Demis Hassabis: Future of AI, Simulating Reality, Physics and Video Games | Lex Fridman Podcast #475

It's hard for us humans to make any kind of clean predictions about

highly nonlinear dynamical systems.

But again to your point, we might be very surprised what classical

learning systems might be able to do about even >> Yes, exactly.

I mean fluid dynamics, Navia Stokes equations, these are traditionally thought of as very

very difficult intractable problems to do on classical systems.

They take enormous amounts of compute,

you know, weather prediction systems,

you know, these kind of things all involve fluid dynamics calculations.

But again, if you look at something like

VO, our video generation model, it can model liquids quite well, surprisingly

well, and materials, specular lighting.

I love the ones where,

you know, there's there's people have generated videos where there's like clear liquids going

through hydraulic presses and then being squeezed out.

I I used to write

uh physics engines and graphics engines and in my early days in gaming.

And I know it's just so painstakingly

hard to build programs that can do that.

And yet somehow these systems are,

you know, reverse engineering

from just watching YouTube videos.

So presumably what's happening is

it's extracting some underlying

structure around how these materials behave.

So perhaps there is some kind of lower dimensional

manifold that can be learned if we actually fully understood

what's going on under the hood.

That's maybe, you know, maybe true of most of

The following is a conversation

with Demis Hassabis, his second time on the podcast.

He is the leader of Google Deep Mind

and is now a Nobel Prize winner.

Demis is one of the most brilliant

and fascinating minds in the world today,

working on understanding and building intelligence

and exploring the big mysteries of our universe.

This was truly an honor and a pleasure for me.

This is the Lex Freedman podcast.

To support it, please check out our sponsors in the description

and consider subscribing to this channel.

And now, dear friends,

here's Deus In your Nobel Prize lecture,

you propose what I think is a super interesting conjecture

that quote any pattern

that can be generated or found in nature can be efficiently

discovered and modeled by a classical learning algorithm.

What kind of patterns of systems

might be included in that?

Biology, chemistry, physics, maybe cosmology, >> neuroscience.

What what are we talking >> Sure.

Well, look, I I felt that it's sort of a tradition, I think, of

Nobel Prize lectures that you're supposed to be a little bit provocative

and I wanted to follow that tradition.

What I was talking about there is if you take a step back and

you look at um all the work that we've done especially with the alpha

x projects so I'm thinking alpho

of course alpha fold

what they really are is we're building models of very

combinatorily highdimensional spaces that you know if you tried to brute force a solution

find the best move and go

or find the the exact shape of a protein

and if you enumerated

all the possibilities you there wouldn't be enough time in the in the you

know the time of the universe.

So you have to do something much smarter

and what we did in both cases was

build models of those environments.

Um and that guided the search in a in a smart way and that makes it tractable.

So if you think about protein folding which is obviously a natural system

you know why should that be possible?

How does physics do that?

You know proteins fold in milliseconds in our bodies.

So somehow physics solves this problem that we've now also solved computationally.

And I think the reason that's possible is that in nature,

natural systems have structure

because they were subject to evolutionary

processes that that shape them.

And if that's true, then you can maybe learn

uh uh what that structure is.

So this perspective, I think, is really interesting one.

You've hinted it at it,

which is almost like uh crudely stated.

Anything that can be evolved

can be efficiently modeled.

Think there's some truth to that.

Yeah, I sometimes call it survival of the stablest

or something like that because

you know it's it's of course there's evolution for life

uh living things but there's also

you know if you think about geological

time so the shape of mountains

that's been shaped by weathering processes

right over thousands of years but then you can even take it cosmological

the orbits of planets

the um shapes of asteroids

these have all been survived

kind of processes that have acted on them many many times

so if that's true then there should some sort of pattern

um that you can kind of reverse learn

and uh a kind of manifold really that helps you

uh uh search to the right solution

to the right shape

um and actually allow you to predict things about it uh in an efficient

way because it's not a random pattern

right so um it may not be possible for for man-made

things or abstract things like factorizing

large numbers because unless there's patterns in the number space which there might be

but if there's not and it's uniform

then there's no pattern to learn there's no model to learn that will help you search.

So you have to do brute force.

So in that case you you know you maybe need a quantum computer something like this.

But in most things in nature that we're interested in

uh are not like that.

They have structure um that evolved for a reason and survived over time.

And if that's true I think that's potentially learnable by a neural network.

>> It's like nature is doing a search process

and it's so fascinating that it's in that search process is creating

systems that could be efficiently modeled. That's right. Yeah. >> So interesting.

>> So they can be efficiently

rediscovered or recovered um because nature is not random, right?

These everything that we see around us, including like the elements that are more

stable, all of those things, they're subject to

um some kind of selection process pressure.

Do you think because you're also a fan of theoretical

computer science and complexity,

do you think we can come up with a kind of complexity

class like a complexity

zoo type of class

where maybe it's the set of learnable systems,

the set of learnable natural systems, lns.

>> this is a deis

new class of systems that could be

actually learnable by classical systems in this kind of way.

Natural systems that can be uh modeled efficiently.

Yeah, I mean I' I've always been fascinated

by the P= MP question and what is

modelable by classical systems

I non-quantum systems you know cheuring machines in effect

and that's exactly what I'm working on actually in kind of my few moments

of spare time with a few colleagues

about is should there be you know maybe a new class of problem

that is solvable by this type of neural network process

and kind of mapped on to

these natural systems so you know the things that exist in and have structure.

So I think that could be a very interesting

uh new way of thinking about it.

And it sort of fits with the way I think about physics in general

which is that you know I think information is primary.

Information is the most sort of fundamental unit of the universe

more fundamental than energy and matter.

I think they can all be converted into each other but I think of

the universe as a kind of informationational system.

>> So when you think of the universe as anformational

system then the P=

NP question is a

is a physics question. >> That's right.

And it's a question that can help us actually

solve the entirety of this whole thing going on.

>> Yeah, I think it's one of the most uh fundamental

questions actually if you think of physics asformational

uh and and the answer to that I think is going to be you know very enlightening.

more specific to the PNNP question.

This again, some of the stuff we're saying is kind of crazy right now.

Just like the Christian

Edinson Nobel Prize speech controversial

thing that he said

sounded crazy and then you went and got a Nobel Prize for this with

John Jumper solved the problem.

So, let me let me just stick to the P equals MP.

Do you think there's something

in this thing we're talking about

that could be shown

if you can do

something like uh polomial

time or constant time compute ahead of time and construct this gigantic

model then you can solve some of these extremely

difficult problems in a theoretical

computer science kind of >> Yeah, I think that there are

actually a huge class of problems that could be couched in this way.

the way we did alpha go and the way we did alpha fold where

you know you you model

what the dynamics of the system is the the the the

properties of that system the environment

that you're trying to understand

and then that makes the search

for the solution or the prediction of the next step

efficient basically polomial time so tractable

by a uh classical

system uh which a neural network is it runs on normal

computers right classical computers

uh chewing machines in effect

and um I think it's one of the most interesting questions there is is

how far can that paradigm go?

You know, I think we've proven

uh and the AI community in general that classical

systems, cheuring machines can go a lot further than we previously thought.

You know, they can do things like model the structures of proteins

and play go to better than world champion level.

And uh you know a lot of people would have thought maybe 10 20

years ago that was decades away

or maybe you would need some sort of quantum machines to to quantum systems

to be able to do things like protein folding.

And so I think we haven't really

uh even sort of scratched the surface yet of what

uh classical systems socalled

uh uh could do.

And of course AGI

being built on a on a neural network system on top of a neural

network system on top of a classical computer

would be the ultimate expression of that.

And I think the limit

you know the the what what the bounds of that kind of system what

it can do it's a very interesting question and and and

directly speaks to the P equals MP question.

What do you think again hypothetical

might be outside of this

maybe emergent phenomena like if you look at cellular automa

some of the you have extremely simple systems and then some complexity

emerges yes >> maybe that would be outside or even

would you guess even that might be

>> to efficient modeling by a classical

>> yeah I think those systems would be right on the boundary

right so um I think most emergent systems cellular automter

things that could be modelable

by a classical system.

You just sort of do a forward simulation of it and it probably be efficient enough.

Um, of course there's the question of things like chaotic

systems where the initial conditions

really matter and then you get to some,

you know, uncorrelated end state

those could be difficult to model.

So I think these are kind of the open questions.

But I think when you step back and

look at what we've done

with the systems and the and the problems that we've solved

and then you look at things like V3

on like video generation

sort of rendering physics and

lighting and things like that, you know, really core fundamental things in physics.

Um it's pretty interesting.

I think it's telling us something quite fundamental about how the universe is structured in my opinion.

Um so you know in a way that's what I want to build AGI

for is to help

uh us uh as scientists answer these questions uh like p=mp.

>> Yeah I think we might be continuously

surprised about what is modelable by classical computers.

I mean alpha fold 3

on the interaction side is surprising

that you can make any kind of progress on that direction.

Alpha genome is surprising that you can map

the genetic code to the function

kind of playing with the emergent kind of phenomena.

You think there's so many combinatorial

options that and then here you go.

You can find the kernel that is efficiently >> Yes.

Because there's some structure there's some landscape

you know in the energy landscape or whatever it is that you can follow

some gradient you can follow.

And of course what neural networks are very good at is following gradients.

And so if there's one to follow and object and you can specify the

objective function correctly you know you don't have to

deal with all that complexity

which I think is how we maybe have naively

thought about it for decades

those problems if you just enumerate

all the possibilities it looks totally intractable

and there's many many problems like that and then you think well it's like

10^ the 300 possible protein

structures uh it's 10^ theund

you know 70 possible go positions

all of these are way more than atoms in the universe so how could

one possibly find the the right solution

or predict the next step

and and it but it turns out that it is possible and of course

reality nature does do it

right proteins do fold so that that gives you confidence that there must be

if we understood how physics was doing that

uh in a sense

uh then and we could mimic that process

I model that process

uh it should be possible

on our classical systems is is is

basically what the conjecture is about >> and of course there's nonlinear

dynamical systems, highly nonlinear dynamical

systems, everything involving >> You know, I recently had a conversation with Terrence Ta who

mathematically uh it contends

with a very difficult

aspect of systems that have some singularities

in them that break the mathematics

and it's just hard for us humans to make any kind of clean predictions

about highly nonlinear dynamical systems.

But again to your point we might be very surprised what classical

learning systems might be able to do about even fluid. Yes, exactly.

I mean fluid dynamics, Navia Stokes equations, these are traditionally thought of as very

very difficult intractable kind of problems to do on classical systems.

They take enormous amounts of compute,

you know, weather prediction systems,

you know, these kind of things all involve fluid dynamics calculations.

And um but again,

if you look at something like

VO, our video generation model, it can model liquids quite well, surprisingly

well, and materials, specular lighting.

I love the ones where, you know, there's there's people have generated videos where

there's like clear liquids going through hydraulic presses and then being squeezed out.

I I used to write

uh physics engines and graphics engines and in my early days in gaming and

I know it's just so painstakingly

hard to build programs that can do that and yet somehow these systems are,

you know, reverse engineering

from just watching YouTube videos.

So presumably what's happening is

it's extracting some underlying

structure around how these materials behave.

So perhaps there is some kind of lower dimensional

manifold that can be learned if we actually fully understood

what's going on under the hood.

That's maybe you know maybe true of most of reality. >> Yeah.

I've been continuously precisely

by this aspect of V3.

I think a lot of people highlight different aspects,

including the comedic and the meme and all that kind of stuff.

And then the ultra realistic

ability to capture humans in a really nice

way that's compelling and feels close to reality and then combine that with native audio.

All of those are marvelous things about V3.

But the exactly the thing you're mentioning,

which is the physics.

>> it's not perfect, but it's pretty damn good.

And then the really interesting scientific question is what is it understanding

about our world >> in order to be able to do that

because of the cynical

take with diffusion models

there's no way it understands

>> but it seem I mean I don't think you can generate that kind of

video without understanding and then our own philosophical

notion what it means to understand

then is like brought to the surface like do

to what degree do you think V3 understands our world?

I think to the extent that it can predict the next frames

you know in a coherent way

that's some that is a form you know of understanding

right not in the anthropomorphic

version of you know it's not some kind of deep

philosophical understanding of what's going on I don't think these systems have that

but they they certainly have

uh modeled enough of the dynamics

you know put it that way that they can

pretty accurately generate whatever it is 8 seconds of consistent

video that by eye at least you know at a glance is quite hard

to distinguish what the issues are and imagine that in two or three more years time.

That's the thing I'm thinking about and how incredible that there will look

uh given where we've come from, you know, the early versions of that

uh one or two years ago.

And so, um the rate of progress is incredible.

And I think um I'm like you is like a lot of people love

all of the the the the standup comedians and the the that actually captures

a lot of human dynamics very well and and body language,

but actually the thing I'm most impressed with and fascinated by is the physics

behavior, the lighting and materials and liquids.

And it's pretty amazing

that it can do that.

And I think that shows

that it has some

notion of at least intuitive physics, right?

um how things are supposed to work

uh intuitively maybe the way that uh a human child would understand physics

right as opposed to a you know a PhD

student really uh being able to unpack all the equations it's more of an

intuitive physics >> well that intuitive

physics understanding that's the base layer that's the thing people sometimes call like common

sense like it it really understands something I think that really surprised a lot

of people it blows my mind that

>> I just didn't think it would be possible to generate that level of realism without understanding.

>> You there's this notion

that you can only understand the physical world by having

an embodied AI system,

a robot that interacts with that world.

That's the only way to construct an understanding of that world.

But V3 is directly

challenging that it feels like >> yes and it's very interesting you know even if

we if you were to ask me 5 10 years ago

I would have said even though I was immersed in all of this I

would have said well yeah you probably need to understand intuitive

physics you know like if I push this off the table this glass it

will maybe shatter you know um and and the liquid will spill out right

so we know all of these things

but I thought that you know and there's a lot of theories in neuroscience

it's called action in perception

where you know you you need to act in the world to really truly

perceive it in a deep way.

And there was a lot of theories about you need embodied intelligence or robotics

or something or maybe at least simulated

action uh so that you would understand

things like intuitive physics.

But it seems like

um you can understand it through passive observation

which is pretty surprising to me and and again I think hints at something

underlying about the nature of

uh reality in in my opinion beyond

um just the you know the cool videos that it generates.

Um and and of course there's next stages is maybe even making those videos interactive.

So uh one can actually step into them and move around them.

Um which would be really mind-blowing

especially given my games background.

So you can imagine

and then and then I think you know you're we're starting to get towards

what I would call a world model

a model of how the world works the mechanics of the world the physics

of the world and the things in that world.

And of course that's what you would need for a true AGI system.

>> I have to talk to you about video games.

So, you were being a bit trolly.

I I think you're you're having more and more fun on Twitter on X,

which is great to see.

So, guy named Jimmy Apples tweeted,

"Let me play a video game of my V3 videos already.

Uh, Google cooked so good

playable world models when spelled

we n question mark."

Uh, and then you quote tweeted that with, "Now, wouldn't that be something?"

So, how how hard is it to build

game worlds with AI?

Maybe can you look out into the future

uh of video games

5 10 years out?

What do you think that looks like?

>> Well, games were my first love really and doing AI for games was the

first thing I did professionally

in my teenage years and and was the first

major AI systems that I built

and uh I always want to have I want to scratch that itch one

day and come back to that.

though, you know, and I will do, I think, and

um I think I'd sort of dream about, you know, what would I have

done back in the '90s if I'd had access to the kind of AI

systems we have today?

And I think you could build absolutely mind-blowing games.

Um, and I think the next stage is I always used to love making

all the games I've made are openw world games.

So, they're games where there's a simulation

and then there's AI characters

and then the player

uh interacts with that simulation

and the simulation adapts to the way the player plays.

And I always thought they were the coolest games because

uh so games like theme park that I worked on where everybody's

game experience would be unique to them, right?

Because you're kind of co-creating the game, right?

Uh we set up the parameters, we set up initial conditions,

and then you as the player immersed in it, and then you are co-creating

it with the with the simulation.

But of course, it's very hard to program open world games.

you know, you've got to be able to create

uh content whichever direction the player goes in and you want it to be

compelling no matter what the player chooses.

Um, and so it was always quite difficult to build

uh things like cellular automter

actually type of those kind of classical systems which created some emergent behavior.

Um, but they're always a little bit fragile, a little bit limited.

Now we're maybe on the cusp in the next few years, 5 10 years

of having AI systems that can truly

create around your imagination.

um can nar sort of dynamically

change the story and storytell

the narrative around uh and make it dramatic

no matter what you end up choosing.

So it's like the ultimate choose your own adventure sort of game.

And uh you know I think maybe we're within reach if you think of

a kind of interactive

version of VO uh and then

wind that forward 5 to 10 years and

you know imagine how good it's going to be. >> Yeah.

So you said a lot of super interesting stuff there.

So one the open

built into that is a deep personalization

the way you've described

>> So it's not just that it's open world like you can open any door

and there'll be something there.

It's that the choice of which door you open

>> in an unconstrained way

defines the worlds you see.

So some games try to do that to give you choice. Yes.

But it's really just an illusion of choice

>> the only uh like like Stanley Parable

game I recently played.

It's it's it's really there's a couple of doors and it really just takes

you down a narrative.

Stanley Parable is a great video game I recommend people play that kind of

uh in a meta way

uh mocks the illusion of choice and there's philosophical

notions of free will and so on.

But uh I do like one of my favorite games of Elder Scrolls is Daggerfall.

I believe that they really played

with a like random generation of the dungeons.

>> Of you can step in and they give you this feeling of an open

world and there you mentioned interactivity.

You don't need to interact.

That's a first step cuz you don't need to interact that much.

You just when you open the door,

whatever you see is

randomly generated for you. >> Yeah.

And that's already an incredible experience because you might be the only person to ever see that. >> Yeah. Exactly.

And and so but what you'd like is a little bit better than just

sort of a random generation, right?

So you'd like uh and and also better than a simple AB hardcoded choice, right?

That's not really uh open world, right?

As you say, it's just giving you the illusion of choice.

What you want to be able to do is is potentially anything in that game environment.

Um, and I think the only way you can do that is to have

uh generated systems, systems that

uh will generate that on the fly.

Of course, you can't create infinite amounts of game assets, right?

It's expensive enough already how

AAA games are made today.

And that was obvious to to us back in the '9s when I was

working on all these games.

I think maybe Black and White

uh was the game that I worked on, early stages of that that had

the still probably the best AI learning AI in it.

It was an early reinforcement

learning system that you,

you know, you were, you were looking after this mythical creature and growing it

and nurturing it and depending how you treated it, it would treat the villagers

in that world in the same way.

So if you were mean to it, it would be mean.

If you were good, it would be protective.

And so it was really a reflection of the way you played it.

So actually all of the

uh I've been working on sort of simulations

and AI through the medium of games at the beginning of my career and

and really the whole of what I do today is still a follow on

from uh those early

more hardcoded ways of doing the AI to now you know fully general learning

systems that that are trying to achieve the same thing.

>> Yeah, it's been uh interesting,

hilarious, and uh fun to watch you and Elon

obviously itching to create games because you're both gamers.

And one of the sad aspects

of your uh incredible success in so many domains of science

like serious adult stuff.

>> That you might not have time to really create a game.

You might end up creating the tooling

that others would create the game.

You have to watch

>> other others create the thing you've always dreamed of.

Do you think it's possible you can somehow in your extremely busy schedule actually

find time to create something like black and white?

some some an actual video game

where like you could

make the childhood dream come become reality.

>> You know, there's two things way to think about that is maybe

with vibe coding as it gets better and there's a possibility that I could,

you know, one could do that actually in in your spare time.

So, I'm quite excited about that as a as that would be my project

if if I got the time to do some vibe coding.

Um I'm actually itching to do that.

And then the other thing is, you know, maybe it's a sbatical

after agi has been safely

stewarded into the world and delivered into the world.

You know, that and then working on my physics theory

as we talked about at the beginning.

Those would be the two

my my two post AGI projects.

Let's call it that way.

>> I I would love to see which

game post AGI which you choose.

Solving uh the the problem that

some of the smartest people in human history contended with, you know, P equals

MP or creating a cool video. Yeah.

Well, but they might but in my world they'd be related because it would

be an openw world

simulated game uh as realistic as possible.

So, you know what what is what is the universe?

That's that's that's speaking to the same question, right? NPL MP.

I think all these things are related, at least in my mind.

I mean in a really serious way

like video games sometimes are looked down upon

as just this fun side activity

but especially as AI does more and more of

the difficult uh boring

tasks something we in in modern world call work.

You know video games

is the thing in which we may find meaning in which we may find

like what to do with our time.

You could create incredibly rich, meaningful experiences.

Like that's what human life is.

And then in video games, you can create

more sophisticated, more diverse

ways of living, >> I think so.

I mean, those of us who love games, and I still do, is is

is um you know, it's almost can let your imagination run wild, right?

Like I I used to love games

um and working on games.

so much because it's the fusion especially in the '9s and early 2000s the

sort of golden era maybe the 80s of of of

game of the games industry

and it was all being discovered new genres were being discovered

we weren't just making games we felt we were we were creating a new

entertainment medium that never existed before

especially with these open world games and simulation games where you were co-create you

as the player were co-creating

the story there's no other

media uh entertainment media where you do that where you as the audience actually

co-create the the story

and of course Now with multiplayer

games as well, it can be a very social

activity and can explore all kinds of interesting worlds in that.

But on the other hand,

you know, it's very important to

um also enjoy and experience

uh the physical world.

But the question is then, you know, I think we're going to have to

kind of confront the question again of what is the fundamental nature of reality?

uh what is the going to be the difference between these increasingly

realistic simulations and uh multiplayer

ones and emergent um and what we do in the real

>> Yeah, there's clearly a huge amount of value to experiencing

the real world nature.

There's also a huge amount of value in experiencing

other humans directly in person the way we're sitting here today.

>> But we need to really

scientifically rigorously answer the question why. >> Yeah.

And which aspect of that can be mapped

into the virtual world.

>> It's not it's not enough to say,

"Yeah, you should go touch grass and hang out in nature."

It's like, why exactly is that valuable? >> Yes.

And I guess that's maybe the thing that's

been uh haunting me, obsessing me from the beginning of my career.

If you think about all the different things I've done, that's they're all related in that way.

This simulation, nature of reality,

and what is the bounds of, you know, what can be modeled.

Sorry for the ridiculous question, but so far, what is the greatest video game of all time? What's up there?

>> Well, my favorite one of all time is Civilization.

I have to say that that was the the the Civilization

1 and Civilization 2.

My favorite games of all time.

Um >> I can only assume you've avoided the most recent one because

it would probably you would that would be your sobatical that you would disappear. >> Yes, exactly.

They take a lot of time these Civilization games.

So, I got to be careful with them. >> Fun question.

You and Elon seem to be somehow solid gamers.

Uh is there a connection between

being great at gaming and and

uh being great leaders of AI >> I don't know.

I It's an interesting one.

I mean uh we both love games

and uh it's interesting he wrote games as well to start off with.

It's probably especially in the era I grew up in where home computers were

just became a thing, you know, in the late ' 80s and '9s, especially in the UK.

I had a Spectrum and then a Commodore Omega

500 which is my my favorite computer ever and that's why I learned all

my programming and of course it's a very fun thing

uh to program is to program games.

So I think it's a great way to learn programming

probably still is and

um and then of course I immediately took it in directions of AI and

simulations which so I may was able to express my

interest in in games

and my sort of wider scientific interests alto together.

And then the final thing I think that's great about games is it fuses

um artistic design, you know, art

with the the the

most cutting edge programming.

Um so again, in the '90s,

all of the most interesting

uh technical advances were happening in gaming, whether that was AI,

graphics, physics engines, uh hardware, even GPUs of course were designed for gaming originally.

Um so everything that was pushing computing forward

in the in the '9s

was due to gaming.

So interestingly that was where the forefront of research was going on and it

was this incredible fusion with with art

um you know graphics

but also music and just the whole new media of storytelling

and I love that.

For me it's this sort of multi-disiplinary

kind of effort is again something I've enjoyed my whole my whole life.

I have to ask you, I almost forgot

about one of the many and

I would say one of the most incredible things recently

uh that somehow didn't yet get enough attention is alpha evolve.

>> We talked about evolution a little bit but it's the

Google deep mind system that evolves algorithms.

>> Are these kinds of evolution-like

techniques promising as a component of future super intelligence system?

So for people who don't know,

it's kind of um I don't know if it's fair to say it's LLM evolution search.

>> So evolutionary algorithms are doing the search

and LLMs are telling you where. >> Yes. Exactly.

So LLMs are kind of proposing

some possible solutions and then you do you use evolutionary

computing on top to to to

find some novel part of the of the search space.

So actually I think it's an example of

very promising directions where you combine

LLMs or foundation models

with other computational techniques.

Evolutionary methods is one but you could also imagine Monte Carlo research

basically many types of search algorithms

or reasoning algorithms sort of on top of or using

the foundation models as a basis.

So, I actually think there's quite a lot of interesting

uh things to be discovered probably with these sort of hybrid systems, let's call

>> But not to romanticize evolution.

Yeah, >> I'm only human.

But you think there's some value in whatever that mechanism is because we already

talked about natural systems.

Do you think where there's a lot of lowhanging

fruit of us understanding

being being able to model

uh being able to simulate evolution and then using that

whatever we understand about that

nature inspired mechanism to to then do surge better and better and >> Yes.

So if you think about uh again

breaking down the sort of systems we've built

uh to their really fundamental

core, you've got like the model of the of the underlying

dynamics of the system.

Uh and then if you want to discover something new, something novel that hasn't

been seen before, um then you need some kind of search process on top

to take you to a novel region of the of the of the search space.

And um you can do that in a number of ways.

Evolutionary computing is one.

um with Alph Go we just use Monte Carlo research

right and that's what found move 37

the new kind of never seen before

strategy in go and so that's how you can go beyond potentially what is

already known so the model can model everything that you currently know about right

all the data that you currently have but then how do you go beyond

that so that starts to speak about the ideas of creativity

how can these systems

create something new discover something new obviously this is super relevant for scientific

discovery or pushing met science and medicine forward, which we want to do with these systems.

And you can actually

bolt on some uh fairly

simple search systems on top of these models

and get you into a new region of space.

Of course, you also have to

um make sure that

you're not searching that space totally randomly.

It would be too big.

So, you have to have some objective function that you're trying to optimize and

hill climb towards and that guides that search.

But there's some mechanism

of evolution that are interesting

maybe in the space of programs.

But then the space of programs is an extremely important space because you can

probably generalize to to everything

you know for example mutation.

So it's not just

Monte Carlo tree search where it's like a search.

>> You could every once in a while >> combine things. Yeah.

>> Combine things alter like sub like a components of a thing. Yes.

So then you know what evolution is really good at is not just the natural selection.

It's combining things and building increasingly

complex hierarchical >> So that component is super interesting

especially like with alpha evolve in the space of programs. >> Yeah. Exactly.

So there's a you can get a bit of an extra property out of

evolutionary systems which is

some new emergent capability

may come about but of course like happened with life.

Interestingly, with naive uh sort of traditional evolutionary

computing methods without LLMs

and the modern AI,

the problem with them,

there was they were very well studied in the 90s and and and and early

2000s and some promising results, but the problem was they could never work out

how to evolve new properties, new emergent properties.

You always had a sort of subset of the properties that you put into the system.

But maybe if we combine them with these foundation models,

perhaps we can overcome that limitation.

Obviously uh natural evolution

clearly did because it it did evolve new capabilities

right so bacteria to where we are now.

So clearly that it must be possible

with evolutionary systems to generate

uh new patterns you know going back to the first thing we talked about

and uh new capabilities

and emergent properties and maybe we're on the cusp of discovering how to do that.

>> Yeah listen uh alpha evolve is one of the coolest things I've ever seen.

I've I've on my desk at home, you know, most of my time is

spent behind that computers just programming.

And next to the the three screens is a skull of a

tectalic, which is one of the early

organisms that crawled out of the water onto land.

And I just kind of watch that little guy.

It's like you whatever the computation

mechanism of evolution is is quite incredible.

It's truly truly incredible.

Now whether that's exactly the thing we need to do to do our search

but never dismiss the power of nature

what it did >> Yeah.

And it's amazing um

which is a relatively simple algorithm

right effectively and it can generate all of this immense complexity

emerges obviously running over

you know 4 billion years of time

but but it's it's it's

you know you can think about that as again a pro a search process

that ran over the physics substrate of the universe for

a long amount of computational

time but then it generated all this incredible uh rich diversity.

>> So uh so many questions I want to ask you.

But one, you do have a dream.

One of the natural systems you want to uh try to model is a is a cell.

>> that's a beautiful dream.

Uh I could ask you about that.

I also just for that purpose

on the AI scientist front just broadly.

So there's a essay

uh from Daniel Cocatalio,

Scott Alexander, and others that outlines steps along the way to get to ASI

and has a lot of interesting ideas in it.

one of which is

uh including a superhuman

coder and a superhuman

AI researcher and in that

there's a term of research

taste that's really interesting.

So in everything you've seen,

do you think it's possible for AI systems to have

research taste to help you in the way that AI co-scientist

does to help steer

human um human brilliant scientists

and then potentially by itself to figure out

what are the directions

where you want to generate truly novel ideas

because that seems to be like a

really important component how to do great science.

Yeah, I think that's going to be one of the hardest things to to

uh mimic or model is is this this idea of taste or or judgment.

I think that's what separates the

you know the the great scientists from the good scientists like all all professional

scientists are good technically right otherwise they wouldn't have made it

that far in in academia and things like that but then

do you have the taste to sort of sniff out what the right direction

is what the right experiment is what the right question is.

So the it's the it's picking the right question is is the hardest part of science.

Um and and making the right hypothesis

and um that's what you know today's systems definitely they can't do.

So you know I often say it's harder to come up with a conjecture

a really good conjecture than it is to solve it.

So we may have systems soon that can solve pretty hard conjectures.

um you know I I um mass Olympiad

problems where we we you know alpha proof last year our system got you

know silver medal in that really hard problems

maybe eventually we'll be able to solve a millennium prize kind of problem

but could a system have come up with a conjecture

worthy of study that someone like Terren Tower would have gone you know what that's

a really deep question

about the nature of maths or the nature of numbers or the nature of

physics and that is

far harder type of creativity

and we don't really Oh,

systems clearly can't do that and we're not quite sure what that mechanism would be.

This kind of leap of imagination

like like Einstein had when he came up with, you know, special relativity and

then general relativity with the knowledge he had at the time.

>> As for conjecture, the

you want to come up with a thing that's interesting

and amenable to >> So like it's easy to come up with a thing that's extremely difficult.

>> It's easy to come up with a thing that's extremely easy.

at that at that very >> that sweet spot, right, of of basically advancing the

science and splitting the hypothesis

space into two ideally, right?

Whether if it's true or not true, you you've learned something really useful

and um and and that's hard

and and and and

making something that's also

uh you know falsifiable

and within sort of the technologies

that you have you currently have available.

So it's a very creative

process actually highly creative process that

um I think just a kind of naive search on top of a model

won't be enough for that. >> Okay.

The idea of splitting the hypothesis

space in two is super interesting.

So uh I've heard you say that there's basically no failure

in or failure is extremely valuable

if it's done if you construct the questions right if you construct the experiments

right if you design them

right that failure success are both useful.

So perhaps because it splits the hypothesis

basically two, it's like a binary >> That's right.

So when you do like, you know, real blue sky research,

there's no such thing as failure really as long as you're picking experiments and

hypotheses that that that that

meaningfully spit the hypothesis space.

So you know, and you learn something, you can learn something kind of equally

valuable from an experiment that doesn't work.

That should tell you, if you've designed the experiment well and your hypothesis

are interesting, it should tell you a lot about where to go next.

and um and then it's you're effectively doing a search process

um and using that information

in in you know very helpful ways.

So to go to your dream

of uh modeling a cell

uh what are the big challenges that lay ahead for us to make that happen?

We should maybe highlight that

alpha I mean there's just so many leaps.

>> So AlphaFold solved if it's fair to say protein folding and there's so many

incredible things we could talk about there including the open sourcing

uh the everything you've released.

Alpha Fold 3 is doing protein, RNA, DNA interactions,

>> which is super complicated and and fascinating.

That's amendable to modeling.

Alpha genome uh predicts

uh how small genetic

changes like if we think about single mutations, how they link to actual uh function.

So um those are it seems like it's creeping along

to sophistic to to much more complicated

u things like a cell but a cell has a lot of really complicated components. >> Yeah.

So what I've tried to do throughout my career is I have these really

grand dreams and then I try to as you've noticed and then I try

to break but I try to break them down

any you know it's easy to have a kind of

a crazy ambitious dream

but the the the trick is how do you break it down into manageable

achievable uh interim steps

that are meaningful and useful in their own right

and so virtual cell which is what I call the project of modeling a

cell I've had this idea you know of wanting to do that for maybe

more like 25 is

and I used to talk with Paul Nurse

who is a bit of a mentor of mine in biology.

He runs the the you know founded the Craig Institute

and and won the Nobel Prize in in 2001.

uh is is we've been talking about it since

you know before the you know in the '90s

and um and I come used to come back to every 5 years is

like what would you need to model the full internals of a cell so

that you could do experiments

on the virtual cell

and what those experiment

you know in silicone

and those predictions would be useful for you to save you a lot of

time in the wet lab right that would be the dream maybe you could

100x speed up experiments

by doing most of it in silicone the search in silicico and then you

do the validation step in the wet lab.

That would be that's the that's the dream.

And so u but maybe now finally

uh so I was trying to build these components alpha fold being one

that that would allow you eventually

to model the full interaction

a full simulation of a cell and I'd probably start with a yeast cell

and partly that's what Paul nurse studied because a yeast cell is like a

full organism that's a single cell right so it's the kind of simplest

single cell organism and so it's not just a cell it's a full organism

and um and yeast is very well understood

And so that would be a good candidate for

uh a a kind of full simulated model.

Now alpha fold is the is the solution to the kind of static

picture of what does a what does a protein look 3D structure protein look

like a static picture of it.

But we know that biology

all the interesting things happen with the dynamics the interactions

and that's what alpha

3 is is the first step towards is modeling those interactions.

So first of all pairwise

you know proteins with proteins proteins with RNA and DNA

but then um the next step after that would be modeling maybe a whole

pathway maybe like the to pathway that's involved in cancer or something like this

and then eventually you might be able to model you know a whole cell

>> also there's another complexity here that

stuff in a cell happens at different time scales

is that tricky like there you know protein

uh folding is you know

super fast >> um I don't know all the bi ological mechanisms, but some of

them take a long time.

And so is that that's an level.

So the levels of interaction

has a different temporal scale that you have to be able to model.

>> So that would be hard.

So you'd probably need several

simulated systems that can interact at these different temporal dynamics

or at least maybe it's like a hierarchical system.

So um you can jump up and down the the different temporal stages.

So can you avoid I mean one of the challenges here is

not avoid simulating for example

the the the quantum mechanical aspects of any of this right you want to

not overm model you can skip ahead

to just model the

really highlevel things that get you a really good estimate of what's going to

>> so you you got to make a decision when you're modeling any natural system

what is the cutoff level of the granularity

that you're going to model it to that

then captures the dynamics that you're interested in.

So probably for a cell I would hope that would be the protein level

uh and that one wouldn't have to go down to the atomic level.

Um so you know of course that's where alpha volt stock kicks in.

So that would be kind of the basis

and then you'd build these

um uh higher level simulations

that um take those as building blocks

and then you get the emergent behavior.

Apologize for the pthead

questions ahead of time, but uh will

do you think uh we'll be able to

simulate and model the origin of life.

So being able to

simulate the first from from non-living

organisms the the birth of a living organism.

>> I think that's a one of the of course one of the deepest and most fascinating questions.

Um I love that area of biology.

you know, uh, people like there's a great book by Nick Lane, one of

the top top experts in this area called the the 10 great inventions of of of evolution.

I think it's fantastic and it also speaks to what the great filters might

be, you know, prior or are they ahead of us.

I think I think they're most likely in the past if you read that

book of how unlikely

to go, you know, have any life at all and then single cell to

multisell seems an unbelievably

big jump that took like a billion years, I think, on Earth to do, right?

So it shows you how hard it was, right?

>> Bacteria were super happy for a very long time, >> a very long time before

they captured mitochondria somehow, right?

I don't see why not

why AI couldn't help with that some kind of simulation.

Again, it's again, it's a bit of a search process through a combinatorial space.

Here's like all the chem, you know, the chemical soup

that that you start with, the primordial

soup that, you know, maybe was on Earth near these hot vents.

Here's some initial conditions.

Can you uh generate

something that looks like a cell?

So perhaps that would be a next stage after the virtual cell project is

well how how could you actually

um something like that emerge from the chemical soup?

>> Well, I would love it if there was a move 37

for the origin of life.

Yeah, >> I think that's one of the sort of great mysteries.

I think ultimately what we will figure out is their continuum.

There's no such thing as a line between non-living and living.

But if we can make that rigorous Yes.

>> that that the very thing from the be big bang

to today has been the same process.

If we can break down that wall that we've constructed in our minds of

the actual origin of

from non-living to living and it's not a line that it's a continuum

that connects physics and chemistry and biology. There's no line.

>> I mean this is my whole reason why I've worked on AI and AGI

my whole life because I think it can be the ultimate tool to help

us answer these kind of questions.

And I don't really understand why

um you know the average person doesn't

think like worry about this stuff more like how how

can we not have a good definition of life and not and not living

and non-living and the nature of time

and let alone consciousness

and gravity and all these things.

It's it's just and quantum mechanics weirdness.

It's just to me it's I've always had this sort of screaming

at me in my face the whole and that it's getting louder

you It's like how what is going on here?

You know, in in

I mean that in the deepest sense like in the you know the nature

of reality which has to be the ultimate question

uh that would answer all of these things.

It's sort of crazy if you think about we can stare at each other

and all these living things all the time.

We can inspect it with microscopes

and take it apart

uh almost down to the atomic level and yet we still can't answer that

clearly in a simple way that question of how do you define living?

>> it's kind of amazing.

Yeah, living you can kind of talk your way out of thinking about but

like consciousness like we have this very obviously

subjective conscious experience like we're at the center of our own world and it

it feels like something and then

h how how are you not screaming

>> at the mystery of it all

I mean but really humans have

been contending with the mystery of the world around them

for long there's a lot of mysteries

like what's up with the sun and and

the rain, >> like what's that about?

And then like last year we had a lot of rain and this year

we don't have rain.

Like what did we do wrong?

Humans have been asking that question for a long time. >> Exactly.

So we're quite I guess we've developed a lot of mechanisms to cope with

this these deep mysteries that we can't

fully we can see but we can't fully understand and we have to have

to just get on with daily life

and and and we get we keep ourselves busy right in a way.

Do we keep ourselves distracted?

I mean weather is one of the most important questions of human history.

We still that's that's the go-to

small talk direction of of the weather >> especially in England

>> and then it's which is

you know famously is an extremely difficult

system to model and

uh even that system

uh Google deep mind has made progress on.

Yes, we've yeah, we've created the the best weather prediction systems

in the world and they're better than traditional

fluid dynamics sort of systems that usually calculated

on massive supercomputers takes days to calculate it.

And we've managed to model a lot of the weather dynamics

with neural network systems

with our weather next system.

And again, it's interesting that those kinds of dynamics

can be modeled even though they're very complicated,

almost bordering on chaotic systems in some cases.

A lot of the interesting aspects of that

um can be modeled by these neural network systems, including

very recently we had, you know, cyclone prediction of where, you know, paths of hurricanes might go.

of course super useful

super important for the world

and and and it's super important to do that very timely and very quickly

and as well as accurately

and uh I think it's very promising direction

again of you know simulating

and uh uh so that you can run forward predictions and simulations

of very complicated real world systems.

>> I should mention that uh I've got a chance in uh Texas

to meet a community of folks called the stormchasers. >> Yes.

And what's really incredible about them, I need to talk to them more, is

they're extremely tech-savvy because what they have to do is they have to use

models to predict where the storm is.

So they're it's just it's it's this beautiful mix of like crazy

enough to like go into the eye of the storm

and >> in order to protect your life and predict where the extreme events are

going to be, they have to have increasingly

sophisticated models of of >> Yeah.

It's it's a a beautiful balance of like

being in it as living organisms

and the the cutting edge of science.

So they actually might be using

uh deep mind system.

So that's >> Yeah, they hopefully they are and I I'd love to join them

on one of those chases.

They look amazing, right?

To actually experience it one time. >> Exactly.

And then also to experience

the correct prediction where something will come

and how it's going to evolve.

It's >> You've estimated that we'll have AGI

by Um so there's interesting questions around that.

How will we actually know that we got there?

Uh and uh what

maybe the move quote move 37 of AGI.

>> My estimate is sort of 50%

chance by in the next 5 years.

So you know by 2030

let's say and uh so I think there's a good chance that that could happen.

Part of it is what what is your definition of AGI?

Of course, people are arguing about that now and and

uh mine's quite a high bar and always has been of like can we

match the cognitive functions that the brain has, right?

So, we know our brains are pretty much

general cheuring machines approximate.

And of course, we've created

incredible modern civilization with our minds.

So, that also speaks to how general the brain is.

And um for us to know we have a true AGI,

we would have to like make sure that it has all those capabilities.

it isn't kind of a jagged intelligence

where some things it's really good at like today's systems but other things it's

really uh flawed at

and and that's what we currently have with today's systems they're not consistent

so you'd want that consistency

of intelligence across the board

and then we have some missing

I think capabilities like sort of uh the true invention capabilities

and creativity that we were talking about earlier so you'd want to see those

how you test that

um I think you just test it one way to do it would be

a kind of brute force test of tens of thousand thousand of cognitive

tasks that um you know we know that humans can do

uh and maybe also

make the system available

to uh a few hundred of the world's top experts

uh the terren towers of each each subject area

and see if they can find you know give them give them a month

or two and see if they can find

an obvious flaw in the system and if they can't

then I think you're you're pretty

uh you know pretty you can be pretty confident we have a a fully

general >> maybe to push back a little bit it seems like humans are really

incredible as the the intelligence improves across all domains

to take it for granted.

>> Uh like you mentioned Terrence Tao

uh these brilliant experts

they might quickly in a span of weeks take for granted all the incredible

things it can do and then focus in well haha right there.

You know I I consider myself

uh first of all human.

>> Uh second I identify as human.

Um I you know some people listen to me talk and they're like that

guy is not good at talking the stuttering the you know

so like even humans

have obvious across domains

limits even just outside of

mathematics and physics and so on

it I I I

wonder if it will take something like a move 37

so on the positive side versus

>> a barrage of 10,000

cognitive tasks where it would be one

or two where it's like yes, holy

this is >> I think exactly.

So I think there's the sort of blanket testing to just make sure you've

got the consistency, but I think there are the

sort of lighthouse moments

like the move 37 that

I would be looking for.

So one would be

inventing a new conjecture

or new hypothesis about physics like Einstein did.

So maybe you could even run the back test of that very rigorously

like have a cut off of knowledge cutff of 1900

and then give the system everything that was you know that was written up

to 1900 and then and then see if it could come up with special

relativity and general relativity right like Einstein did that that would be an interesting

test another one would be can it invent a game

like go not just come up with move 37 a new strategy but can

it invent a game that's as deep as aesthetically

beautiful as elegant as go

and those are the sorts of things I would be looking out for.

Uh and probably a system being able to do uh uh several of those

things, right, for it to be very general.

Um not just one domain.

And so I think that would be the signs

at least that I would be looking for

that we've got a system that's a GI level.

And then maybe to fill that out, you would also check the consistency,

you know, make sure there's no holes

in that system >> Yeah.

Something like a new conjecture or scientific discovery.

That would be a cool feeling.

Yeah, that would be amazing.

So, it's not not just helping us do that, but actually coming up with

something brand new >> and you would be in the room for that.

So, it would be like

probably 2 or 3 months before announcing it.

>> And you would just be sitting there

trying not to tweet >> something like that. Exactly.

It's like what is this amazing new

you know physics idea?

And then we would probably check it with world experts in that domain, right?

and validate it and kind of go through its workings

and it I guess it would be explaining its workings too.

Um yeah be an amazing

>> Do you worry that we as humans

even expert humans like you might miss it

might miss >> it may be pretty complicated.

So it could be the analogy I give there is I don't think it

will be um uh uh totally

mysterious to the to the best human scientists but it may be a bit

like for example in chess if I was to talk to Gary Kasparov

or Magnus Carlson and play a game with them and they make a brilliant

move I might not be able to come up with that move but they

could explain why afterwards

that move made sense and we were to understand it to some degree

not to the level they do but in you know if they were good

at explaining which is actually part of intellg igence too is being able to

explain in a simple way

that what you're thinking about.

Um uh I I think that that would be very possible for the best human scientists.

>> But I wonder maybe you can you can educate me on the side of go.

I wonder if there's moves for Agnes or Gary where they at first

will dismiss it as a bad >> Yeah, sure. It could be.

But then afterwards they'll figure out with their intuition

that that this why this works.

And then and then and then empirically

the nice thing about games is one of the great things about games is

you can it's a sort of scientific test.

Does it do you win the game or not win?

And then um that tells you

okay that move in the end was good.

That strategy was good.

And then you can go back and analyze that and and and and

explain even to yourself a little bit more

why explore around it.

And that's how chess analysis

and things like that works.

So perhaps that's why my brain works like that cuz I I've been doing

that since I was four

and you're train you know it's sort of hardcore

training in that way.

But even even now like when I generate

code there there is this kind of nuanced

fascinating con contention that's happening

where I might at first identify

as a set of generated code is incorrect

in in some interesting nuanced ways

but then I'm always have to ask the question

is there a deeper insight here that

that I'm the one who's incorrect

>> and that's going to as the systems get more and more intelligent

you're going to have to contend with that.

It's like what what

what do you is this a bug or a feature of what you just came up with? >> Yeah.

And they're going to be pretty complicated to do.

But of course it will be you can imagine also AI systems that are

producing that code or whatever that is and then

human programmers looking at it but also not unaded

with the help of AI tools as well.

So it's going to be kind of an interesting you know maybe different AI

tools to the ones

that the more you know kind of monitoring tools to the ones that generated it.

So if we look at a AGI

system, sorry to bring it back up, but alpha super cool.

So Alpha Evolve enables

on the programming side

something like recursive self-improvement

uh potentially like what who can imagine

what that AGI system maybe not the first version but

a few versions beyond that.

What does that actually look like?

Do you think it would be simple?

You think it'll be something like a self-improving

program in a simple one?

>> I mean, potentially that's possible.

I would say um I'm not sure it's even desirable

because that's a kind of like hard takeoff scenario.

But but you you

these current systems like Alpha Evolve,

they have, you know, human in the loop deciding on various things.

They're separate hybrid systems that interact.

Uh one could imagine

eventually doing that end to end.

I don't see why that wouldn't be possible

but right now um you know I think the systems are not good enough

to do that in terms of coming up with the architecture of the code.

Um and again it's a little bit connected to this idea of coming up

with a new conjectural hypothesis.

How like they're good if you give them very specific instructions about what you're trying to do.

Um, but if you give them a very vague high level instruction,

that wouldn't work currently.

Like, uh, and I think that's related to this idea of like invent a

game as good as go, right?

Imagine that was the prompt.

That's that's pretty underspecified.

And so the current systems wouldn't know, I think, what to do with that,

how to narrow that down to something tractable.

And I think there's similar like, look, just make a better version of yourself

that's too that's too unconstrained.

But we've done it in, you know, and as you know with Alpha Evolve,

like things like faster matrix multiplication.

So when you when you hone it down to very specific thing you want

um it's very good at incrementally

improving that but at the moment these are more like incremental

improvements sort of small

iterations whereas if you know if you wanted a big leap

in uh understanding you need a you need a much larger uh advance. >> Yeah.

But it could also be sort of to push back against hard takeoff scenario.

It could be just

a sequence of um

incremental improvements like matrix multiplication

like it has to sit there for days

thinking how to incrementally

improve a thing and that it does so recursively

and as you do more and more improvement

it'll slow down so there'll be like a

like uh the path to AGI won't be like a

it'll be a gradual improvement over time. >> Yes.

If it was just incremental

improvements that's how it would look.

So the question is could it come up with

a new leap like the transformers

architecture right could it have done that back in 2017

when you know we did it and brain did it and it's it's not

clear that that these systems

something like Alpha wouldn't be able to do make such a big leap

so for sure these systems are good we have systems I think that can

do incremental hill climbing

and that's a kind of bigger question about is that all that's needed from

here or do we actually need one or two more

um uh big breakthroughs

>> and can the same kind of

systems provide the breakthroughs also.

So make it a bunch of scurves

like incremental improvement but also every once in a while leaps. >> Yeah.

I don't think anyone has systems that can have shown

unequivocally those big leaps

that the the right.

We have a lot of systems that do the hill climbing of the S-curve

that you're currently on. >> Yeah.

And that would be the move 37 is a >> Yeah.

I think would be a leap.

Um something like that.

Uh do you think the scaling laws are holding strong

on the pre-training, post- training, test time, compute?

Uh do you uh on the flip side of that anticipate

AI progress hitting a

>> We certainly feel there's a lot more room just in the scaling.

So um actually all steps pre-training,

post-training and inference time.

So uh there's sort of three scalings that are happening concurrently.

Um and we again there it's about how innovative

you can be and we you know we pride ourselves on having the broadest

and um deepest research bench.

uh we have amazing you know incredible

uh researchers and uh people like Nam Shazir who you know came up with

transformers and and Dave Silva

you know who led the Alph Go project and so on

and um it's it's it's

that research base means that if some new

new breakthrough is required

like an Alph Go or Transformers

uh I would back us to be the place that does that.

So I'm actually quite like it when the terrain gets harder, right?

Because then it veers more from just engineering

to to true research and you know re or research plus engineering and that's our sweet spot.

And I I think that's harder

it's harder to invent things than to than to

um you know fast follow.

And um so you know we don't know I would say it's a it's

kind of 50/50 whether

new things are needed or whether the scaling the existing stuff is going to be enough.

And so in true kind of empirical

fashion, we're pushing both of those as hard as possible.

The new blue sky ideas

and you know maybe about half our resources are on that and then and

then uh scaling to the max

the the current the current capabilities

and um we're still seeing some

you know fantastic progress on

uh each different version of Gemini.

That's interesting the way you put it in terms of the deep bench

that if uh progress towards AGI

is more than just

scaling compute so the engineering

side of the problem

and is more on the scientific

side where there's breakthroughs

needed then you feel confident deep mind as well

Google deep mind is well positioned to

kick kick ass in that domain

>> well I mean if you look at the history of the last decade or

15 years um it's been I you know maybe I don't know 80 90% of

the breakthroughs that that underpins modern AI field today was from you know originally

Google brain Google research and deep mind so

yeah I would back that to continue hopefully

>> uh so on the data side are you concerned about running out of highquality

data especially high quality human data

>> I'm not very worried about that partly because I think there's enough data

uh or and it's been proven to get the systems to be pretty good

and this goes back to simulations

again if you do you have enough data to make simulations

or so that you can create more synthetic

data that are from the right distribution.

Obviously, that's the key.

So, you need enough real world data in order to be able to

uh uh create those kinds of generator

data generators and um I think that we're at that step at the moment.

>> Yeah, you've done a lot of incredible stuff on the side of science and

biology doing a lot with not so much data.

>> I mean, it's still a lot of data, but I guess enough

>> take off that going. Exactly. Yeah.

>> So exactly >> uh how crucial is the scaling of compute to building AGI?

This is a question that's an engineering question.

It's a almost geopolitical

question >> because it also integrated

into that is the supply chains and energy

a thing that you care a lot about which is um potentially fusion.

So innovating on the side of energy also.

Do you think we're going to keep scaling compute?

>> I think so for several reasons.

I think compute there's there's the amount of compute you have for training

often it needs to be colloccated

so actually even like

you know uh bandwidth constraints between data centers can affect that

so it's it's it's

there's additional constraints even there

and that that's important for training obviously the largest models you can

but there's also because now

AI systems are in products and being used by billions of people around the

world you need a ton of inference compute now

um and then on top of that there's the thinking

systems, the new paradigm

uh of the last year that

uh where they get smarter the longer amount of inference time you give them at test time.

So all of those things

need a lot of compute

and I don't really see that slowing down.

Um and as AI systems become better, they'll become more useful and there'll be

more demand for them.

So both from the training side, the training side actually is is only just

one part of that.

It may even become the smaller part

of of what's needed

um uh in the overall compute that that's required.

Yeah, that's one sort of almost memey

kind of thing which is like the success and the incredible aspects of V3

there people kind of make fun of like the more successful it becomes the

you know the servers are

>> Yes, exactly the difference in >> Yeah. Yeah. Exactly.

We did a little video of of

the servers frying eggs and things and

um that's right and and and we're going to have to figure out how to do that.

Um there's a lot of interesting hardware innovations that we do as you know

we have our own TPU line and we're looking at like inference

only things inference only chips and how we can make those more efficient.

We're also very interested in building AI systems and we have done the help

with energy usage so help

um data center energy like for the cooling systems be efficient

um grid um and then eventually things like helping with

plasma containment fusion reactors.

We've done lots of work on that with Commonwealth

Fusion and also uh one could imagine reactor design.

Um and then material design I think is one of the most exciting

new types of solar material solar panel material

super room temperature superconductors

has always been on my list of dream breakthroughs

and um optimal batteries

and I think a solution to any you know one of those things would

be absolutely revolutionary for

you know climate and

energy usage and we're probably close

you know again in the next 5 years to having AI systems that can

materially help with those problems.

If you were to bet, sorry for the ridiculous question, but what what is

the main source of energy

in like 20, 30, 40 years, do you think it's going to be nuclear fusion?

>> I think fusion and solar are the two that I I would bet on.

Um solar, I mean, you know, it's the fusion reactor in the sky, of

course, and I think really

the problem there is is is batteries and transmission.

So you know as well as more efficient more and more efficient solar material

perhaps eventually you know in space

you know these kind of Dyson sphere type ideas

and fusion I think is

definitely doable seems uh if we have the right design

of reactor and we can control the plasma and uh fast enough and so

on and I think

both of those things will actually get solved

so we'll probably have at least those will probably be the two primary

sources of renewable clean

almost free or perhaps free energy

What a time to be alive.

If I uh traveled into the future with you

100 years from now,

how much would you be surprised if we've passed

a type one card scale civilization?

I would not be that surprised if there was a like a 100redyear time scale from here.

I mean, I think it's pretty

clear if we crack the energy problems in one of the ways we've just

discussed, fusion or or very efficient solar,

um, then if energy is kind of free and renewable and clean,

um, then that solves a whole bunch of other problems.

So, for example, the water access problem

goes away because you can just use desalination.