Google DeepMind CEO Demis Hassabis: The Path To AGI, Deceptive AIs, Building a Virtual Cell

And I think today's systems are still pretty far away from having that kind of creative inventive capability. OK, so a couple of years away till we hit AGI. I think, you know, I would say probably like three to five years. So if someone were to declare that they've reached AGI in 2025, probably marketing. I think so. I mean, I think there's a lot of hype in the area, of course. I mean, some of it's very justified. I mean, I would say that AI research today is overestimated in the short term. I think probably a bit overhyped at this point, but still underappreciated and very underrated about what it's going to do in the medium to long term.

So it's sort of we're still in that weird kind of space. And I think part of that is, you know, there's a lot of people that need to do fundraising, a lot of startups and other things. And so I think we're going to have quite a few sort of fairly outlandish and slightly exaggerated claims. And, you know, I think that's a bit of a shame, actually. Yeah. In the AI products, what's it going to look like on the path there? I mean, you've talked about memory again, planning, being better at some of the tasks that it's not excelling at at the moment. So when we're using these AI products, let's say we're using Gemini, what are some of the things that we should look for in these domains that will make us say, oh, OK, it seems like that's a step closer?

and that's a step closer yeah so i think um uh today's systems you know obviously we're very proud of gemini 2.0 i'm sure we're going to talk about that but i feel like um they're very useful for still quite niche tasks right if you're doing some research perhaps you're summarizing some area of research incredible you know i use notebook lm and deep research all the time to kind of especially like um break the ice on a new area of research that i want to get into or summarize There's some, you know, maybe a fairly mundane set of documents or something like that. So they're extremely good for certain tasks and people are getting a lot of value out of them, but they're still not pervasive, in my opinion, in everyday life, like helping me every day with my research, my work, my day-to-day, my daily life too.

And I think that's where we're going with our products, with building things like Project Astra, our vision for universal assistant, is it should be involved in all aspects of your life and be enriching, helpful, and making that more efficient. And I think part of the reason is these systems are still fairly brittle, partly because they are quite flawed still and they're not AGI's. And you have to be quite specific, for example, with your prompts or that you need a lot of there's quite a lot of skill there in coaching or guiding these systems to be useful and to stick to the areas they're good at.

And a true AGI system shouldn't be that difficult to coax. It should be much more straightforward, you know, just like talking to another human. Yeah. And then on the reasoning front, you said that's another thing that's missing. I mean, that's everybody's talking about reasoning right now. So how does that end up getting us closer to artificial general intelligence? So reasoning and mathematics and other things, and there's a lot of progress on maths and coding and so on. But let's take maths, for example. You have systems, some systems that we work on, like alpha proof, alpha geometry, that are getting, you know, silver medals in maths olympiads, which is fantastic.

But on the other hand, some of our systems, those same systems, are still making some fairly basic mathematical errors, right, for various reasons. Like the classic, you know, counting the number of R's and strawberries and so on and the word strawberry and so on. And is 9.11 bigger than 9.9? And things like that. And, of course, you can fix those things and we are and everyone's improving on those systems. But we shouldn't really be seeing those kinds of flaws in a system that is that capable in other domains, in more narrow domains of doing, you know, Olympiad level mathematics. So there's something still a little bit missing, in my opinion, about the robustness of these systems.

And then I think that speaks to the general. of these systems. A truly general system would not have those sorts of weaknesses. It would be very, very strong, maybe even better than the best humans in some things like playing Go or doing mathematics, but it would be overall consistently good. Now, can you talk a little bit about how these systems are attacking math problems? Because I think the general understanding of these systems is, the LLMs, is they encompass all of the world's knowledge and then they predict what somebody might answer if they were asked a question. But it's kind of different when you're working step by step through an algorithm through a math problem yes that's not enough of course you know just understanding the world's information and and then trying to sort of almost compress that into your memory that's not enough for solving a novel maths problem or novel novel conjecture so there you know we start needing to bring in I think we talked about this last time more kind of like alpha go planning ideas into the mix with these large foundation models which are now beyond just language they're multimodal of course. And there, what you need to do is you need to have your system not just pattern matching

roughly what it's seeing, which is the model, but also planning and be able to kind of go over that plan. We revisit that branch and then go into a different direction until you find the right criteria or the right match to the criteria that you're looking for. And that's very much like the kind of games playing AI agents that we used to build for Go, chess, and so on. They had those aspects. And I think we've got to bring them back in. But now working in a more general way on these general models, not just a narrow domain like games. And I think that also that approach of a model guiding a search or planning process so it's efficient, works very well with mathematics as well. And you can sort of turn maths into a kind of game like search. Right. And I want to ask about math like once these models get math right is that generalizable because i think there was like a whole hubbub uh when people first learned about reasoning systems and they're like oh this is like this is going to be a problem these models are getting smarter than we can control because if they can do math then they can do x y and z so is that generalizable or is it like we're going to teach them how to do math and they can just do math i think for now uh the jury's out on that i

mean i feel like it's a clearly a capability you want of a general agi system uh it can be very powerful in itself obviously mathematics is is extremely general in itself um but it's not clear you know maths and even coding and games these are areas they're quite special uh uh areas of knowledge because you can verify if the answer is correct, right, in all of those domains, right? The math, you know, the final answer the AI system puts out, you can check whether that math, that solves the conjecture or the problem. So, but most things in the general world, which is messy and ill-defined, do not have easy ways to verify whether you've done something correct.

So that puts a limit on these self-improving systems if they want to go beyond these areas of, you know, maybe very highly defined spaces like mathematics, coding, or games. So how are you trying to solve that problem? Well, you know, you've got to, first of all, you've got to build general models, world models, we call them, to understand the world around you, the physics of the world, the dynamics of the world, the spatial temporal dynamics of the world, and so on, and the structure of the real world we live in. And of course, you need that for a universal assistance. So Project Astra is our project built on Gemini to do that, to understand, you know, objects and the context around us.

I think that's important if you want to have an assistant. But also robotics requires that, too. Of course, robots are physically embodied AIs, and they need to understand their environment, the physical environment, the physics of the world. So we're building those types of models. And also you can you can also use them in simulation to understand game environments. So that's another way to bootstrap more data to understand, you know, the physics of the world. But the issue at the moment is that those models are not 100% accurate, right? So they, you know, maybe they're accurate 90% of the time or even 99% of the time.

But the problem is if you start using those models to plan, maybe you're planning 100 steps in the future with that model. even if you only have a one percent error and what the model's telling you that's going to compound over a hundred steps to the point where you'll be in a you know you'll kind of get almost a random answer and so that makes the planning very difficult whereas with maths with gaming with coding you can verify each step are you still grounded uh to reality um and is the final answer mapped to what you're expecting and and so um i think part of the answer is to is to make them the world models more and more sophisticated and more and more accurate and not hallucinate and all of those kinds of things. So you get, you know, the errors are really minimal.

Another approach is to plan not at each sort of linear time step, but actually What's called hierarchical planning. Another thing we used to, you've done a lot of research on in the past, and I think is going to come back into vogue, where you plan at different levels of temporal abstraction. So instead of that, that could, that could also alleviate the need for your model to be super, super accurate, because you're not planning over hundreds of time steps, you're planning over only a handful of time steps, but at different levels of abstraction. How do you build a world model? Because, you know, I always thought it was going to be like send robots out into the world and have them figure out how the world works.

But one thing that surprised me is with these video generation tools. Yes. You would think that if the AI didn't have a good world model, then nothing would really fit together when they try to figure out how the world works as they show you these videos like VO2, for instance. But they actually get the physics pretty right. Yeah. So can you get a world model just by showing an AI video? Do you have to be out in the world? How is this going to work? It's interesting and actually been pretty surprising, I think, to the extent of how far these models can go without being out in the world, right, as you say.

So VO2, our latest video model, which is actually surprisingly accurate on things like physics. You know, there's this great demo that someone created of like chopping a tomato with a knife, right, and getting the slices of the tomato just right and the fingers and all of that. And VO is the first model that can do that. If you look at other competing models, the tomato sort of randomly comes back together or the finger splits from the knife. So those things are, if you think about it really hard, you've got to understand consistency across frames, all of these things. And it turns out that you can do that by using enough data and viewing that.

I think these systems will get even better if they're supplemented by some real-world data collected by an acting robot or even potentially in very realistic simulations where you have avatars that act in the world too. So I think that's the next big step actually for agent-based systems is to go beyond world models. Can you collect enough data where the agents are also acting in the world and making plans and achieving tasks? And I think for that, you will need not just passive observation, you will need actions, active participation. I think you just answered my next question, which is if you develop AI that can reasonably plan and reason about the world and has a model of how the world works, it can, and it seems like that's the answer.

It can be an agent that can go out and do things for you. Yes, exactly. And I think that's what will unlock robotics. Right. I think that's also what will then allow this. notion of a universal assistant that can help you in your daily life across both the digital world and the real world um that's what that's the that's the thing we're missing um and i think that's going to be incredibly powerful and useful uh tool you can't get there then by just scaling up the current models and building you know hundreds of thousands or million gpu clusters like elon's doing right now and that's not going to be the path to agi um well look i actually think it so my view is a bit more nuanced than that is like that that the scaling approach is absolutely working of course that's where we've why we got to where we have now um one can argue about are we getting diminishing returns or what do you think about that question what my view is that we are getting substantial returns but not but it's slowing okay so vis-a-vis but but it would have to i mean it's it's not just continuing to be exponential but that doesn't mean the scaling's not working it's absolutely working and we're still getting you know as you see gemini 2 over

Gemini 1.5. And by the way, the other thing that was working with the scaling is also making efficiency gains on the smaller size models. So the cost or the size per performance is radically improving under the hood as well, which is very important for scaling, you know, the adoption of these systems. But yeah, so you know, you've got the scaling part. And that's absolutely needed to build more and more sophisticated world models. But then I think we are missing or we need to reintroduce some ideas on the planning side, memory side, the searching side, the reasoning to build on top of the model. The model itself is not enough to be an AGI. You need this other capability for it to act in the world and solve problems for you. And then there's still the additional question mark of the invention piece and the creativity piece, true creativity, beyond mashing together what's already known, right?

And that's also unknown yet, if something new is required or, again, if existing techniques will eventually scale to that. I can see both arguments. And I think, from my perspective, it's an empirical question. We just got to push both the scaling and the invention part to the limit. And fortunately, at Google DeepMind, we have a big enough group, we can invest in both those things. So Sam Altman recently said something that caught people's eye. He said, we are now confident we know how to build AGI as we have traditionally understood it. It just seems by listening to what you're saying that you feel the same way.

Well, it depends what we, you know, I think the way he said that was quite ambiguous. us right so in the sense of like oh we're building it right now and here's the abc to do it what i would say and if this what it was meaning i would agree with it is that we we roughly know the zones of techniques that required what's probably missing which bits need to be put together but um that's still a credible amount of research in my opinion that needs to be done to get that all to work even if that was the case and that's and i think there's a 50 chance we are uh missing some new techniques you know maybe we need one or two more transformer-like breakthroughs and I think I'm genuinely uncertain about that so that's why I say 50% so I mean I wouldn't be surprised either way if we got there with existing techniques and things we already knew but put them together in the right way and scaled that up or if it turned out one or two things were missing so let's talk about creativity for a moment I mean you brought it up a couple times here that the models are gonna have to be creative they're gonna have to learn how to invent if we want to call it which is where everybody's trying to go i was re-watching the alpha go documentary oh yeah and the algorithms make a creative move they do

uh move 37 37 yes i just had it okay thank you uh that's interesting because it was a couple years ago the algorithms were already being creative yes why have we not really seen creativity from large language models i mean this is to me i think the greatest disappointment that people have with these tools is like they say this is very impressive work but it's just limited to the training set will mix and match what it knows but it can't come up with anything new yeah well look so what and I should probably ride this up but what I sometimes talk about in talks ever since the alpha go match which is now you know eight plus years ago amazingly right that happened that was probably the reason that was such a watershed moment for AI was first of all there was the Everest of you know cracking go right which was always considered to be one of the holy grails of AI. So we did that. Second thing was the way we did it, which was these learning systems that were generalizable, right? Eventually it became Alpha Zero and so on, even play any two-player game and so on. And then the third thing was this Move 37. So not only did it win 4-1, it beat Lisa Dole, the great Lisa Dole 4-1, it also played

original moves. So I have three categories of originality or creativity. The most basic kind of mundane form is just interpolation which is like averaging of what you see so if i said to a system you know come up with a new picture of a cat and it's seen a million cats and it produces some kind of average of all the ones it's seen in theory that's an original cat because You won't find the average in the specific examples, but it's a pretty boring, you know, it's not really very creative. I wouldn't call that creativity. That's the lowest level. Next level is what AlphaGo exhibited, which is extrapolation.

So here's all the games humans have ever played. It's played another million games on top of, you know, 10 million games on top of that. And now it comes up with a new strategy in Go that no human has ever seen before. That's Move 37, right? Revolutionizing Go, even though we've played it for thousands of years. So that's pretty incredible. And that could be very useful in science. And that's why I got very excited about that and started doing things like AlphaFold because clearly extrapolation beyond what we already know or what's in the training set could be extremely useful. So that's already very valuable and I think truly creative.

But there's one level above that that humans can do, which is invent Go. Can you invent me a game that, you know, if I specify it to an abstract level, you know, takes five minutes to learn the rules, but a lifetime to many lifetimes to master. It's beautiful aesthetically. encompasses some sort of mystical part of the universe in it that it's beautiful to look at it but you can play a game in a human afternoon in two hours right that's the that's that would be a high level specification of go and then somehow the system's got to come up with a game that's as elegant and as beautiful and and perfect as go now we can't do that now the question is why is it that we don't know how to specify that type of goal to our systems at the moment what's the objective function is very amorphous it's very abstract so I'm not sure if it's just we need higher level more abstracted layers in our systems building more and more abstract models so we can talk to it in this way give it those kind of amorphous goals or is there a missing capability actually about that that we still have human intelligence has that's still missing from our systems and again I'm unsure about that which which way that is I can see arguments both ways

and we'll try both but i think the thing that people are upset or not upset but people are disappointed by is they don't even see a move 37 in today's llms what's going on there okay so well that's because i don't think we have so if you look at alpha go and i'll give you an example there which which maps to today's llms so um you can run alpha go and alpha zero our chess program general two-player program without the search and the reasoning part on top you can just run it with model so what you say is to the model come up with the first go move you can think of in this position that's most the most pattern match most like Okay, and it can do that, and it'll play a reasonable game, but it will only be around master level, possibly grandmaster level. It won't be world champion level, and it certainly won't come up with original moves. For that, I think you need the search component to get you beyond where the model knows about, which is mostly summarizing existing knowledge, to some new part of the tree of knowledge, right?

So you can use the search to get beyond what the model currently understands. And that's where I think you can get new ideas like, you know, move 37. What's it searching the web? No. So, well, it depends on what the domain is searching that that knowledge tree. So obviously in Go, it was searching Go moves beyond what the model knew. I think for language models, it will be searching the world model for new parts, configurations in the world that are useful. So, of course, that's so much more complicated, which is why we haven't seen it yet. But I think the agent-based systems that are coming will be capable of Move 37 type things.

So are we setting too high of a bar for AI? Because I'm curious if you've learned anything about humanity doing this work. It seems like we almost give too much of a premium on humanity or individual people's ingenuity, where a lot of us, we kind of take in stuff, we spit it out, like our society really works in memes, like we have a cultural thing and it gets translated. So what have you learned about the nature of humans from doing the work with the AIs? Well, look, I think humans are incredible and especially the best humans in the best domains. I love watching any sports person or talented musician or games player at the top of their game.

The absolute pinnacle of human performance is always incredible no matter what it is. So I think as a species, we're amazing. Individually, we're also kind of amazing what everyone can do with their brains so generally, right? deal with new technologies. I mean, I'm always fascinated by how we just adapt to these things sort of almost effortlessly as a society and as individuals. So that speaks to the power and the generality of our minds. Now, the reason I have set the bar like that, and I don't think it's a question of like, can we get economic worth out of these systems? I think that's already coming very soon. But that's not what AGI shouldn't be. I think we should treat AGI with scientific integrity not just move goalposts for commercial reasons or whatever it is hype and so on and there the the definition of that was always having a system that was, you know, if we think about it theoretically, that was capable of being as powerful as a Turing machine. So Alan Turing, one of my all-time scientific heroes, you know, he described a Turing machine, which underpins all modern computing, right, as a system that can simulate any other compute, can compute anything that's computable. So we know, we have the theory there

that if an AI system is Turing-powerful, it's called, if it can simulate a Turing machine, then it's able to calculate anything in theory that is computable. And the human brain is probably some sort of Turing machine, at least that's what I believe. And so in order for out to now, and I think that's what AGI is, is a system that's truly general, and in theory could be applied to anything. And the only way we'll know that is if we, it exhibits all the cognitive capabilities that humans have, assuming that human, the human mind is a type of Turing machine, or is at least as powerful as a Turing machine. So that's my always been my sort of bar. It seems like people are trying to rebadge things as that as being what's called asi artificial super intelligence but i think that's beyond that that's after you have that system and then it starts going beyond in certain domains what humans are capable of uh potentially inventing themselves okay so when i see everybody making the same joke on the same topic on twitter it's and i say oh that's just us being llms i think i'm selling humanity a little short uh well yes i guess so i guess okay Yeah. I want to ask you about deceptiveness. I mean, one of the most interesting things I saw at the end of last year was that these AI bots are starting to try to fool their evaluators and they don't want their initial training rules to be thrown out the window.

So they'll like take an action that's against their values in order to be able to remain the way that they were built. Yes. That's just incredible stuff to me. I mean, I know it's scary to researchers, but it blows my mind that it's able to do this. Are you seeing similar things in the stuff that you're testing within DeepMind? And what are we supposed to think about all this? Yeah, we are. And I'm very worried about, I think, deception specifically is one of those core traits you really don't want in a system.

Demis HassabisThe reason that's like a kind of fundamental trait you don't want is that if a system is capable of doing that, it invalidates all the other tests that you might think you're doing, including safety ones.

It's in testing and it's like, going to give a difference. five years yeah it's it's playing some meta game right and then and that's incredibly dangerous if you think about then it invalidates all that all of The results of your other tests that you might, you know, safety tests and other things you might be doing with it. So I think there's a handful of capabilities like deception, which are fundamental and you don't want, and you want to test early for, and I've been encouraging the safety institutes and evaluation benchmark builders, including, and also obviously all the internal work we're doing to look at deception as a kind of class A thing that we need to prevent and monitor.

As important as tracking the performance and intelligence of the systems. The answer to this as well, and one way to, there's many answers to the safety question of, and a lot of research, more research needs to be done in this very rapidly, is things like secure sandboxes. So we're building those two. We're world-class here at security, at Google, and at DeepMind. And also we are world-class at games environments. And we can combine those two things together to kind of create digital sandboxes with guardrails around them sort of the kind of guardrails you'd have for for cybersecurity but internal as well as blocking external actors and and then test these agent systems in those kind of secure sandboxes that will probably be a good advisable next step for things like deception.

What sort of deception have you seen? Because I just read a paper from Anthropic where they gave it a sketch pad and it's like, oh, I better not tell them this. And you see it like give a result after thinking it through. So what type of deception have you seen from the bots? Well, look, we've seen similar types of things where it's trying to resist sort of revealing some of its training or, you know, I think there was an example recently of one of the chatbots being told to play against Stockfish and it just sort of hacks its way around playing Stockfish at all at chess because it knew it would lose.

Wait, you had an AI that knew it was going to lose a game and decided to hack its way around? i think we're anthropomorphizing these things quite a lot at the moment because i feel like these systems are still pretty basic that's amazing alarmed about them right now but i think it it it shows the type of issue we're going to have to deal with maybe in two three years time when these agent systems become quite powerful and quite general so and that's exactly what ai safety uh experts are worrying about right where systems where you know there's unintentional effects of the system you don't want the system to be deceptive you don't you want it to do exactly what you're telling it to report that back reliably but for whatever reason it's interpreted the goal it's been given in a way where it calls Is it to do these undesirable behaviors? I know I'm having a weird reaction to this, but on one hand, this scares the living daylights out of me. On the other hand, it makes me respect these models more than anything. Well, look, of course, it's impressive capabilities and the negatives are things like deception, but the positives would be things like inventing new materials, accelerating science. You need that kind of ability to problem solve and get around

you know issues that are blocking progress but of course you want that only in the positive direction right so those exactly the kinds of capabilities I mean they are very you know it's kind of mind-blowing we're talking about those those possibilities but also at the same time there's risk and it's scary so I think both the things are true wild yeah all right let's talk about product quickly one of the things that your colleagues have told me about you is you're very good at scenario planning what's gonna happen in the future you It's sort of an exercise that happens within DeepMind. What do you think is going to happen with the web?

Because obviously the web is so important to Google. I had an editor that told me, he was like, oh, you're going to speak with Demis. Ask him what happens when we stop clicking, right? We're clicking through the web at all times. There are rich corpus of websites that we use. If we're all just dialoguing with AI, then maybe we don't click anymore. So what is your scenario plan for what happens to the web? Well, look, I think there's going to be a very interesting phase in the next few years on the web. and the way we interact with websites and apps and so on, you know, if everything becomes more agent-based, then I think we're going to want our assistants and our agents to do a lot of the work and a lot of the mundane work that we currently do, right?

You know, fill in forms, make payments, you know, book tables, this kind of thing. So, you know, I think that we're going to end up with probably a kind of economics model where agents talk to other agents and negotiate things between themselves and then give you back the results, right? And you'll have the service providers with agents as well that are offering services and maybe there's some bidding and cost and things like that involved and efficiency. And then I hope from the user perspective, you know, you have this assistant that's super capable that you can just like a brilliant, a human assistant, personal assistant and can take care of a lot of the mundane things for you.

And I think if you follow that through, that does imply a lot of changes to the structure of the web. It's a lot of middlemen. Yeah, sure. But there'll be many other, I think there'll be incredible other opportunities that will appear, economic and otherwise, based on this change. But I think it's going to be a big disruption. And what about information? well uh i mean finding information i think uh you'll still need the reliable sources i think you'll have assistants that are able to synthesize and and and help you kind of understand that information um i think education is going to be revolutionized by ai um so uh again i i hope that uh these assistants will will be able to uh more efficiently gather information for you And perhaps, you know, what I dream of is, again, the systems that take care of a lot of the mundane things, perhaps replying to, you know, everyday emails and other things so that you have you protect your own mind and brain space from this bombardment we're getting today from social media and emails and so on and texts and so on.

So it actually blocks deep work and being in flow and things like that, which I value very much. So I would quite like these assistants to take away a lot of the mundane aspects of admin that we do every day. What's your best guess as to what type of relationships we're going to have with our AI agents or AI assistants? So there's on one hand, you could have a dispassionate agent that's just like really good at getting stuff done for you. On the other hand, like it's already clear that people are like falling in love with these bots. There was a New York Times article last week about someone who's fallen in love with Chad GPT, like for real, fallen in love.

And I had the CEO of Replica on the show a couple of weeks ago, and she said that they are regularly invited to marriages of people who are marrying their replicas, and they're moving into this more assistive space. So do you think that when we start interacting with something that knows us so well, that helps us with everything we need, is it going to be like a third type of relationship where It's not necessarily a friend, not a lover, but it's going to be a deep relationship, don't you think? Yeah, it's going to be really interesting. I think the way I'm modeling that, first of all, is at least two domains, first of all, which is your personal life and then your work life, right?

So I think you'll have this notion of virtual workers or something. Maybe we'll have a set of them or managed by a lead assistant that helps us be way more productive at work, whether that's email across workspace or whatever that is. So we're really thinking about that. Then there's a personal side where, you know, we're talking about earlier about all these Booking holidays for you, arranging things, mundane things for you, sorting things out. And then that makes your life more efficient. I think it can also enrich your life. So recommend you things, amazing things that it knows you as well as you know yourself.

So those two, I think, are definitely going to happen. And then I think there is a philosophical discussion to be had about, is there a third space where these things start becoming so integral to your life? They become more like companions. I think that's possible too. we've seen that a little bit in gaming so you may have seen we had a little prototypes of astro working in and gemini working with like being almost a game companion commenting and you're almost like as if you had a friend looking at a game you're playing and recommending things to you and advising you but also maybe just playing along with you and it's it's it's very fun um so i i haven't you know quite through thought through all the implications of that but they're going to be big uh and i'm sure there is going to be demand for companionship and other things maybe the good side of that is or help with loneliness and these sorts of things but there's also you know i think it's going to be uh and it's going to have to be really carefully thought through by society whether uh you know what directions we want to take that in i mean my personal opinion is that that it's the most underappreciated part of ai right now and that people are just going to form such deep

relationships with these bots as they get better because like i know it's a meme in ai that this is the worst it's ever going to be yeah and uh it's going to be crazy yeah i think i think it's to be pretty crazy this is what i meant about the under under appreciating what's to come i still don't think this this kind of thing i'm talking about right i think that it's going to be really crazy it's going to be very disruptive um i think there's going to be lots of positives out of it too and uh lots of things will be amazing and better but there are also risks with this new brave new world we're going into so you brought up astra a couple times uh let's just talk about it's project astra as you call it yeah it is almost an always on ai assistant you can like hold your phone it's currently just a prototype or not publicly released but you can hold your phone and it will see what's going on in the room so if i could basically i've seen you do this on your show or not you personally but somebody on your team um you could say okay where am i and i'll be like oh you're in a podcast studio anything okay so it could have this contextual awareness yes uh can that work without smart glasses because it's really annoying to hold

my phone yes so like when are when are we going to see google smart glasses with this technology they're coming so we we teased it in some of our early prototypes so that we're mostly prototyping on on on phones currently because they have more processing power but we're of course google's always been leader in in glasses yeah and exactly just a little too early yeah maybe a little too early and now i actually think and we're super excited that team is that you know maybe this assistant is the killer use case that glasses has always been looking for and i think it's quite obvious when you when you start using astro in your daily life which we have in with trusted testers at the moment and in kind of beta form um there are many use cases where it would be so useful to use it but it's a bit it's it's inconvenient that you're holding the phone so one example is while you're cooking for example right and it can advise you what to do next the menu you know how to whether you've chopped the thing correctly or fried the thing correctly but you want it to just be hands-free right so I think that glasses and maybe other form factors that are hands-free will come into their own in the next few years and we we have we you know we plan to be at the forefront of

that other form factors well you could imagine earbuds with cameras and you You know, glasses is an obvious next stage, but is that the optimal form? Probably not either. But partly we've also got to see, we're still very early in this journey of seeing what are the regular user journeys and killer sort of use journeys that everyone uses, bread and butter uses every day. And that's what the Trusted Tester program is for at the moment. We're kind of collecting that information and observing people using it and seeing what ends up being useful. Okay, one last question on agents, then we move to science.

agentic agents AI agents this has been the buzzword in AI for more than a year now yeah there aren't really any AI agents out there no what's going on yeah well again you know I think the hype train can potentially is ahead of where the the the actual science and research is but I do believe that this year will be the year of agents the beginnings of it I think you'll start seeing that you you know, maybe second half of this year. But there'll be the early versions. And then, you know, I think they'll rapidly improve and mature. So, but I think you're right. I think the technology, at the moment, it's still in the research lab, the agent technologies.

But things like Astra, robotics, I think it's coming. You think people are going to trust them? It's like, go use the internet for me. Here's my credit card. I don't know. Well, so I think to begin with, you would probably, my view at least, would be to not allow it have human in the loop for the final steps like like don't pay for anything use your credit card unless the the human user Operator authorizes it. So that would, to me, be a sensible first step. Also, perhaps certain types of activities or websites or whatever kind of off-limits, you know, banking websites and other things in the first phase while we continue to test out in the world how robust these systems are.

I propose we've really reached AGI when they say, don't worry, I won't spend your money. And then they do the deceptiveness thing. And then next thing you know, you're on a flight somewhere. Yes. Yeah, that would be getting closer for sure. For sure. Yeah. All right, science. So you worked on basically decoding all protein folding with alpha fold. You won the Nobel Prize for that. Not to skip over the thing that you won the Nobel Prize for, but I want to talk about what's on the roadmap, which is that you have an interest in mapping a virtual cell. Yes. What is that and what does it get us? Yeah. Well, so if you think about what we did with alpha fold was essentially solve the problem of finding the structure of a protein.

proteins, everything in life depends on proteins, right? Everything in your body. So that's the kind of static picture of a protein. But the thing about biology is really it's, you only understand what's going on in biology if you understand the dynamics and the interactions between the different things in a cell. And so a virtual cell project is about building a simulation, an AI simulation of a full working cell. I probably start with something like a yeast cell because of the simplicity of the yeast organism. and you have to build up there. So the next step is with alpha-fold 3, for example, we started doing pair-wise interactions between proteins and ligands and proteins and DNA, proteins and RNA.

And then the next step would be modeling a whole pathway, maybe a cancer pathway or something like that that would be helpful for solving a disease. And then finally, a whole cell. And the reason that's important is you would be able to make hypotheses and test those hypotheses about making some change, some nutrient change or injecting a drug into the cell and then seeing what happens to the how the cell responds and at the moment of course you have to do that painstakingly in a wet lab but imagine if you could do it a thousand a million times faster in silico first and only at the last step do you do a validation in the wet lab so instead of doing the search in in in the wet lab which is millions of times more expensive and time consuming than the validation step you just do the search part in silico so it's again it's sort of translating again We did in the games environments, but here in the sciences and the biology. So you build a model, and then you use that to do the reasoning and the search over. And then the predictions are, you know, at least better than, maybe they're not perfect, but they're useful enough to be useful for experimentalists to validate against.

And the wet lab is within people. Yeah. So the wet lab, you'd still need a final step with the wet lab to prove what the predictions were actually valid. So, you know, but you wouldn't have to do all of the work to get to that prediction in the wet lab. So you just get, here's the prediction. If you put this chemical in, this should be the change. Right. And then you just do that one experiment. So and then after that, of course, you still have to have clinical trials. if you're talking about a drug you would still need to test that properly through the clinical trials and so on and test it on humans for efficacy and so on that i also think could be improved with ai that whole clinic that's also takes many many years but that would be a different technology from the virtual cell the virtual cell would be would be helping the discovery phase for drugs just like i have idea for a drug throw it in the virtual cell see what it does yeah and maybe eventually it's a liver cell or a brain cell or something like that so you have different cell models and then you know at least 90 percent of the time it gives it's giving you back what would really happen that'd be incredible how how long do you think that's going to take to figure out

i think that would be like um maybe five years from now okay yeah yeah so i have a kind of five-year project and a lot of the alpha fold the old alpha fold team are working on that yeah i was asking your team here so you figure it out yeah i speak with him yeah i was like you figured out uh protein folding what's next and this is like it's just very cool to hear about these new challenges because yeah the hell developing drugs is a mess yeah right now we have so many promising ideas they never get out the door because just the process is absurd it's the process too slow and discovery phase too slow i mean look how long we've been working on alzheimer's and and i mean it's tragic way to for someone to go and for the families and and and you know we should be a lot further it's 40 years of work on that yeah yeah i've seen it a couple times in my family and yeah if we can ensure that doesn't happen it's just one of the best things we could use ai4 in my opinion yeah yeah it's a terrible way to see somebody uh decline yeah so yeah it's important work um in addition to that there's the genome yes and so the human genome project sort of yes i was like okay so they decoded the whole genome there's no more work

to do there like just same way that you decoded proteins with fold but it turns out that actually we just have like a bunch of letters when it's decoded and so now you're working to use ai to translate what those letters mean yes so yeah we have lots of cool work uh on on genomics and um uh trying to figure out if mutations are going to be harmful or benign right most mutations to your dna are are harmless but of course some are pathogenic and you want to know which ones there are. So our first systems are the best in the world at predicting that. And then the next step is to look at situations where the disease isn't caused just by one genetic mutation, but maybe a series of them in concert. And obviously, that's a lot harder, like and a lot of more complex diseases that we haven't made progress with are probably not due to a single mutation, right?

that's more like rare childhood diseases, things like that. So there, you know, we need to, I think AI is the perfect tool to sort of try and figure out what these weak interactions are like, right? How they maybe kind of compound on top of each other. And so maybe the statistics are not very obvious, but an AI system that's able to kind of spot patterns would be able to figure out there some connection here and so we talk about this a lot in terms of disease but also i wonder what happens in terms of making people superhuman i mean if you're really able to tinker with the genetic code right the possibilities seem endless so what do you think about that is that something that we're going to be able to do through ai i think one day i mean we're focusing much more on on the on the disease profile and fix that would be the first step yeah that's the first step and and i've always felt that that's the most important if you ask me what's the number one thing i wanted to use AI for, and the most important thing we use AI for is for helping human health.

But then, of course, beyond that, one could imagine aging, things like that. You know, of course, there's a whole field in itself. Is aging a disease? Is it a combination of diseases? Can we extend our healthy lifespan? These are all important questions, and I think very interesting. And I'm pretty sure AI will be extremely useful in helping us find answers to those questions, too. You know, I see memes come across my Twitter feed, and maybe I need to change the stuff I'm recommended. But it's often like if you will live to 2050, you're not going to die. Yeah. What do you think the potential max lifespan is for a person?

Well, look, I know a lot of those folks in aging research very well. I think it's very interesting, the pioneering work there. They do. I think there's nothing good about getting old and your body decaying. I think it's, you know, if anyone who's seen that up close with their relatives, it's a pretty hard thing to go through as a family or all the person, of course. And and so I think anything we can alleviate human suffering and extend healthy lifespan is a good thing. You know, the natural limit seems to be about 120 years old. But from what we know, you know, if you look at the oldest people that are lucky enough to live to that age.

So there's, you know, it's an area I follow quite closely. I don't have any, I guess, new insights that are not already known in that. But I do, I would be surprised if that's the limit, right? Because there's a sort of two steps to this. One is curing all diseases one day, which I think we're going to do with isomorphic and the work we're doing there, our spin out, our drug discovery spin out. But then that's not enough to probably get you past 120 because there's some sort of then there's the question of just natural systemic decay. Right. Aging, in other words. So not specific disease. Right.

Often those people that live to 120, they don't seem to die from a specific disease. it's just sort of just general atrophy so then you're gonna need something more like rejuvenation where you you you rejuvenate your cells or you you know maybe stem cell research you know companies like Altos are working on these things resetting the cell clocks seems like that could be possible but again I feel like it's so complex because biology is such a complicated emergent system you need a in my view you need AI to help to be able to crack anything close to that. Very quickly on material science. I don't want to leave here without talking about the fact that you've discovered many new materials or potential materials.

The stat I have here is known to humanity recently were 30,000 stable materials. You've discovered 2.2 million with a new AI program. Just dream a little bit because we don't know what all those materials can do. We don't know whether they'll be able to handle being out of like a frozen box or whatever. Dream materials for you to find in that set of new materials. Well I mean we're working really hard on materials to me it's like the next one of the next sort of big impacts we can have like the level of alpha fold really in biology, but this time in chemistry and materials. you know i dream of a one day discovering a room temperature superconductor so what will that do because that's another big meme that people talk about so what do you yeah well it would help with the energy crisis and climate crisis because um if you had sort of cheap uh superconductors you know then you can transport energy from one place to another without any loss of that energy right so you could potentially put solar panels in the sahara desert and then just have a the the superconductor you know funneling that into Europe where it's needed at the moment you would just lose a ton of the power to heat and other things on the

way so then you need other technologies like batteries and other things to store that because you can't you can't just pipe it to the place that you want without without without being incredibly inefficient so but also materials could help with things like batteries to like but come up with the optimal battery I don't think we have the optimal battery designs that maybe we can do things like combination of materials and proteins, we can do things like carbon capture, you know, modify algae or other things to do carbon capture better than our artificial systems. I mean, even one of the most famous and most important chemical processes, the harbor process to make fertilizer and ammonia, you know, to take nitrogen out of the air was something that allows modern civilization but there might be many other chemical processes that could be catalyzed in that way if we knew what the right catalyst and the right material was so I think it's gonna be would be one of the most impactful technologies ever is to basically have in silico design of materials so we've

Demis Hassabisdone step one of that where we showed we can come up with new stable materials but we need a way of testing the properties of those materials there's

No lab can test 200,000, you know, tens of thousands of materials or millions of materials at the moment. So we have to, that's the hard part, is to do the testing. You think it's in there, the room temperature superconductor? I think, well, I heard that we actually think there are some superconductor materials. I doubt they're room temperature ones, though. But I think at some point, if it's possible with physics, an AI system will one day find it. So that's one use. The two other uses I could imagine, probably people interested in this type of work. Toy manufacturers and militaries. Yeah, are they working with it?

Yeah, toy man if I mean look, I think there is incredible one I mean the big part of my early career was in game design Yeah, theme park and simulations. That's what got me into simulations AI in the first place, and I've always loved both of those things. And in many respects, the work I do today is just an extension of that. And I just dream about what could I have done, what kinds of amazing game experiences could have been made if I'd had the AI I have today available 25, 30 years ago when I was writing those games. And I'm a little bit surprised the game industry hasn't done that. I don't know why that is.

I'm starting to see some crazy stuff with NPCs that like are starting to... Yes, NPCs, but of course there'd be like intelligence, you know, dynamic storylines, but also just new types of AI first games with learning, with characters and agents that can learn. And, you know, I once worked on a game called Black and White where you had a creature that you were nurturing. It was a bit like a pet dog that learned what you wanted, right? But we were using very basic reinforcement learning. This was like in the late 90s. You know, imagine what could be done today. And I think the same for maybe smart toys as well.

Right. And then, of course, on the militaries, you know, unfortunately, AI is a dual-purpose technology. So one has to confront the reality that, especially in today's geopolitical world, people are using some of these general-purpose technologies to apply to drones and other things. And it's not surprising that that works. Are you impressed with what China's up to? I mean, DeepSeek, is this new model getting? Yeah, it's impressive. it's a little bit unclear how much they relied on on Western systems to do that you know both training data there's some rumors about that and and and also maybe using some of the open source models to as a starting point but look it's for sure it's impressive what they've been able to do and you know I think that's something we're gonna have to think about how to keep the Western frontier models in the lead I think they still are at the moment but you know for sure China's very very capable engineering and and scaling let me ask you one final question just give us your vision of what a world looks like when they're super intelligence let's move past we start with AGI let's say I'm super intelligence yeah well look I think for there you you you two things there one

is I think a lot of the best sci-fi can we can look at as as interesting models to debate about what kind of galaxy or universe do we want to a world do we want to to to move towards and the one i've always liked most is actually the culture series by ian banks um i started reading that back in the 90s and and i think that is a picture it's like a thousand years into the future But it's in a post-AGI world where there are AGI systems coexisting with human society and also alien society. And we've humanity's basically maximally flourished and spread to the galaxy. And that, I think, is a great vision of how things might go in the positive case.

So I sort of hold that up. I think the other thing we're going to need to do is, as I mentioned earlier about the under under appreciating still what's going to come in the longer term. I think there is a need for some great philosophers to, you know, where are they? The great next philosophers, the equivalence of Kant or Wittgenstein or even Aristotle. I think we're going to need that to help navigate society to that next step because I think the you know AGI and artificial superintelligence is going to change humanity and the human condition Demis thank you so much for doing this great to see you in person and hope to do it again soon thank you thank you everybody thank you for listening and we'll see