Satya Nadella

work we do as the only work that needs to be done, then everything you're saying is true. But that has never been true in human history, right? It's not like our work has not changed, right? What I do today is different than what I did even 10 years ago. So you have to, and it's, you have to then exponentially think about what is the work. Let's just take, you know, our GitHub Copilot example. I started by saying code completion is great, AI is fantastic.

Then I said, okay, let me use it as a search engine. Right in the flow, helps me without getting out. Then the agent mode came in. Great, I can now go have it do multi-file edits for me, give it little tasks. Now with the coding agent, I can give it a full issue and it goes off. In its own branch, it'll do a full PR and then it'll put back. And then, by the way, I can even have another agent, which is a code review agent, review it and what have you. So that's kind of where we are.

Nu kun je zeggen, oké, wat is de rol van een software engineer? Kijk eens wat, een software engineer is nog steeds, zoals dit is waar mijn idee is, elke software engineer is een software architect nu. De volledige metacognitie van de repo is nog steeds heel belangrijk, zelfs in de level 4 stage, laten we het noemen.

Because I need humans, like by the way, one of the fundamental great limiters in all of this will be all of the legal and liability laws. Because I think humans are going to be pretty liable and human organizations are going to be liable for whatever is the output of AI. And that's only going to be possible if there is enough human agency to verify, to attest. And what are those things that you will be doing? Those will be the jobs of the future, which is I'm attesting stuff.

I think about the opportunity for a kid to be able to now have no fear in learning anything. I mean, if you think about it, right? I always think about it. Like I'm a visual learner. I'm actually an electrical engineer in my undergrad. And I never understood the Maxwell's equations.

Until much later in life somebody showed me a JavaScript rendering of Maxwell's equations and it clicked to me. And I wish I had clicked when I was doing my undergrad. So think about a 5-year-old, an 8-year-old learning physics, chemistry, the conceptual understanding, the ability to go to something like Copilot and ask it the questions. Have an empathetic personalized tutor on the other end. So I feel that... This is your rational optimism again.

Ja, nee, dus dan kom ik terug naar het. Aan de andere kant van het, er zullen altijd de onbedoelde consequenties zijn, toch? Want als je denkt dat dezezelfde technologie wordt gebruikt door iemand, door een cybercriminal, om een cybersecurity risico te introduceren.

Some bioterrorism risks. So we are going to have to deal as a society with the unintended consequences with very clear eyes. But I don't want the fear to get in the way of what could be real benefits in our own daily lives.

Do you feel there is a necessity for fear in this moment even a little bit to make sure that people are aware? I think we should not be just optimists or pessimists. We should just meet the reality and really make sure we deal with it. With both the advantages and the benefits.

TV Gelderland 2021

Another one would be a societal check on what is the social permission to use AI, right? For example, at Microsoft, I think about it as, hey, we've got to earn it by showing that the benefits far outweigh any of the unintended consequences. Otherwise, absolutely the world should stop this. Why would they? Like, why would we? So I have more confidence, I would say, in our...

political and social systems, that if something is not working, it will not work. So therefore, my confidence is within the confines of that, how do we make sure that we all participate? Thank you so much for your time. Oh, thank you so much.

Yeah, I mean, the thing that...

It's exciting for me, Dwarkesh.

First of all, it's fantastic to be on your podcast.

I'm a big listener and it's just fun to be... And I love the way that you do these interviews and the broad topics that you explore.

It sort of, to me, reminds me a little bit of my, I'd say, first few years even in the tech industry starting in the 90s, where...

there was like real debate about whether it's going to be RISC or CISC or, hey, are we really going to be able to build servers using even x86 or, you know, when I joined Microsoft, that was the, you know, in the beginning of what was Windows NT.

So everything from the core silicon platform to the operating system, to the app tier, that full stack,

I mean, the entire thing is being litigated.