Microsoft CEO Satya Nadella talks A.I. on Freakonomics

 Most C.E.O.s of big technology firms are not loving every day right now. They’ve been facing all sorts of headwinds and backlash. But you can see why Satya Nadella might be the exception. He has worked at Microsoft for more than 30 years, nearly 10 as C.E.O. At the start of the personal-computer era, Bill Gates’s Microsoft was a behemoth, eager to win every competition and crush every rival. But the internet era put the company on its heels; newer firms like Google, Facebook, and Amazon were more nimble, more innovative — and maybe hungrier. Jeff Bezos of Amazon would reportedly refer to Microsoft as “a country club.”

But under Nadella, Microsoft has come roaring back. He invested heavily in what turned out to be big growth areas, like cloud computing.

Microsoft has always been in the business of acquiring other companies — more than 250 over its history — but some of the biggest acquisitions have been Nadella’s: LinkedIn, Nuance Communications and, if regulators allow, the gaming firm Activision Blizzard. And there have been many more key acquisitions, like GitHub, where computer programmers store and share their code. Once again, Microsoft is a behemoth, the second-most valuable company in the world, trailing only Apple; its stock price is up nearly 50 percent since the start of 2023.

But that’s not even the reason why Microsoft has been all over the news lately. They’re in the news because of their very splashy push into artificial intelligence, in the form of ChatGPT, the next-level chatbot created by a firm called OpenAI. Microsoft has invested $13 billion in OpenAI, for a reported 49 percent stake in the company, and they quickly integrated OpenAI’s tech into many of their products — including the Microsoft search engine Bing.

For years, Bing was thought of as something between footnote and joke, running a very distant second to Google. But suddenly, Bing with ChatGPT is on the move, and Google is trying to play catchup, with its own chatbot, called Bard. So how, exactly, did Satya Nadella turn the country club into a bleeding-edge tech firm with a valuation of more than two-and-a-half trillion dollars?

In this week's Freakonomics Radio episode, Microsoft CEO Satya Nadella talks for an hour with host Stephen Dubner about a range of topics related to A.I. search: Whether this will finally make Bing competitive with Google, safety and regulation and the "doomsday scenario," and how he uses ChatGPT to read translate poetry and summarize Heidegger.

Nadella also discusses negotiations to try to buy TikTok, his business philosophy of cooperation, potential conflicts with being both CEO and Chair of the board, and succession plans.

Here are just two of the many questions -- and responses -- Dubner asked during the interview.

DUBNER: I’d like you to walk us through Microsoft’s decision to bet big on OpenAI, the firm behind ChatGPT. There was an early investment of $1 billion, but then much, much more since then. I’ve read that you were pretty upset when the Microsoft Research Team came to you with their findings about OpenAI’s L.L.M., large language model. They said that they were blown away at how good it was and that it had surpassed Microsoft’s internal A.I. research project with a much smaller research team in much less time. Let’s start there. I’d like you to describe that meeting. Tell me if what I’ve read, first of all, is true. Were you surprised and upset with your internal A.I. development?

NADELLA: Yeah, I think that this was all very recent. This is after GPT-4 was very much there, and then that was just mostly me pushing some of our teams as to, “Hey, what did we miss? You got to learn…” You know, there are a lot of people at Microsoft who got it and did a great job of, for example, betting on OpenAI and partnering with OpenAI. And to me four years ago, that was the idea. And then as we went down that journey, I started saying, “Okay, let’s apply these models for product-building.” Models are not products. Models can be part of products. The first real product effort which we started was GitHub Copilot. And, quite frankly, the first attempts on GitHub Copilot were hard because the model was not that capable. But it is only once we got to GPT-3 when it started to learn to code that we said, “Oh wow, this emergent phenomena, the scaling effects of these transformer models are really showing promise.”

DUBNER: Google still handles about 90 percent of online global search activity. An A.I. search-enabled model is a different kind of search, plainly, than what Google has been doing. Google’s trying to catch up to you now. How do you see market share in search playing out via Bing, via ChatGPT, in the next five and ten years? And I’m curious to know how significant that might be to the Microsoft business plan overall.

NADELLA: This is a very general purpose technology, right? So beyond the specific use cases of Bing Chat or ChatGPT, what we have are reasoning engines that will be part of every product. In our case, they’re part of Bing in ChatGPT, they’re part of Microsoft 365, they’re part of Dynamics 365. And so in that context, I’m very excited about what it means for search. After all, Google, as you said, rightfully, they’re dominant in search by a country mile, and we’ve hung in there over the decade. We’ve been at it to sort of say, “Hey, look, our time will come where there will be a real inflection point in how search will change.” We welcome Bing versus Bard as competition. It’ll be like anything else, which is so dominant in terms of share and also so dominant in terms of user habit. We also know that defaults matter, and obviously Google controls the default on Android, default on iOS, default on Chrome. And so they have a great structural position. But at the same time, whenever there is a change in the game, it is all up for grabs again to some degree, and I know it’ll come down to users and user choice. We finally have a competitive angle here, and so we’re going to push it super-hard. 

 Listen and find a transcript at freakonomics.com or wherever you get podcasts. (Or read a summary of highlights at Thurrott.com)