and the code has got good quality um and so to me that is still a thing uh and so
23:19
there will be a level of abstraction uplift uh even in a world of all of AI agents because one thing that we don't talk about is the legal liability by the way until some real laws change are going to be with humans and institutions humans build. Um, and as long that is
23:36
true, we're going to have to really make sure the human is in the loop uh at a fundamental level. And that means we will need a lot of tools for humans to be in the loop in order to figure out what these things are doing.
In AI development, you see so much, you know, what do you think is underestimated and
23:53
uh, you know, and what is overhyped by the broader tech industry from where you're sitting? It's not short of overhyping.
Let me just put it that way, right? We're at the uh everything is AI all the time.
So, it's good. Uh you know, for us all in this industry, we live and die by our ability to get into
24:10
a frenzy about something new, right? What is the Steve Jobs thing or the Bob Dylan thing, which is you're either busy being born or busy dying.
It's better to be being busy being born. So, that's good.
The I think the thing that we have to most um worry about and most work on
24:27
as a as a tech community I would say is that how do we earn that social permission? If there is one thing that I feel to me one of the demos I saw which completely really blew me away was I think in the beginning of 23
24:45
um when I was in India and I saw a local developer Daisy chain essentially at that time either GPD3 or 35 uh with one of these India stack speechto text texttospech uh open-source things and
25:00
then showed a local Indian farmer who was able to sort of use a a chatbot that was built in WhatsApp uh to be able to get some agricultural subsidy, right? Uh by going to a government website.
That to me was unbelievable, right? I I I
25:17
felt like, man, how could something that was built in the west coast of the United States get to a real use case that fast uh thanks to sort of the diffusion rate and basically people uh everywhere. That is the story that needs
25:33
to be told uh right at scale. Uh that is the underhyped story I think because right now the overhyped thing is the model capability and the model capability is fantastic but man if we can somehow get the world to recognize
25:49
that this is making a real difference in the lives of people everywhere we're in good shape. If that doesn't happen, this is all about some valuations of us uh our companies and our industry and it's the same repeat then that is not going
26:04
to end well. I love that example.
Um I mean you can I don't it feels like Microsoft is sort of full of examples of things that uh lower the floor so that you know a lot more people can get access to technology. I mean you could
26:20
argue go GitHub Copilot is one of the biggest. Yeah.
By the way, one of the other ones just you brought it up. There was a World Bank study they did, I think in Nigeria and now they've taken it to Peru or Chile, one of those in in South America.
It's, you know, we've been working at Microsoft forever on can
26:36
there be an intervention in education, right? That's been the dream, man.
We've been at it at it for decade after decade and it's made a difference. uh but this study said by access to something like a co-pilot is probably the best tech
26:52
intervention in education uh in Africa or in Latin America and that's you know that's been the dream I think that we've all had in tech and it's right there within our grasp I guess are there any um interesting observations I'm curious because you know your co-pilot in
27:09
Windows is uh you know often you know here in tech like maybe people are really obsessed with the latest frontier models, but it's easy to forget like you know Windows and the integration with Windows is actually the first interaction people have with uh you know
27:26
pre- AGI sort of AI today. Are there any observations from like people using that and like Yeah.
No, we're very very excited about Clippy being back as co-pilot. Uh but seriously, I mean like look at to me um the the thing that I find
27:44
is even in the form factor that we know and love and uh work in which is a good old computer with a mouse and a keyboard, right? The dream has always been in fact the first research group Bill built at Microsoft research was speech
28:00
um in 1995. Um, and so since then we've been saying, God, like when will speech be first class on PCs.
But right now with Copilot, the two things that are just pretty surreal to me, it's kind of like a new browser moment, right? There
28:15
is both vision and speech. I leave it on all the time.
It can see what I see and I can speak to it. That seems like a precision mouse movement, right, to me, right?
So that is where I think even on
28:32
existing form factors there is a way to change uh the complete computer use and then there will be new form factors right uh so I think it's an exciting time to be building uh both hardware and modifying existing hardware for what is
28:50
I think possible in terms of computer use yeah computer use is fascinating in that I mean you have the intelligence and then computer use is actually the superset of uh all the data like your personal data, your work data, your you know all your office docs like
29:06
everything is accessible right there. Was uh the movie her correct and that you know literally the operating system is going to embed itself with uh your most trusted agent.
Yeah, I mean I think that has been the dream which is can these agents become your computers. Um
29:24
and they do the computer use for you. Uh and that absolutely I think is the direction of travel.
Uh and I think you mentioned the most operative thing which is trust. Uh which is can I trust this to delegate what I want and that means
29:41
it's about precision. uh it is about sort of uh the privacy uh it's about a lot of these considerations and I think that these all will in time will have to work out.
I mean in that respect you know when you look at um you know both
29:56
your company and you could argue Apple they sort of have to be on the front lines of protecting privacy for all computer users in the world actually. Yeah.
I mean so to us you know there are many it's not even sort of there's privacy there is security there's
30:13
sovereignty these are three big big considerations right privacy every user cares about it uh security is what every tenant or every customer will care about it on top of privacy and then every country will care about sovereignty
30:29
security and privacy so that's the way to think about it right so you really need to build any product or any system you need to be able to answer the questions on for the people and for organizations and for countries how you cross all those three boundaries. Sachi,
30:46
you've had an absolutely extraordinary journey at Microsoft starting as an engineer all the way up to CEO. What lessons from that path would you share for the next generation of builders?
It's not like um you start any journey
31:03
um with sort of a specific goal of where you want to end up but you do start uh with this uh goal of taking the first spot and sort of having the highest ambition for yourself on what you want
31:19
to get done. Right?
I always say u it's not like uh I was waiting to become CEO to do my best work. The first job I had uh I felt was the greatest job I could ever have when I joined the company in '92.
I felt like w if I retired in that
31:35
job that would be fantastic. Um and that was a great mental model when I look back at it, right?
Which is it's not I was not waiting for my next promotion to do something but using the opportunity I was given to do everything I could. Um and I think that that's what uh people
31:51
who are starting out or who are founders or who are researchers or students today have. And so I would say keep that alive.
Uh don't wait for the next big thing. You take the thing that you have as the biggest thing and then make it expansive.
Um and then the other thing
32:08
that I would say is big things are achieved uh by having a team around you. uh learn how to work in teams, making teams great.
Uh one of the things that I feel at Microsoft I learned was what it
32:24
means to be in a project, what it means to work. In fact, that's kind of the big difference between school and work is that right, which is you join a team.
Um and you got to figure out you how to make the team successful. the incentives are actually pretty clear except I think
32:42
the thing that is least thought uh is how do you really make sure you can compose as a team and what's your role in it. Every one of us sort of looks and say somebody else's job is to align the team.
It's your job to align the team. So I would say if you get those two
32:57
things high ambition for your own impact, how to work in a team and make a team effective that's magical. Uh here's a fun story.
Um, I actually did learn how to uh do product management and project management as a PM on Windows Mobile. And uh when I was employee
33:14
number 10 at Palunteer, I taught them uh actually how to run a project zero bug bounds and you know all of the sort of you know my PM training at Microsoft turned into the thing that created you know how even Palunteer uh you know runs
33:31
their product org today which is pretty wild. So, you know, thank you to Microsoft for that.
Um, I'm curious, you know, what are the qualities that you look for in uh, you know, sort of people and teams just because AI is becoming a really key piece of, you know, creative
33:47
work and engineering work. Um, it's sort of changing the way even you might interview someone and evaluate them for technical or, you know, broader skills.
Yeah. I mean, look, I I I'm always looking for uh three qualities in people.
Um, one
34:04
is, uh, in fact, you know, Bill turned me on to this, which is he was describing at one point who are good architects and who are bad architects. And he had this, you know, nice way to uh, uh, summarize it, which is good
34:20
architects uh, bring clarity and bad architects bring confusion, right? Even if they're equally smart.
So I sort of always go um to people who innately to can drop into an ambiguous uncertain
34:36
situation and bring clarity uh it's an understated quality right I mean you just think about the number of conversations you have in a day um about some tough situation tough context uh and people who can bring clarity on what
34:52
to do what to do next what's the next step that's at a premium So I always am looking for people who bring clarity in uncertain times. The second thing I'm looking for is people who create energy.
Uh right in the other words um it's like
35:08
not just they bring energy but they're also really able to bring multiple constituents. Right?
Anybody who comes to me as a leader at Microsoft who says my team is great everybody else sucks. That's not really useful.
uh I need people who can bring people together across the company outside the company
35:25
create energy right innately and then the last thing is people who are good at solving over constrained problems that's why I think my favorite interview question always is um asking someone to describe like a project they worked on
35:42
which really was going nowhere and they figured out a path right and the way they go about it uh problems solving uh because essentially what are what do people do who are successful? They take an over constrained problem and figure out how to unconstrain it.
Um and that
36:01
magical sort of three things right which is bringing clarity, creating energy and driving success by solving over constrained problems is what I think leadership is about. But leadership is not about something that you do later in life.
You do it every step of the way.
36:16
Uh I want to cover quantum briefly. I mean you guys just released your um Majorana 1 in February.
Um is there an interaction um with the future of AI and you know I think there are probably some quantum researchers in the house. So curious what the future will be.
Yeah to
36:32
me it's pretty exciting to see what's happening. I mean we've been at it man for like it's like I'm the third CEO at Microsoft who's been writing checks on quantum.
Um um and we've been at it for 20 plus years and the dream at least or the
36:51
focus we always had was if we really want to build a quantum quant computer which is a general purpose computer uh you got to solve u for really stable cubits um and error corrected cubits um so a fault tolerant quantum computer and
37:08
we bet on this bas basically a physical property which was envisioned by these Italian physicists Mayorana and that is what uh we went after and finally we've had a physics breakthrough and we were able to actually fabricate that particle and so
37:24
therefore uh that's what has led to this chip. uh so we feel like one of the big things that we needed to achieve uh has been achieved and the way I think about it is if you say um takes you know if you want to understand the language of nature which is simulation
37:40
um I think the best way to do it is through a quantum computer because after all uh you know physics and nature is quantum and so therefore but AI is I think of it as an emulator of that simulator right so that's another way to
37:57
uh perhaps even use AI today with HPC. In fact, a lot of what we are seeing is pretty good advances in uh using um basically HPC plus AI as a way to accelerate advances in chemistry, in
38:13
physics, in material science. And so quantum would be the next step in it.
But we're very excited about what AI plus quantum and HPC in a loop can do. Very cool.
Uh we're running out of time. I feel like we could go for another hour if we if we had the time.
Um, so just to
38:29
close, I just wanted to get your sense. You know, let's do a simulation of a sort.
You know, you're uh 22 years old and you're level 59 at Microsoft. You're starting your career.
You just graduated. Um, what are you working on given, you know, in 2025?
You know, if
38:46
you started over knowing what you know now, what would you be working on? How would you be approaching it?
you know what would you be excited about? If you look back at the history of Microsoft, how office got built is a you know it's an unbelievable uh story in the sense of
39:04
thinking of these tools um right a word processor a spreadsheet a slidem tool um what those tools are meant to all of us right I mean I mean ex I mean that's why I always say what's your if somebody
39:19
asks me what's my favorite product it's always you know VS code is one and the other one is excel out. Uh it's just your you feel so good when you use the tool.
It's all about the sense of empowerment you have, the numbers sense
39:35
you have, the analytical power you have with something simple like a spreadsheet like what an unbelievable scaffolding uh it is right uh columns and rows with some sort of tuning machine in the middle um is just breakthrough. And so I
39:51
would want to work on what are the next set of tools like when I see even copilot today that's kind of where I feel like you know researcher analyst creator these are like the word excel powerpoint right every day I go to them so to me that's what I would love to
40:07
what are the tools that we can put in the hands of people that will give them that sense of empowerment that's what I would love to work on I have a feeling the people who make those tools are sitting in in this audience right now. Please give it up for Sachin Nadella.
40:23
Thank you so much. Thank you.
Thank you. Incredible.
[Music]