STEVE INSKEEP, BYLINE: We have a portrait of one of the brilliant minds whose ambition is to create an even more brilliant mind. Demis Hassabis is a leader in artificial intelligence. He co-founded DeepMind, the company that's now part of Google. In a corporate podcast late last year, the British researcher recalled the years since he started the company in 2010.
(SOUNDBITE OF ARCHIVED RECORDING)
DEMIS HASSABIS: When we started DeepMind, no one believed in it. No one thought it was possible. People were wondering, what's AI for anyway? And then now - fast-forward 10, 15 years, and now, obviously, it seems to be the only thing people talk about in business.
INSKEEP: Hassabis himself is part of the reason. He shared the Nobel Prize in Chemistry for AI-driven research in biology. The journalist Sebastian Mallaby has written a new biography of Hassabis. It's called "The Infinity Machine." The book describes Hassabis as a competitor determined to win in both science and business.
SEBASTIAN MALLABY: And that combination is pretty rare. If you look at Sam Altman, he's a great entrepreneur, but not a scientist. If you look at someone like Geoffrey Hinton of the University of Toronto, he is a great scientist, but not an entrepreneur. Demis combines those things. And he is the one who had the vision to set up an AI lab back in 2010, before AI could even recognize a cat. So there's a sort of fascination with somebody. How do they develop conviction quite that early?
INSKEEP: Where did that vision come from?
MALLABY: Well, it came, amazingly enough, from his teen years. He was a chess player - you know, the best young chess player in Britain, second-best in the world for his age group. But then he got bored of chess and sort of segued into coding. And he coded up an amazing early video game, and his boss at the time kept on talking about AI. And he got this idea that he could only advance science - and he had these enormous ambitions to be the new Einstein - that he could only do this if he invented a new scientific tool. After all, telescopes have helped science in the past. He thought AI would be an even more powerful tool. So right when he was 17, 18 back in the mid-'90s, he was already fixated on building artificial intelligence.
INSKEEP: What extraordinary ambition - I want to be the next Einstein.
MALLABY: Right. Exactly. I mean, he said to me in an early conversation, well, you know, people like Newton ultimately failed.
(LAUGHTER)
MALLABY: And what he meant was, you know, he'd explained some of reality, but not all of it, and that wasn't good enough. And, you know, one of the mysteries here was, why would an individual be quite that ambitious about understanding the deep mysteries of science? Because this is what really motivates Demis. And the answer, I think, is that for him, it's a quasi-spiritual quest. He would use religious language in terms of his feeling of compulsion to understand nature more deeply because that would be getting closer to his version of God.
INSKEEP: But there's even something grander than that in this ambition because you are effectively not just saying, I want to have my own God. I want to be the creator of God. I want to be in charge of God unless it gets out of my control.
MALLABY: Yes, and Demis is highly competitive. I would say he is the most competitive person I've ever met, period. And the competition extends to kind of silly things. Like, you know, he once said to me, when he was in college, he was the best table football - foosball - player in the entirety of Cambridge University. So I laughed politely. And he said, no, no, I'm serious. I'm serious. I've watched the best players on YouTube, and they don't have the snake shot, which you - which I know how to do.
And so even on something as silly as that, he took it incredibly seriously. But when it comes to building AI, it's obviously 10 times more serious, and he is determined to be the one to bring it into the world. When I met him right after ChatGPT was released by the rival company OpenAI in November 2022, he told me, Sebastian, the opposition has parked their tanks in our front yard. And that's how he viewed it.
INSKEEP: I would say that at this point, OpenAI is still more famous than DeepMind, if that is a measure of anything. Does he still feel like he's behind?
MALLABY: No. I think that although - OpenAI has probably more users, although it depends how you count it, because DeepMind's AI models reach Google users in enormous numbers. So you don't see the DeepMind brand when you used AI mode in Google Search, but that is what you're using.
INSKEEP: I want to ask about some of the questions that are raised about his work, which you certainly do in the book. You mentioned Robert Oppenheimer and the atomic bomb within a few pages, for example. We had recently on this program Tristan Harris. He's a former tech worker who is now at the Center for Humane Technology. And let's listen to one of his warnings about the concentration of corporate power through AI.
(SOUNDBITE OF ARCHIVED NPR CONTENT)
TRISTAN HARRIS: The only way they can justify the amount of money that's been invested into this AI industry is if they race to replace all economic labor in the economy. They're not racing to augment and support human workers. They're racing to replace human workers. And what that will lead to is unprecedented levels of concentration and wealth and power because essentially, all the money in the economy - instead of being paid to individual laborers, you're going to pay, you know, five to 10 AI companies to do all of the work.
INSKEEP: And Hassabis is there working for what would be one of those five to 10 AI companies. What does he think about a handful of firms, including his own, absorbing more and more of the wealth and power in society?
MALLABY: He would be unhappy with that. He is unhappy with that. I mean, he has pretty strong social democratic values. One of the reasons why he has refused to relocate to Mountain View and to Google headquarters is that he kind of views Britain as a more egalitarian place. You know, you go to the DeepMind office in London, and there are kids from the local housing projects playing soccer on the public grounds. You know, and you wouldn't find that in Silicon Valley, where things are all private. So I think he deeply would like AI to be used in a more egalitarian way. But the question now is, you know, does what he prefers, you know, make a difference? I mean, if Google is effectively his overlord, you know, does his own preference matter?
INSKEEP: Is he like Robert Oppenheimer, then, in that he's going to do the science, and ultimately, it'll be up to somebody else to do the arms control?
MALLABY: Yes, exactly. That's just what it is. And the question is, will they do the arms control? In the case of nuclear technology, it took until 1968 for the Nuclear Non-Proliferation Treaty to be agreed. And I just hope that it won't be quite such a long lag this time.
INSKEEP: Sebastian Mallaby, it's a pleasure talking with you. Thank you so much.
MALLABY: You're welcome.
INSKEEP: His book is called "The Infinity Machine: Demis Hassabis, DeepMind, And The Quest For Superintelligence."
(SOUNDBITE OF LUDWIG GORANSSON'S "QUANTUM MECHANICS") Transcript provided by NPR, Copyright NPR.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.
