Google DeepMind CEO Demis Hassabis + Google Co-Founder Sergey Brin: Scaling AI, AGI Timeline, Simulation Theory

0:00All right, everybody, we have an amazing crowd here today. we're gonna be live streaming this so let's hear you make some noise everybody can hear that you're here let's go I'm Alex Cantrowitz I'm the host of big technology podcast and I'm here to speak with you about the frontiers of AI with two amazing guests Demis Asabas the CEO of DeepMind is here Google DeepMind good to see you Demis see you too and we have a special guest Sergey Brin the co-founder of Google, is also here.

0:33All right, so this is going to be fun. Let's start with the frontier models. Demis, this is for you. With what we know today about frontier models, how much improvement is there left to be unlocked? And why do you think so many smart people are saying that the gains are about to level off? I think we're seeing incredible progress. We've all seen it today, all the amazing stuff we showed in the keynote. So I think we're seeing incredible gains with the existing techniques, pushing them to the limit. But we're also inventing new things all the time as well. And I think to get all the way to something like AGI, I think may require one or two more new breakthroughs.

1:16And I think we have lots of promising ideas that we're cooking up and we hope to bring into the main branch of the Gemini branch. All right, and so there's been this discussion about scale. You know, does scale solve all problems or does it not? So I want to ask you, in terms of the improvement that's available today, is scale still the star or is it a supporting actor? I think I've always been of the opinion you need both. You need to scale to the maximum the techniques that you know about. You want to exploit them to the limit, whether that's data or compute scale. and at the same time you want to spend a bunch of effort on what's coming next maybe six months a year down the line so you have the next innovation that might do a 10x leap in some way to kind of intersect with the scale so you want both in my opinion I don't know Sergey what do you think I mean I agree it takes both you know you can have algorithmic improvements and simply compute improvements better chips more chips more power bigger data centers I think that historically if you look at things like the n-body problem and simulating you know just gravitational bodies and things like that as you plot it the algorithmic advances have actually beaten out the computational advances even with Moore's law if I had to guess I would say the algorithm that advances are probably going to be even more significant than the computational advances but both of them are coming up now so we're kind of getting the benefits of both and Demis do you think the majority of your improvement is coming from building bigger data centers and using more chips like there's talk about how the world will be just wallpapered with data centers is that your vision well no look I mean it we're definitely gonna need a lot more data centers it's amazing that you know it still amazes me from a scientific point of view, we turn sand into thinking machines.

3:15It's pretty incredible. But actually, it's not just for the training. It's now we've got these models that everyone wants to use. And actually, we're seeing incredible demand for 2.5 Pro. And I think Flash, we're really excited about how performant that is for the incredible low cost. I think the whole world is going to want to use these things. And so we're going to need a lot of data centers for serving. And also for inference time compute, Giving you know you saw you saw deep think today 2.5 pro deep think the more time you give it the better It will be and certain tasks very high value very difficult tasks You want to it will be worth letting it think for a very long time and we're thinking about how to push that even further and Again, that's going to require a lot of chips at runtime.

4:00Okay, so you brought up test time compute We've been about a year into this reasoning paradigm and you and I have spoken about it twice in the past as something that you might be able to add on to traditional LLMs to get gains. So I think this is like a pretty good time for me to be like, what's happening? Can you help us contextualize the magnitude of improvement we're seeing from reasoning? No, we've always been big believers in what we're now calling this thinking paradigm. If you go back to our very early work on things like AlphaGo and AlphaZero, our agent work on playing games, they will all have this type of attribute of a thinking system on top of a model.

4:37And actually you can quantify how much difference that makes if you look at a game like chess or go You know we had versions of alpha go now for zero with the thinking turned off So it was just the model telling you its first idea and you know, it's not bad It's maybe like master level something like that But then if you turn the thinking on it's be way beyond world champion level You know, it's like a 600 ELO plus difference between the two versions So you can see that in games let alone for the real world which is way more complicated and And I think the gains will be potentially even bigger by adding this thinking type of paradigm on top.

5:13Of course, the challenge is that your models, and I talked about this earlier in the talk, need to be a kind of world model. And that's much harder than building a model of a simple game, of course. And it has errors in it, and yet those can compound over longer-term plans. But I think we're making really good progress on all those fronts. Yeah, look, I mean, as Demis said, DeepMind really pioneered a lot of this reinforcement learning work and what they did with AlphaGo and AlphaZero, as you mentioned. It showed, as I recall, something you would take 5 ,000 times as much training to match what you were able to do with still a lot of training and the inference time compute that you were doing with Go.

5:58so it's obviously a huge advantage and obviously like most of us we get some benefit by thinking before we speak and although not always I was reminded to do that but I think that the the AI is obviously are much stronger once you add that capability and I think we're just at the tip of the iceberg right now in that sense. It's been less than a year than these models have really been around. Especially if you think about obviously with an AI, during its thinking process it can also use a bunch of tools or even other AIs in during that thinking process to improve what the final output is. So I think it's going to be an incredibly powerful paradigm.

6:46Deep think is very interesting. I'm going to describe it, I'm trying to describe it right. It's basically a bunch of parallel reasoning processes working and then checking each other and then it's like reasoning on steroids now Demis you mentioned that the industry needs a couple more advances to get to AGI where would you put this type of mechanism is this one of those that might get the industry closer I think so I think it's it's maybe part of one okay shall I shall we say and there are others too that we need to we know maybe this can be part of improving reasoning where Where does true invention come from where you're not just solving a mass conjecture, you're actually proposing one or hypothesizing a new theory in physics.

7:30I think we don't have systems yet that can do that type of creativity. I think they're coming. And these types of paradigms might be helpful in that, things like thinking and then probably many other things. I think we need a lot of advances on the accuracy of the world models that we're building. I think you saw that with VO, the potential VO3 of how it amazes me like the how it can intuit the physics of the light and the gravity having someone who I used to work on on computer games not just the AI but also graphics engines in my early career and remember having to do all of this by hand you know and program all of the lighting and the shaders and all of these things incredibly complicated stuff we used to do in early games and now it's It's just intuiting it within the model.

8:17It's pretty astounding. I saw you shared an image of a frying pan with some onions and some oil. There was no subliminal messaging about that? No, not really. Not really. Just maybe a subtle message. Okay. So we said the word AGI or the acronym AGI a couple times. There's, I think, a movement within the AI world right now to say, let's not say AGI anymore. The term is so overused as to be meaningless. But Demis, it seems like you think it's important. Why? Yeah, I think it's very important. But I think, I mean, maybe I need to write something about this also with Shane Legg, who's our chief scientist, who was one of the people who invented the term 25 years back.

8:59I think there's sort of two things that are getting a little bit conflated. One is, like, what can a typical person do, an individual do? and we can you know we're all very capable but we can only do however capable you are there's only a certain slice of things that one is expert in right and or you know you could say what can you do what like 90 % of humans can do that is obviously going to be economically very important and I think from a product perspective also very important so it's a very important milestone so maybe we should say that's like you know typical human intelligence but what I'm interested in and what I would call AGI is really a more theoretical construct which is what is the human brain as an architecture able to do?

9:40And the human brain is an important reference point because it's the only evidence we have maybe in the universe that general intelligence is possible. And there, you would have to show your system was capable of doing the range of things even the best humans in history were able to do with the same brain architecture. It's not one brain, but the same brain architecture. So what Einstein did, what Mozart was able to do, what Marie Curie, and so on. And that, it's clear to me today's systems don't have that. And then the other thing that why I think it's sort of overblown the hype today on AGI is that our systems are not consistent enough to be considered to be fully general yet.

10:15They're quite general, so they can do, you know, thousands of things. You've seen many impressive things today. But every one of us have experience with today's chatbots and assistants. You can easily, within a few minutes, find some obvious flaw with them. Some high school math thing that it doesn't solve. some basic game it can't play. It's not very difficult to find that, those holes in the system. And for me, for something to be called AGI, it would need to be consistent, much more consistent across the board than it is today. It should take a couple of months for maybe a team of experts to find a hole in it, an obvious hole in it.

10:55Whereas today, it takes individual minutes to find that. Sergey, this is a good one for you. Do you think that AGI is going to be reached by one company and it's game over or could you see Google having AGI Open AI having AGI and thropic having AGI China having AGI Wow, that's a great question. I mean I Guess I would suppose that one Company or country or entity will reach AGI first now. It is a little bit of a spectrum. It's not like a completely precise thing, so it's conceivable that there will be more than one roughly in that range at the same time. After that, what happens, I mean, I think it's very hard to foresee, but you could certainly imagine there's going to be multiple entities that come through.

11:45And in our AI space, we've seen whatever, when we make a certain kind of advance, like other companies are quick to follow and vice versa when other companies make certain advances it's you know it's a kind of a constant leapfrog so I do think there's an inspiration element that you see and that would probably encourage more and more entities to cross that threshold then miss what do you think well I think we we probably do I think it is important for the field to agree on a definition of AGI so I will maybe we should try and help that's a coalesce. Assuming there is one, there probably will be some organizations that get there first.

12:26And I think it's important that those first systems are built reliably and safely. And I think after that, if that's the case, we can imagine using them to shard off many systems that have safe architectures sort of provably underneath them. And then you could have personal AGI's and all sorts of things happening. But it's You know, it's quite difficult as Sergei says it's pretty difficult to predict Sort of see beyond the event horizon to predict what that's going to be like right So we talked a little bit about the definition of AGI and a lot of people have said AGI must be knowledge Right the intelligence of the brain.

13:05What about the intelligence of the heart? Demis briefly? Does does AI have to have emotion to be considered AGI? Can it have emotion? I think you will need to understand emotion. I don't know if I think it will be a sort of almost a design decision if we wanted to mimic emotions. I think there's no, I don't see any reason why it couldn't in theory. But it might be different or it might be not necessary or in fact not desirable for them to have the sort of emotional reactions that we do as humans. So I think, again, it's a bit of an open question as we get closer to this AGI timeframe sort of events, which I think is more on a five to 10 year timescale.

13:44So I think we have a bit of time not much time but some time to research those kinds of questions When I think about how the timeframe might be shrunk I wonder if it's going to be the creation of self-improving systems and last week I almost fell out of my chair reading this headline about something called Alpha Evolve Which is an AI that helps design better algorithms algorithms and even improve the way LLM's train so Demis are you trying to cause an intelligence explosion no not an uncontrolled one I look I think it's an interesting first experiment it's amazing system great team or that's working on that where it's interesting now to start pairing other types of techniques in this case evolutionary programming techniques with the latest foundation models which are getting increasingly powerful and I actually want to see in our exploratory work a lot more of these combinatorial systems and pairing different approaches together.

14:43And you're right. That is one of the things, a self-improvement, someone discovering a self-improvement loop would be one way where things might accelerate further than they're even going today. And we've seen it before with our own work, with things like AlphaZero learning chess and Go and any two-player game from scratch within less than 24 hours, starting from random, with self improving processes so we know it's possible but again those are in quite limited game domains which are very well described so the real world is far messier and far more complex so remains to be seen if that type of approach can work in a more general way Sergey we've talked about some very powerful systems and it's a race it's a race to develop these systems is that why you came back to Google?

15:34I mean I think as a computer scientist it's a very unique time in history like honestly anybody who's computer scientist should not be retired right now should be working on AI that's what I would just say I mean there's just never been a greater sort of problem an opportunity a greater cusp of technology so I don't I wouldn't say it's because of the race although we fully intend that Gemini will be the very first AGI clarify that but to be immersed in this incredible technological revolution I mean it's unlike you know I went through sort of the web 1.0 thing it was very exciting and whatever we had mobile we have this we have that but I think this is scientifically far more exciting and I think I think ultimately the impact on the world is going to be even greater and as much as you know the web and mobile phones have had a lot of impact I think AI is going to be vastly more transformative so what do you do day to day I think I torture people like Demis It was amazing, by the way, he tolerated me crashing this fireside.

16:56I'm across the street pretty much every day, and there are just people who are working on the key Gemini text models, on the pre-training, on the post-training. Mostly those I periodically delve into some of the multimodal work. VO3, as you've all seen. but I tend to be pretty deep in the technical details and that's a luxury I really enjoy fortunately because guys like Demis are you know minding the shop and yeah that's just where you know my scientific interest is it's deep in the algorithms and how they can evolve okay let's talk about the products a little bit, some that were introduced recently.

17:45I just want to ask you a broad question about agents, demos, because when I look at other tech companies building agents, what we see in the demos is usually something that's contextually aware, has a disembodied voice, is often interacted with, you often interact with it on a screen. When I see DeepMind and Google demos, oftentimes it's through the camera. It's very visual. There was an announcement about smart classes today. so talk a little bit about if that's the right read why why Google is so interested in having an assistant or a companion that is something that sees the world as you see it well it's for several reasons several threads come together so as we talked earlier we've always been interested in agents that's actually the the heritage of deep mine actually we started with agent based systems in games we are trying to build a GI which is a full general intelligence clearly that would have to understand the physical environment the physical world around you.

18:42And two of the massive use cases for that, in my opinion, are a truly useful assistant that can come around with you in your daily life, not just stuck on your computer or one device. We want it to be useful in your everyday life for everything. And so it needs to come around you and understand your physical context. And then the other big thing is I've always felt for robotics to work, you sort of want what you saw with Astra on a robot. And I've always felt that the bottleneck in robotics isn't so much the hardware, although obviously there's many, many companies working on fantastic hardware, and we partner with a lot of them, but it's actually the software intelligence that I think is always what's held robotics back.

19:21But I think we're in a really exciting moment now where finally, with these latest versions, especially 2.5 Gemini, and more things that we're going to bring in, this kind of VO technology and other things, I think we're going to have really exciting algorithms to make robotics finally work and sort of realize its potential, which could be enormous. So I think this... And then in the end, AGI needs to be able to do all of those things. So for us, and that's why you can see we always had this in mind, that's why Gemini was built from the beginning, even the earliest versions, to be multimodal. And that made it harder at the start because it's harder to make things multimodal than just text only.

19:59But in the end, I think we're reaping the benefits of those decisions now, and I see many of the Gemini team here in the front row, of the correct decisions we made. They were the hardest decisions, but we made the right decisions and now you can see the fruits of that with all of what you've seen today actually sorry I've been thinking about whether to ask you a Google Glass question oh far away what did you learn from Glass that Google might be able to apply today now that it seems like smart glasses have made a reappearance wow yeah a great question I learned a lot I mean that was I I definitely feel like I made a lot of mistakes with Google Glass, I'll be honest.

20:40I am still a big believer in the form factor, so I'm glad that we have it now. And now it looks like normal glasses, doesn't have the thing in front. I think there was a technology gap, honestly. Now in the AI world, the things that these glasses can do to help you out without constantly distracting you that capability is much higher there's also just I just didn't know anything about consumer electronics supply chains really and how hard it would be to build that and have it be at a reasonable price point managing all the manufacturing so forth this time we have great partners that are helping us build this.

21:25So that's another step forward. What else can I say? I do have to say I miss the airship with the wingsuiting skydivers for the demo. Honestly, it would have been even cooler here at Shoreline Amphitheater than it was up in Moscone back in the day, but maybe we'll have to... We should probably polish the product first this time. We'll Make sure it's ready and available, and then we'll do a really cool demo. So that's probably a smart move. Yeah, what I will say is, I mean, look, we've got obviously an incredible history of glass devices and smart devices. We can bring all those learnings to today.

22:04And I'm very excited about our new glasses, as you saw. What I was always talking to our team and Sharam and the team about is that, I mean, I don't know if Sergei would agree, but I feel like the universal assistant is the killer app for smart glasses. And I think that's what's going to make it work, apart from the fact that the hardware technology has also moved on and improved a lot. I feel like this is the actual killer app, the natural killer app for it. Okay. Briefly on video generation, I sat in the audience in the keynote today and was fairly blown away by the level of improvement we've seen from these models.

22:43And I mean, you had filmmakers talking about it in the presentation. I want to ask you Demis specifically about model quality if the internet fills with video that's been made with artificial intelligence Does that then go back into the training and lead to a lower quality model than if you were training just from human generated content? Yeah, well look we know there's a lot of worries about this so-called like model collapse I mean video is just one thing but in any modality text as well There's a few things to say about that first of all we're very rigorous with our data quality management and curation We also at least for all of our generative models.

23:24We attach synth ID to them So there's this invisible AI actually made watermark that is pretty very robust as held up now for you know Year 18 months since we released it and all of our images and videos are embedded with this watermark. So we can detect, and we're releasing tools to allow anyone to detect these watermarks and know that that was an AI-generated image or video. And, of course, that's important to combat deepfakes and misinformation, but it's also, of course, you could use that to filter out, if you wanted to, whatever was in your training data. So I don't actually see that as a big problem.

24:03Eventually, we may have video models that are so good, you could put them back into the loop as a source of additional data, synthetic data it's called. And there you've just got to be very careful that you're actually creating from the same distribution that you're going to model. You're not distorting that distribution somehow. The quality is high enough. We have some experience of this in a completely different way with things like AlphaFold, where there wasn't actually enough real experimental data to build the final AlphaFold. So we had to build an earlier version that then predicted about a million protein structures and then we selected it had a confidence level on that we selected the top three four hundred thousand and put them back in the training data so there's lots of it's very cutting-edge research to like mix synthetic data with real data so there are also ways of doing that but on the terms of the video sort of generator stuff you can just exclude it if you want to at least with our own work and hopefully other the gen media companies follow suit and put robust watermarks in also obviously first and foremost to combat deep fakes and misinformation.

25:05Okay we have four minutes I got four questions left we now move to the miscellaneous part of my questions let's see how many we can get through and as fast as we can get through them. Let's go to Sergey with this one what does the web look like in ten years? What does the web look like in ten years? I mean one minute boy I think ten years because of the rate of progress in AI is so far beyond anything we can see. Let's guess. Not just the web yeah I mean I I don't think we really know what the world looks like in 10 years. Okay. Demis? Well, I think that's a good answer. I do think the web, I think in nearer term, the web is going to change quite a lot if you think about an agent-first web.

25:47Does it really need to, you know, it doesn't necessarily need to see renders and things like we do as humans using the web. So I think things will be pretty different in a few years. Okay. This is kind of an under-over question. AGI before 2030 or after 2030? 2030, boy, you really kind of put it on that fine line. I'm going to say before. Before? Yeah. Dennis? I'm just after. Just after. Yeah. Okay. No pressure, Dennis. Exactly. I have to go back and get working harder. Is that... I can ask for it. He needs to deliver it. Exactly. Stop sandbagging. We need that next week. That's true. I'll come to the review.

26:32All right, so would you hire someone that used AI in their interview? Demis? Oh, in their interview? Depends how they used it. I think using today's models tools, probably not. But I think that would be, well, it depends how they would use it, actually. I think it's probably the answer. Sergey? I mean, I never interviewed at all. So, I don't know. I feel it would be hypocritical for me to judge people exactly how they interview. Yeah, I haven't either, actually. So snap on that. I've never done a job with you. Okay. So, Demis, I've been reading your tweets. You put a very interesting tweet up where there was a prompt that created some sort of natural scene.

27:20Oh, yeah. Here was the tweet. Nature to simulation at the press of a button does make you wonder with a couple of emojis, and people ran with that and wrote some headlines saying, Demis thinks we're in a simulation. Are we in a simulation? Not in the way that Nick Bostrom and people talk about. I do think, though, this... So I don't think this is some kind of game, even though I wrote a lot of games. I do think that ultimately underlying physics is information theory. So I do think we're in a computational universe, but it's not just a straightforward simulation. I can't answer you in one minute.

27:56but but I think I think the fact that these systems are able to model real structures in nature is quite interesting and telling and I've been thinking a lot about our work we've done with alpha go and alpha fold in these types of systems I spoken a little bit about it maybe at some point I'll write up a scientific paper about what I think that really means in terms of what's actually going on here in reality Sergei you want to make a headline Well, I think that argument applies recursively, right? If we're in a simulation, then by the same argument, whatever beings are making the simulation are themselves in the simulation for roughly the same reasons, and so on and so forth.

28:39So I think you're going to have to either accept that we're in an infinite stack of simulations or that there's got to be some stopping criteria. And what's your best guess? I think that we're taking a very anthropocentric view, like when we say simulation in the sense that some kind of conscious being is running a simulation that we are then in and that they have some kind of semblance of desire and consciousness that's similar to us. I think that's where it kind of breaks down for me. So I just don't think that we're really equipped to reason about sort of one level up in the hierarchy. Okay, well, Demis, Sergey, thank you so much.

29:27This has been such a fascinating conversation. Thank you. Thank you all. All right. Thanks, Alex. Thank you. Sergey. Pleasure.

29:44We'll see you next time.