{"url": "https://www.youtube.com/watch?v=muwZu5iG_Ww", "page_url": "https://www.youtube.com/watch?v=muwZu5iG_Ww", "title": "Sir Demis Hassabis in conversation with Dr Mohamed A. El-Erian - Fireside Chat at Queens' College", "show_or_channel": "Queens' College Cambridge", "date": null, "duration_s": 1522, "type": "interview", "speakers": ["Demis Hassabis", "Mohamed A. El-Erian"], "confidence": "high", "notes": "Fireside chat at Queens' College, March 2025. Approximate date: 2025-03.", "people": ["demis-hassabis"], "id": "0000_queens-college-cambridge-sir-demis-hassabis-in-conversation-with", "audio_duration_s": 1521.127625, "worker": "jc-w23", "transcribe_s": 69.7, "model": "large-v3", "segments": [{"start": 7.25, "end": 36.24, "text": "Thank you, everybody, for coming. We are extremely honored and delighted to have with us Sir Demis Hassabis, fellow of the Royal Society, the most respected and admired leader in AI, and one of the most highly regarded scientists in the world. But much more importantly, he's an alumnus of Queen's, an honorary fellow, and a fellow benefactor of this college."}, {"start": 36.24, "end": 73.65, "text": "This college first Nobel laureate. We are extremely proud of you. So I could spend the next hour telling you about Demis and simply going through his CV, but I'm not going to do this. Thankfully. Instead, through questions about him, we're going to talk about Demis, the chess protege, Demis, the founder of DeepMind, Demis' work on AI and solving the longstanding protein folding problem."}, {"start": 73.68, "end": 94.35, "text": "And then the Nobel Prize and what's ahead. And this way we're trying to complement the incredible lecture you gave, which Alice introduced, that I see quite a few people were here, so I had to rewrite my notes quite a bit. Okay, we're then gonna open it up to your questions. So we are gonna continue for about 35 minutes and then open it up to your questions."}, {"start": 94.45, "end": 121.79, "text": "Well, thanks for, I mean, it's amazing to be here and thanks for hosting this, Mohammed. It's a real honor to be here and thanks everyone for coming. It's amazing to see so many old friends and supporters and mentors and also new friends and colleagues. It's going to be great to chat to you after this at dinner. And it's amazing to have this in this amazing venue. You're the only one who could bring in the vice-chancellor."}, {"start": 121.79, "end": 152.3, "text": "Reid Hoffman, Stephen Fry. I mean, I can go on. That's very distinguished. And most importantly, Lorie Twel. Lorie, well, of course, yes. And Dr. Walker. And Dr. Walker. We're going to come to Dr. Walker. I thought you were going to say Dr. Walker. So you've said from the very start that your quest in life is to solve AGI. People have an image of what tech is. Tech is in Silicon Valley. It's not in London."}, {"start": 153.26, "end": 183.07, "text": "Tech, when they're offered a million pounds not to go to university, take the million pounds and don't go to university. When Google comes along to merge, rather than think about what computing power that brings to them, they'll think, no, I'm going to stay outside and maximize the value of my startup. Tech goes fast and breaks things. You've done none of these."}, {"start": 183.07, "end": 209.38, "text": "So explain to us a little bit how you've gone against every single conventional wisdom in your world. Well, look, I mean, there's lots of aspects to that. I mean, the first fundamental thing is I've always dreamed of doing this from the beginning. Maybe if those of you in the lecture can see, I've been thinking about this for more than 30 years. And I would have done this no matter what."}, {"start": 209.38, "end": 232.1, "text": "So work on AI, try to advance the frontiers of science. I always dreamed of, and places like this, Cambridge inspired me to dream of maybe doing something one day worthy of the greats that I read about, you know, and idolized in some ways when I was growing up, you know, the Trurings and the Babbage's and the Cricks and so on."}, {"start": 232.1, "end": 252.69, "text": "And which this place represents. And I think all of you who came to Cambridge and are involved with Cambridge feel that. It's in the bones of the buildings, I feel, and the grounds. I really feel that somehow the intellectual atmosphere over the thousand years that Cambridge has been here really matters."}, {"start": 252.69, "end": 265.25, "text": "And so it was inspiring. That's why it was easy to turn down the million pounds because I was not going to, you know, once I saw that Crick and Watson film with Jeff Goldblum in it, I was like, that's where I'm going. How many people told you that was crazy?"}, {"start": 265.25, "end": 285.36, "text": "Well, I mean, lots of people and today, like, you know, with the way that Silicon Valley is today, it's like people that, you know, it's a badge of honor to drop out of universities or not go at all. I actually totally disagree with that, especially if you want to do deep, what's now called deep tech startups. So startups where you're, it's not just the latest, you know."}, {"start": 285.36, "end": 314.3, "text": "Photo app or gaming app or mobile app. You actually want to solve something important. I think startups are actually one of the best ways of doing that, but you also need the training. You need to understand what research really is. I think PhD is vital in that, actually understanding how to deal with the research discourse and to engage with that and learn from it. You can overturn it, that's fine, but you also need to actually understand how the current systems work."}, {"start": 314.3, "end": 329.34, "text": "So that was easy for me. And money, I've never done, money's like the least important thing for me. The only important thing about money is it gives you free time and resources to do the things that you want to do. So it frees you up."}, {"start": 329.34, "end": 357.36, "text": "Clearly. And that's why I couldn't even come to Cambridge if I hadn't earned the money doing theme park so that I could come here. And in those days, fortunately, the university was free and you actually got scholarships. I mean, you know, today, I think it's pretty bad that the amount people have to pay to get this experience, which is why we try to help with our personal philanthropy in scholarships and things for underrepresented, disadvantaged kids, because I think it's so important for them to get that opportunity that I had and others like me had."}, {"start": 357.36, "end": 370.22, "text": "So I've always wanted to do that and the reason is I've always been obsessed and maybe troubled by the biggest questions and why more people aren't."}, {"start": 370.22, "end": 396.78, "text": "Worrying about them all the time. I used to read all these, maybe I read too much science fiction and read too many biographies of these scientific greats, but I could never wonder. We don't know what time is. We don't know what consciousness is. We don't know how quantum mechanics works. How can everyone go about their daily lives when we don't know any of these super basic things? We literally don't know what time is. That's just incredible. How can we live with that in that situation? That's how I feel."}, {"start": 396.78, "end": 409.81, "text": "This burning desire to have to know. Maybe I'm sort of incurable, inquisitive about just how does the world work? And when you follow that line, then you get to AI as..."}, {"start": 409.81, "end": 435.28, "text": "In my view a natural answer to that question of like I need to build there's so many great questions to that I want to know the answer to then maybe if you did that just specialize in physics or psychology you get to work on one of those questions but they're all fascinating time consciousness all of these things so my answer to that was to build a general tool that in itself was a very interesting intellectual pursuit but then had this double thing of also being a great tool"}, {"start": 435.28, "end": 462.98, "text": "For all the other questions that I wanted to answer. This takes us to Queens. So you come here younger than your peers, having worked at Bullfrog, right? Yes. For a year. Yeah. And then you're around all these freshers. Yeah. What was it like? Well, look, it was actually the most amazing three years of my life, still now. I've been working, you've got to understand, I've effectively been working since the age of four."}, {"start": 462.98, "end": 478.96, "text": "Right, because I've never had a summer holiday. So ever since I was a kid, every summer holiday from the moment school broke up to the start of the next year, I'd be playing chess tournaments around the world, which is pretty intense and kind of, you know, we were sort of doing that professionally."}, {"start": 478.96, "end": 508.83, "text": "And then obviously in my year out and year and a half I did, instead of doing, I sort of slightly regret this, sort of traveling around South America or something, which is what I would have liked. I wish I'd done that as well somehow. I went straight, the day after my A-level results, I went straight to Bullfrog to work all the way to the first day at Cambridge. So when I got here, I was really ready for a break and to sort of really just embrace everything Cambridge had to offer, not just intellectually, but socially."}, {"start": 508.83, "end": 535.46, "text": "And I do think that's why Queen's specifically is amazing. And I was very lucky to get chosen by Dr. Walker to come to Queen's. And we'll tell that story later. But because Queen's is quite unusual in that all three years, at least at the time, I mean, I think it's still true now, but all three years where you were in college. So you really got to know everyone in the year. And it was a very family atmosphere. And we had an amazing year. And still my best friends now are..."}, {"start": 535.46, "end": 559.6, "text": "People, actually they were in the front row of the lecture today, are people I met here at Cambridge. Not only my best friends, lifelong friends, but also some of my most important colleagues, including like David Silver and others who are on AlphaGo, who I met here at Cambridge on the computer science course. So both professionally and socially, it was an amazing time for me here."}, {"start": 559.6, "end": 588.42, "text": "I was ready to embrace everything that Cambridge had to offer and sort of throw myself into that fully and not worry too much about this mission of AI. I sort of picked that up again at the third year, but for a couple of years I was able to be just a normal student and just absorb everything. You mentioned Robin Walker, Andy Koch, my director of studies when I first came here. I was terrified of him. Absolutely terrified of him."}, {"start": 588.42, "end": 615.44, "text": "How did you feel about it, Robin? No, it was inspiring. I mean, well, Dr. Walker, he used to give us incredibly hard maths problems from the, I think they were from the 1960s or something, Dr. Walker, right, just to make them extra hard. And then it made the exam super easy because, you know, they just seemed way harder back in the 70s. And it was inspiring. You know, Dr. Walker is just, the sky's the limit. You can try anything."}, {"start": 615.44, "end": 642.59, "text": "It was amazing. But your body language changed across your arms. I remember you used to have your office in the old tower, right? So you'd come up the tower steps. I think that's part of the intimidation routine. You had the room over the tower, didn't you? It's a beautiful room. We're on the cusp of probably everything I ever dreamed about doing, using it to cure all diseases."}, {"start": 642.59, "end": 664.24, "text": "Help with climate, find new energy sources, find new physics. I think we're on the brink of all of that, the cusp of all of that. So that's amazing. But then, of course, it's a dual purpose technology. It's a general purpose technology. So that means, you know, would be bad actors can repurpose that same technology for harmful ends."}, {"start": 664.24, "end": 693.1, "text": "And that's the real conundrum of how do you embrace the opportunities, allow things like open source alpha fold for everyone to use so that you enable all the amazing work and these two million researchers that are doing incredible work, medical and biological research. How do you enable that kind of thing without also restricting access to the bad actors? And the bad actors could be individuals all the way to rogue nations."}, {"start": 693.1, "end": 720.64, "text": "I've always felt the best way to do this is that as we got closer to AGI, sort of around now, maybe what I have advocated for is we create a CERN-like effort for AGI. There's lots of problems with CERN, but a kind of international research effort like that that could advance the last steps in a very careful, thoughtful way with a lot of research around."}, {"start": 720.64, "end": 748.32, "text": "The guardrails and the safety aspects of this to make sure that we keep control of these increasingly powerful systems. But then that's not enough. You also need something like an international atomic agency to monitor road projects and make sure that that is not proliferation of bad uses and harmful actors bringing access to these types of systems. And then finally,"}, {"start": 748.32, "end": 772.77, "text": "You need some kind of governance mechanism on top. Ideally, that would be some form of UN, but that doesn't seem very possible in today's geopolitics. So I call it like a technical UN, you know, of like 200 of the wisest people in the world from different backgrounds, walks of life and nations representing different aspects of how, so assuming it's technically built correctly."}, {"start": 772.77, "end": 802.54, "text": "How should it be used? That's an ethical and societal question and not one that the technologists and scientists should be making on their own. They should be part of it, but it needs to be civil society, governments, social scientists, ideally philosophers. I think there's a great need for some great new philosophers. I wonder where they are. I hope they're being trained right now. Those of you who are philosophers in the room, I think there's an opportunity for a new Kant or Wittgenstein to arrive and sort of"}, {"start": 802.54, "end": 831.71, "text": "Map out where society should go from here, beyond the technology, because it's going to change everything. And I think it's underestimated. I think it's overestimated and overhyped in the short term, what it's going to do. Various companies and startups are doing that for their own reasons. Also, the timelines, you know, are not quite as short as some people say, but it's still soon, you know, within the next decade. But I think the changes that are going to happen are underestimated still, even now."}, {"start": 831.71, "end": 859.7, "text": "in say 10, 15 years time from now after AGI gets built. Everything's going to change. I think it could be amazing, but it's going to be very different. And we need some really thoughtful, great new thinkers, I think, to kind of map that out to help us as a society navigate that, including great new politicians and philosophers."}, {"start": 859.7, "end": 887.44, "text": "Even as far as theologians, potentially. So, people will agree with you, and then a few people will say, Demis, slow down, because you're way ahead of all the other things that you've cited as desirable, and some people will say it's necessary. So, what do you tell to people who say, just slow down, because you're going too fast? So, look, I actually have some sympathy with that view, and in fact, my original plan for this was..."}, {"start": 887.44, "end": 909.5, "text": "You know, it would be a scientific endeavor and stay a scientific endeavor for a very long time, this kind of CERN idea. And if I had, you know, if I could wave a magic wand and create the ideal setup, it would be everyone would be working on more narrow AI tools like AlphaFold, which are not general intelligences, but they use that same technology."}, {"start": 909.5, "end": 939.36, "text": "And we would be doing a lot of very beneficial things like curing cancer and helping with new materials and some of the scientific stuff while we are learning about what these general systems are. And then we would take that next step. But the technology hasn't turned out that way. It's turned out because the Internet was available. I sometimes think of it as like the Industrial Revolution. This is how I describe it sometimes. What's happened is we're about to have an equivalent of, or maybe it's going to be bigger than the Industrial Revolution."}, {"start": 939.36, "end": 967.28, "text": "But imagine the Industrial Revolution happened, but there was no coal or oil. We just happened to have dead dinosaurs and dead trees in the ground. But it didn't have to be like that. Think about that. That's sort of completely accidental. That's the case. And so if you weren't able to power the Industrial Revolution by just digging up some coal or surface oil, imagine you had to jump straight to solar or nuclear."}, {"start": 967.28, "end": 995.9, "text": "It would take that, how much harder that would have been. You asked a question today about this notion of AI being labor-enhancing and AI being labor-displacing. Do you get a sense of the balance? And what do we do in a society where income gets completely decoupled from employment? Well, look, I feel, first of all, in the next phase, maybe the next 10 years, I think you're going to get what normally happens with technology, which is jobs will be disrupted."}, {"start": 995.9, "end": 1022.5, "text": "But new jobs that we can't even imagine today will appear. And I think I mentioned in the end of the talk, you know, even if we take an example like coding, which so many programmers I know are worried about that because that seems to be, interestingly, one of the things that the UA systems are better at for various reasons, partly because you can create synthetic data on encoding and maths that you can prove are correct or not correct."}, {"start": 1022.5, "end": 1051.17, "text": "So when you're generating the synthetic data, it's very easy to put that back into your training set because you know for sure if it's correct or not. And games are like that too. So there's a few areas of human discipline that are basically amenable to that. But on the other hand, maybe coding and all the creative power that gives you will be available to everyone in society, not just programmers, which is actually a somewhat unnatural thing for our human brains to do."}, {"start": 1051.17, "end": 1080.82, "text": "Logical coding, not that many people can do it. It's beyond the reach of many people. But all of us are creative and maybe could think of great things or applications they would like. Maybe that is becoming into reach now in the next couple of years. So I don't know what that will create. So I think the first phase, the next 10 years, will be like that. Lots of disruption, but net more jobs. Then after that, post-AGI, that becomes a question of then distributing"}, {"start": 1080.82, "end": 1109.65, "text": "The presumably the massively increased productivity gains that you get, then that should imply, you know, if you crack free renewable energy, for example, that changes everything. You know, resource extraction, you can start traveling to space for cheap. We can start human flourishing across the solar system and the galaxy. I mean, it sounds crazy that could be 10 years away, but I believe it is. But then there's different questions of like, how do you distribute that fairly?"}, {"start": 1109.65, "end": 1124.48, "text": "Let's say there's something like universal high basic income or actually even better would be universal basic services perhaps, right? So everyone you can imagine gets the equivalent lifestyle of what a millionaire Western person would have today."}, {"start": 1124.48, "end": 1151.58, "text": "So when we look at the exciting future,"}, {"start": 1151.58, "end": 1176.22, "text": "We were with a friend of yours who had Astra on their phone. And we were in a museum. And they were going around and asking Astra, what am I looking at? Tell people how powerful Astra is. Yeah, it's brilliant. It's still mind-blowing to me when I use it. So you hold your phone camera up, or also you could imagine it on your glasses, and it will understand everything in this room. You could ask it about these pictures. It will tell you about them."}, {"start": 1176.22, "end": 1182.03, "text": "You could ask it, you know, what's nearby, what's a good place to visit. So it'll start becoming..."}, {"start": 1182.03, "end": 1211.68, "text": "Once we get used to these systems, like you have the world's best recommendation system with you, the world's best digital assistant, and you carry it around with you on different surfaces. So, you know, it will be on your desktop, you'll be doing your work with it, put it on your game console, you'll be a game companion, and then you go out into the real world and it will be with you. And I think it will enrich our lives and also make it way more efficient, do a lot of the mundane, everyday things that are a bit of chores for us today, and just take care of that."}, {"start": 1211.68, "end": 1240.88, "text": "Just like anyone who has a great assistant today knows how valuable that is and how amazing that is to save you time and make you more productive. So we kind of want that to be available to everyone. I don't want to say we curse you, but we think of you every time we do online interviews and we can't figure out whether the answers we're getting are genuine or AI. I could go on, but people in this room are much more interesting and intelligent than me. So let me ask one final question."}, {"start": 1240.88, "end": 1269.2, "text": "So we talked about you, the chess protege. We talked about you in the dating game. We talked about you in Queens and everything else. Let's talk about them as the movie star. There's a film coming out called The Thinking Game, which some of us had the incredible privilege to see in prison. I cried, I laughed. It's really touching. Tell us a little bit about how it came together."}, {"start": 1269.2, "end": 1285.86, "text": "Because I want to make sure you get lots of people going to the cinema. Well, the first thing I say, it's hard for me to judge. I hate seeing myself on screen or anything, any interview with me. But there are these two amazing filmmakers, Greg Cozen, Gary Craig."}, {"start": 1285.86, "end": 1314.58, "text": "We've trusted with this. They've actually been following us around, believe it or not, for over a decade. And the reason was they first came to our attention when we were trying to make maybe a documentary about AlphaGo, or at least capture the soul match for posterity. We mostly thought for our own archives, really, and also maybe for future, show future staff what we did. And so they came and they just filmed it all kind of like flies on the wall. And then about a year after that,"}, {"start": 1314.58, "end": 1343.55, "text": "The director came to me and said, look, we've got something really special here. It's a very emotional story. I was like, it's pretty dry. It's like a board game, and then there's some servers, and then there's people programming. It's not even physical, though. How can this be a compelling story? And then they put together, but then, you know, they say, but can you let us do a rough cut, and then you'll see it. And then they did it, and it was amazing. And they captured, because it's a very human story. Actually, if you've seen AlphaGo, the first movie, it's actually about, I think it's a film, not about Go at all."}, {"start": 1343.55, "end": 1367.17, "text": "It's about the spirit of human endeavor and achievement and pushing yourself to the limit, whether that's us coding it or the incredible noble games players, who obviously I empathize with a lot because of my past as a games player. And I know what it would take to be at the level that someone like Lisa Doll is and the absolute mastery of the art that he has."}, {"start": 1367.17, "end": 1395.71, "text": "So it wasn't incredible, and I find that a very emotional film. You know, it's on YouTube for those of you who want to see it. And then because that was so good, they said, can we just follow you around for another few years? And we did fine. And we thought actually, the reason in the back of our minds is we thought it would be important, actually, because AI is going to be so big and so important, it's important for society, I think, to see behind the curtain of what are people like."}, {"start": 1395.71, "end": 1413.95, "text": "And what are their motivations in everyday life when they don't know they're being filmed and so on about why they're trying to do this? You know, what is the motivations? And I would ask all of you, and I say that you should query the motivations of any of the leaders of these technologies or companies. Like, why are they motivated to do it?"}, {"start": 1413.95, "end": 1439.09, "text": "Is it power? Is it money? But in my case, it's about science. And you should understand what their motivation is and then figure out how to interact with them or what to do about that. Because I think that, especially now, because it's so high profile and so many people have got understood now what we understood 20 years ago how important this was going to be, it's attracted a lot of..."}, {"start": 1439.09, "end": 1461.14, "text": "Characters who jump on bandwagons and are doing it for other reasons. I think it's important for society to understand that and make choices about that. What services you use but also whether we want something like CERN to take us through the next steps. I think it's important for..."}, {"start": 1461.14, "end": 1485.86, "text": "We had a problem. How do we choose the students that can come here?"}, {"start": 1485.86, "end": 1514.77, "text": "So we thought the best thing to do is to outsource it to the JCR committee and the MCR committee. And they had to run lotteries and everything else. And despite that, you should see the letters I got. Because people are incredibly inspired by you. So thank you for that. And finally, I want to ask you all to join me in thanking Dennis. We are extremely proud that Dennis is one of us. So please join me in thanking Dennis."}]}