You can listen to the Google DeepMind CEO Demis Hassabis on AI, Creativity, and a Golden Age of Science | All-In Summit using Speak’s shareable media player:
Google DeepMind CEO Demis Hassabis on AI, Creativity, and a Golden Age of Science | All-In Summit Podcast Episode Description
(0:00) Introducing Sir Demis Hassabis, reflecting on his Nobel Prize win
(2:39) What is Google DeepMind? How does it interact with Google and Alphabet?
(4:01) Genie 3 world model
(9:21) State of robotics models, form factors, and more
(14:42) AI science breakthroughs, measuring AGI
(20:49) Nano-Banana and the future of creative tools, democratization of creativity
(24:44) Isomorphic Labs, probabilistic vs deterministic, scaling compute, a golden age of science
Thanks to our partners for making this happen!
Solana: https://solana.com/
OKX: https://www.okx.com/
Google Cloud: https://cloud.google.com/
IREN: https://iren.com/
Oracle: https://www.oracle.com/
Circle: https://www.circle.com/
BVNK: https://www.bvnk.com/
Follow Demis:
https://x.com/demishassabis
Follow the besties:
https://x.com/chamath
https://x.com/Jason
https://x.com/DavidSacks
https://x.com/friedberg
Follow on X:
https://x.com/theallinpod
Follow on Instagram:
https://www.instagram.com/theallinpod
Follow on TikTok:
@theallinpod
Follow on LinkedIn:
https://www.linkedin.com/company/allinpod
Intro Music Credit:
https://rb.gy/tppkzl
https://x.com/yung_spielburg
Intro Video Credit:
https://x.com/TheZachEffect
This interactive media player was created automatically by Speak. Want to generate intelligent media players yourself? Sign up for Speak!
Google DeepMind CEO Demis Hassabis on AI, Creativity, and a Golden Age of Science | All-In Summit Podcast Episode Top Keywords
Google DeepMind CEO Demis Hassabis on AI, Creativity, and a Golden Age of Science | All-In Summit Podcast Episode Summary
Based on the provided context, the phrase “has joined the group” refers to someone becoming a member of a group, band, club, or team. Throughout the conversation, there are multiple references to joining various groups, inviting members, and welcoming new people. Specific examples include:
Continue reading the full guide (click to expand)
– “we joined the band”
– “He should’ve joined the…”
– “Join the team.”
– “Welcome to the club.”
– “add one more bestie.”
– “they’re in, they’re in.”
– “invite you to…”
These statements all indicate the act of someone joining or being added to a group or collective. However, the context does not specify exactly who “has joined the group” in a particular instance. The general meaning is clear: it signifies the addition of a new member to a group. If you are looking for a specific individual who joined a specific group, that information is not explicitly provided in the context.
This summary was created automatically by Speak. Want to transcribe, analyze and summarize yourself? Sign up for Speak!
Google DeepMind CEO Demis Hassabis on AI, Creativity, and a Golden Age of Science | All-In Summit Podcast Episode Transcript (Unedited)
Speaker: 0
00:01
A genius who may hold the cards of our future. CEO of Google DeepMind, which is the engine of the company’s artificial intelligence. After his Nobel and a knighthood from King Charles, he became a pioneer of of artificial intelligence.
Speaker: 1
00:16
Want to run this on your own file?
Upload audio, video, or text and get a transcript, summary, and insights in minutes.
Free trial includes 30 minutes (60 with a work email)
We were the first ones to start doing it seriously in the modern era. AlphaGo was the big watershed moment. I think not just for DeepMind and my company, but for AI in general. This was always my aim with AI from a kid, which is to to use it to accelerate scientific discovery.
Speaker: 0
00:32
Ladies and gentlemen, please welcome Google DeepMind’s Demis Hassabis.
Speaker: 2
00:43
Welcome.
Speaker: 1
00:44
Great to be here.
Speaker: 2
00:44
Thanks for thanks for following Tucker, Mark Cuban, vatsal. First off, congrats on winning the Nobel Ai. Thank you. For the incredible breakthrough of AlphaFold, maybe, you may have done this before, but I know everyone here would love to hear your recounting of how you where you were when you won the Nobel Prize.
Speaker: 2
01:08
Ai you find out? Well, it
Speaker: 1
01:09
was a very surreal moment, obviously. You know, it’s Everything about it is surreal. The way they tell you, they tell you, like, ten minutes before, it all goes ai. It’s just, you know, you can’t really It’s You’re sort of shell shocked when you get that call from Sweden. It’s the call that every scientist dreams about.
Speaker: 1
01:24
And, and then the ceremonies the whole week in Sweden with the royal family. It’s amazing. Obviously, it’s been going for a hundred and twenty years. And the most amazing bit is they bring out the this Nobel book from the from the vaults in the sai, and you get to sign your name next to, you know, all the other greats.
Speaker: 1
01:41
So it’s quite an incredible moment sort of leafing back to the other pages and seeing Feynman and find Marie Curie and Einstein and Niels Bohr. And you just carry on going backwards, and you just get to put your name on that in that book. It’s incredible.
Speaker: 2
01:54
Did you have an inkling you’d been nominated and that this might be coming your way?
Speaker: 1
01:58
Well, you you you get you hear rumors. It’s amazingly locked down actually in in today’s age how they keep it so so quiet. But, it’s sort of like a a national treasure for Sweden. And, and so you hear, you know, maybe AlphaFold is the kind of thing that that would be, worthy of that recognition and it has they look for impact as well as the ai breakthrough, impact in the real world.
Speaker: 1
02:21
And that can take twenty, thirty years to to arrive. So, you just never know, you know, whether how soon it’s gonna be and and and whether it’s gonna be at all. So, it’s Amazing. Ai.
Speaker: 2
02:30
Well, congrats.
Speaker: 1
02:30
Yeah. Thank you.
Speaker: 2
02:31
And and thank you. You let me take a picture with it a few weeks ago, and we had it. So that’s something I’ll cherish. What is DeepMind within Alphabet? Alphabet is a sprawling organization, sprawling business units. What is DeepMind? What are you responsible for?
Speaker: 1
02:45
Well, we sort of see DeepMind now and Google DeepMind as it’s become. We sort of merged a couple of years back all of the different AI efforts across Google and Alphabet, including Ai. Put it all together, the ai of bringing the the strengths of all the different groups together into one division.
Speaker: 1
03:01
And, really, the way I describe the engine room of the whole of Google and the whole of Alphabet. So Gemini, our main model that we’re building, but also many of the other models that we also build, the the video models and interactive world models, we plug them in all across Google now.
Speaker: 1
03:18
So pretty much every product, every surface area has, one of our a AI models in it. So, you know, billions of people now interact with Gemini models, whether that’s through AI overview, AI mode, or the Gemini app. And that’s just the beginning. You know? We’re we’re kind of incorporating into Workspace, into Gmail, and so on.
Speaker: 1
03:37
So it’s a fantastic opportunity, really, for us to do cutting edge research, but then immediately ship it to billions of users.
Speaker: 2
03:45
And, how many people what’s the profile? Are these scientists, engineers? What’s the makeup of your organization?
Speaker: 1
03:50
There’s around 5,000 people in In your org. In my org, in in Google Ai. And and you know, it’s predominantly, I guess, 80% plus engineers and PhD researchers.
Speaker: 2
03:59
Right.
Speaker: 1
03:59
So, yeah, about, you know, three or three or 4,000.
Speaker: 2
04:02
So there’s an evolution of models, a lot of new models coming out and also new classes of models. The other day, you released this, Genie world model. Yes. So what is the Genie world model and, I think we got a video of it. Is it worth looking at and we can talk about it ai?
Speaker: 1
04:18
Yeah. We can watch it. Sure.
Speaker: 2
04:20
I think it you have to see it to understand it because it’s so extraordinary. Can we pull up the video? And then, Demis can narrate a little bit about what we’re looking at.
Speaker: 3
04:28
What you’re seeing are not games or videos. They’re worlds. Each one of these is an interactive environment generated by Genie three, a new frontier for world models. With Genie three, you can use natural language to generate a variety of worlds and explore them interactively, all with a single text prompt.
Speaker: 1
04:48
Meh. So all of these videos, all these interactive worlds that you’re seeing, so you’re seeing someone actually can control the video. It’s not a static video. It’s just being generated by a text prompt, and then people are able to control the three d environment using the arrow keys and the space bar.
Speaker: 1
05:02
So everything you’re seeing here is being fully all these pixels are being generated on the fly. They don’t exist until the player or the in the the person interacting with it goes to that part of the world. So, all of this richness, and then you’ll see in a second so this is fully generated. This is not a real real video.
Speaker: 1
05:21
This is generated someone painting their room, and they’re painting some stuff on the wall. And then the the player is sana look to the right, and then look back. So now this part of the world didn’t exist before, so now it exists. And then they look back, and they see the same painting marks they they left just earlier.
Speaker: 1
05:40
And, again, this is fully every pixel you can see is fully generated, and then you can type things like person in a chicken suit or meh ski, and it will just, in real time, include them in the scene. So, I think Ai know, it’s quite mind blowing really.
Speaker: 2
05:57
But I think what’s hard to grok when looking at this because we’ve all played video games that have a three d element to them when you’re in an immersive world. Meh. But there’s no objects that have been created. There’s no rendering engine. You’re not using Unity or Unreal, which are the three d rendering engines.
Speaker: 1
06:13
Yeah.
Speaker: 2
06:13
This is actually just two d images Yeah. That are being rendered, ai, created on the fly by the AI.
Speaker: 1
06:19
This model is reverse engineering intuitive physics. So, you know, it’s watched many millions of videos and YouTube videos and other things about the world. And just from that, it’s kinda reverse engineered how a lot of the world works. It’s not perfect yet, but it can generate, a consistent minute or two of, interaction as you as the user, in many, many different worlds.
Speaker: 1
06:42
There there are some videos later on where you can control, you know, a dog on a beach or a jellyfish or that’s not limited to just human things.
Speaker: 2
06:49
Because the way a three d rendering engine works is you type in the programmer programs all the laws of physics. How does light reflect off of an object. You create a three d object, light reflects off and then sai what I see visually is rendered by the software because it’s got all the programming on Yeah.
Speaker: 2
07:07
How to create physics, how to do physics. But this this was just trained off of video and it figured it all out.
Speaker: 1
07:13
Yeah. It was trained off of video and some synthetic data from from game engines, and it’s just reverse engineered it. And for me, it’s it’s it’s very close to my heart, this project, but it’s also quite mind blowing because in the nineties in my early career, I used to write, video games and AI for video games and graphics engines.
Speaker: 1
07:30
And I remember how hard it was to do this by hand, program all the polygons and the physics engines. And it’s amazing to just see this do it effortlessly. All of the reflections on the water and the the way materials flow and and and objects behave. And it’s just doing that all out of the box.
Speaker: 2
07:49
I think it’s hard to describe like how much complexity was solved for with that model. It’s it’s it’s really, really, really mind blowing. Where does this lead us? So fast forward this model
Speaker: 1
08:02
Yes. This gen five. Yeah. So so the reason we’re building these kind of models is, we feel and we’ve always felt, we’re obviously progressing on the normal language models ai with our Gemini model. But from the beginning with Gemini, we wanted it to be multimodal. So we wanted it to to input any take any kind of input, images, audio, video, and it can output anything.
Speaker: 1
08:22
And, and and so we’ve been very interested in this because you the for an AI to be truly general, to build AGI, we feel that the AGI system needs to understand the world around us and, the physical world around us, not just the abstract world of languages or mathematics. And, of course, that’s what’s critical for robotics to work. It’s probably what’s missing from it today.
Speaker: 1
08:42
And also things like smart glasses, a smart glasses system that helps you in your everyday life. It’s got to understand the physical context that you’re in and and how the world the intuitive physics of the world works. So we think that building these types of models, these Genie models and also Veo, our the best text to video models, those are expressions of us, building world models that understand the dynamics of the world, the physics of the world.
Speaker: 1
09:06
If you can generate it, then, that’s that’s an expression of your system understanding, those dynamics.
Speaker: 2
09:13
And that leads to a world of robotics, ultimately. One one one aspect, one application, but maybe we can talk about that. What is the state of the art with the vision, language, action models today? Sana generalized system, a box, a machine Yeah. That can observe the world with a camera and then I can use language, I can use text or speech to tell it Ai want you to do it and then it knows how to act physically to do something in the physical world
Speaker: 1
09:40
for me. Yeah. That’s right. So if you if you look at our, Ai Gemini live version of of Gemini where you can hold up your phone to the world around you. Ai recommend any of you try it. It’s kind of magical what it already understands about the physical world. You can think of the next step as as incorporating that in some sort of more handy device like glasses, and then it will be an everyday assistant.
Speaker: 1
10:02
It’ll be able to recommend things to you, as you’re walking the street, so we can embed it into Google Maps. And then with robotics, we’ve we’ve built, something called Gemini robotics models, which are sort of fine tuned Gemini with extra robotics data. And what’s really cool about that is and and we released some demos of this over the summer was, you can have you know, we’ve got these tabletop setups of two hands, interacting with objects on a table, two robotic hands.
Speaker: 1
10:29
And you can just talk to the robot. So you can say, you know, put the the yellow object into the red bucket or whatever it is, and, it will just it will it will interpret that instruction, that language instruction, into motor movements. And that’s the power of a multimodal model rather than just a robotic specific model
Speaker: 2
10:46
Right.
Speaker: 1
10:46
Is that it will be able to bring in real world understanding to the way you interact with it. So in the end, it’ll be the UI, UX that you you need for, as well as the understanding the robotic the robots need to to navigate the world safely.
Speaker: 2
10:59
I asked Sundar this. Does that mean that ultimately you could build what would be the equivalent of call it either a UNIX ai an operating system layer, or like an Android for generalized robotics, at which point, if it works well enough across enough devices, there will be a proliferation of robotics, devices and and companies and product that will suddenly take off in the world Meh.
Speaker: 2
11:23
Because this software exists to do this generally.
Speaker: 1
11:26
Exactly. That’s certainly one strategy we’re pursuing is a is a kind of Android play, if you ai, a crossroad as a kind of robotics, almost an OS layer, cross robotics. But there’s also some quite interesting things about vertically integrating our latest models, with specific robot, ai and robot designs and some kind of end to end learning of that too.
Speaker: 1
11:47
So both are actually pretty interesting and we’re pursuing, both strategies.
Speaker: 2
11:51
Do you think that there’s humanoid robots as a good kind of, form factor? Is that does that make sense in the world? Yeah. Because some folks have criticized it as being good for humans because we’re meant to do lots of different things. But if we sana solve a problem, there may be a different form factor to fold laundry or do dishes or clean the house or whatever.
Speaker: 1
12:08
Yeah. I think I think there’s gonna be a place for both. So so, ai, I used to be of the opinion maybe five five, ten years ago that we’ll have form specific robots for certain tasks. And I think in industry, industrial robots will definitely be like that where you can optimize the robots for the specific task, whether it’s a laboratory or a production line.
Speaker: 1
12:27
You’d want quite different types of robots. On the other hand, for, general use or personal use robotics, and just interacting with the the ordinary world, the humanoid form factor could be pretty important because, of course, we’ve designed the physical world around us, to be for for humans.
Speaker: 1
12:45
And so steps, doorways, all the things that we’ve designed for ourselves, rather than changing all of those in the real world, it might be easier to design the form factor to work seamlessly, with the way we’ve already designed the world. So I think there’s an argument to be made that the humanoid form factor could be very important for for those types of tasks.
Speaker: 1
13:05
But I think there is a place also for ai, robotic forms. Do you
Speaker: 2
13:08
have a view on hundreds of millions, millions, thousands over the next five years, seven years? I mean, do you have a ai, in your head, do you have a vision on Yeah.
Speaker: 1
13:15
I Ai do. And I I speak quite a lot of time on this. And I think we’re we’re still I I feel we’re still a little bit early on robotics. I think in the next couple of years, there’ll be a sort of real wow moment with robotics. But, I think the algorithms need a bit more development.
Speaker: 1
13:29
The general purpose, models that these these these robotics models are built on still need to be better and more reliable, and and better understanding the world around it. Ai think that will come in the next couple of years. And then also on the on the hardware side, the key is, I think, eventually, we will have millions of robots, helping helping helping society and and increasing productivity.
Speaker: 1
13:53
But the key there is when you talk to hardware experts is, at what point, do you have the right level of hardware to go for the scaling, option? Because, effectively, when you start building factories around trying to make tens of thousands, hundreds of thousands of particular robot type, you know, it’s harder for you to update, quickly iterate the the robot design.
Speaker: 1
14:13
Sai it’s one of those ai of questions where if you call it too early, then then then the next generation of robot might be invented in six months’ time that’s just more reliable and better and and more dexterous.
Speaker: 2
14:24
Sounds like using a computing analogy, we’re kind of in the seventies era PC DOS kind of,
Speaker: 1
14:28
Yeah. Potentially. But, of course, I think the the the the maybe that’s where we are, but I I think the except that ten years happens in one year ai. Right. So you’re gonna have four ai be
Speaker: 2
14:38
one of those years.
Speaker: 1
14:39
Ai. Exactly.
Speaker: 2
14:40
Yeah. So let’s talk about other applications, particularly in in sai, true to your heart as Yes. As a as a scientist, as the Nobel Ai winning scientist. I always felt like the greatest thing things that we would be able to do with Ai would be the problems that are intractable to humans with our current technology and capabilities and our brains and whatnot and we can unlock all of this potential.
Speaker: 2
15:07
What are the areas of science and breakthroughs in science that you’re most excited about and what kinds of models do we use to get there?
Speaker: 1
15:15
Yeah. I mean, Sai to accelerate scientific discovery and and help with things like human health is the reason I spent my whole career on AI. And I think, it’s the most important thing we can do with Ai. And I feel like if we build AGI in the right way, it’ll be the ultimate tool for science.
Speaker: 1
15:31
And I think we’ve been showing at DeepMind a lot of the way of that, obviously, AlphaFold, most famously. But, actually, we’ve we’ve, applied, our AI systems to many branches of science, whether it’s material design, helping with controlling plasma and fusion reactors, predicting the weather, solving, you know, mass Olympiad, math problems.
Speaker: 1
15:53
And, the same types of systems, with some, extra fine tuning can basically, solve a lot of these complex problems. So I think we’re just scratching the surface of what AI will be able to do, and there are some things that are missing. So, AI today, I would say, doesn’t have true creativity in the sense that it can’t come up with a new conjecture yet or a new hypothesis.
Speaker: 1
16:15
It can maybe prove something, that you give it, but it’s not able to come up with a sort of new idea or new theory itself. So, I think that would be one of the tests actually for AGI.
Speaker: 2
16:26
What is that? Creativity as a human. Meh. What is creativity then?
Speaker: 1
16:29
Well, I think it’s this sort of intuitive leaps that we often celebrate with the best scientists in history and and and artists, of course. And, you know, maybe it’s done through analogy or analogical reasoning. There are many theories in psychology and neuroscience and as to how, we as human scientists do it.
Speaker: 1
16:46
But a good test for it would be something like, give, one of these modern AI systems a knowledge cutoff of nineteen o one and see if it can come up with special relativity like Einstein did in nineteen o five. Right? If it’s able to do that, then I think, we’re onto something really, really important where perhaps we’re nearing an AGI.
Speaker: 1
17:06
Another example would be with our, AlphaGo program that beat the world champion at Go. Not only did it win in, you know, back ten years ago, it it invented new strategies that had never been seen before, for the game of Go, this famously move 37 in game two two that is now studied.
Speaker: 1
17:23
But can an AI system come up with a game as elegant, as satisfying, as aesthetically beautiful as Go, not just a new strategy? And the answer to those things at the moment is no. So that’s one of the things Ai think that’s missing, from, a true general system, an AGI system, is it should be able to do, those kinds of things as well.
Speaker: 2
17:42
Can you break down what’s missing and maybe relate it to the point of view shared by Dario, Sam, others about AGIs a few years away? Yeah. Do you not subscribe to that belief? And maybe help us understand what is it in your understanding of structure, in your understanding of the system architecture, what what’s lacking?
Speaker: 1
18:03
Well, so Ai think the fundamental aspect of this is, can we mimic these intuitive leaps rather than meh, advances that that the best human scientists seem to be able to do? I always say, like, what separates a great scientist from a good scientist is they’re both technically very capable, of course, but the great scientist is more creative.
Speaker: 1
18:23
And so maybe they’ll spot some pattern from another subject area that can be, can sort of have an analogy or some sort of pattern matching to the area they’re trying to solve. And I think one day Sai will be able to do this, but it doesn’t have the reasoning, capabilities and and some of the, thinking capabilities that, are gonna be needed to to make that kind of breakthrough.
Speaker: 1
18:46
I also think that we’re lacking consistency. So you often hear some of our competitors talk about, you know, these modern systems that we have today are PhD intelligences. I think that’s a nonsense. They’re not they’re not PhD intelligences. They have some capabilities that are PhD level, but they’re not in general, capable, and and that’s what exactly what general intelligence should be of of performing across the board at the PhD level.
Speaker: 1
19:12
In fact, as we all know, interacting with today’s chatbots, if you pose the question in a certain way, they can make simple mistakes with even, like, high school maths, and and simple counting. So, that shouldn’t be possible for a true AGI system. So I think that we are maybe, you know, I would sai, sort of five to ten years away, from having, an Ai system that’s capable of doing those things.
Speaker: 1
19:37
Another thing that’s missing is continual learning, this ability to, like, online teach the system something new, or or some or adjust its behavior in some way. And sai, a lot of these, I think, core capabilities are still missing. And maybe scaling will get us there, but I feel if I was to bet, I think there are probably one or two missing breakthroughs that are still required, and will come over the next, five five or so so years.
Speaker: 2
20:00
In the meantime, some of the reports and the the the scoring systems that are used seem to be demonstrating two things. One perhaps, and tell me if we’re wrong on this, a convergence of performance of large language models And number two, perhaps, is a slowing down or a flatlining of improvements in performance on each generation.
Speaker: 2
20:19
Are those two statements generally true or not so much?
Speaker: 1
20:22
No. I mean, we’re not we’re not seeing that internally and and, we’re still seeing a huge rate of progress. But also, we’re sort of looking at things more broadly. You see with our genie models and VO models and Nanobana.
Speaker: 2
20:35
Nanobana. Nanobana is in sync. It’s bananas.
Speaker: 1
20:38
Yes. It’s bananas. It’s bananas. Has
Speaker: 2
20:39
anyone here? Can can I see who’s used it? Has anyone used Nanobana? It’s incredible. Right? I mean Yeah. I’m I’m sai nerd who used to use Adobe Photoshop as a kid and Kai’s power tools and I was telling you Ai three d.
Speaker: 1
20:51
Yes. Still
Speaker: 2
20:52
like the graphic systems and like recognizing what’s going on there was just, ai, mind blowing.
Speaker: 1
20:57
Well, I think that’s the future of, a lot of these creative tools is you’re just gonna sort of vibe with it or just talk to them. And it’ll be consistent enough where, like, with Nanobanana, what’s amazing about it is that it’s an image generator. It’s best in best, you know, it’s state of the art and best in class. But it’s one of the things that makes it so great is the it’s consistency.
Speaker: 1
21:15
It’s able to under instruction follow what you want changed and keep everything else the same. And so you can iterate with it, and eventually get the kind of output that you want. And that’s, I think what the future of a lot of these creative tools is gonna be, and and sort of signals the direction. And people love it.
Speaker: 1
21:32
And and they love creating with it.
Speaker: 2
21:34
So democratization of creativity, I Ai think is really powerful. Ai having to buy books on Adobe Photoshop as a kid and then you’d read them to learn how to Right. Remove something from ram image and how to fill it in and feather and all this stuff. Now, anyone can do it with nano banana and just they can explain to the software what they want it to do, and it just does it.
Speaker: 1
21:53
Yeah. I think you’re gonna see two things, which is the, the sort of democratization of these tools for everybody to just use and and create with without having to learn, you know, incredibly complex UXs and UIs, like like we had to do in the past. But on the other hand, I think we’re and we’re also collaborating with filmmakers and top creators and artists.
Speaker: 1
22:14
So they’re helping us design what these new tools should be, what features would they want. People like, the director Darren Aronofsky, who’s a good friend of mine, ai director. And and he’s been making and his team will be making films using Veo and some of our other tools.
Speaker: 1
22:27
And we’re learning a lot by observing them and and collaborating them. And what we find is that it’s it also superpowers and turbocharges the best professionals too because they’re suddenly, the best creatives, the professional creatives, they’re suddenly able to be 10 x, 100 x more productive.
Speaker: 1
22:42
They can just try out all sorts of ideas they have in mind, you know, very low cost, and then get to the beautiful thing that they wanted. So I actually think it’s sort of both things are true. We’re we’re democratizing it for everyday use, for YouTube creators and so on. But on the other hand, at the high end, the people who, who understand these tools and it’s and it’s not everyone can get the same output out of these tools.
Speaker: 1
23:03
There’s a skill in that as well as, the vision and the storytelling and the narrative style of, the top creatives. And Ai think it just allows them they really enjoy, using these tools. It allows them to iterate way faster.
Speaker: 2
23:16
Do we get to a world where each individual describes what sort of content they’re interested in? Play meh music like Dave Matthews and it’ll play some new ram.
Speaker: 1
23:26
Yes.
Speaker: 2
23:27
Or I wanna play a video game set, you know, in the movie Braveheart and I wanna be in that movie. Yes. And I just have that experience. Do we end up there or do we still have a one to many creative process in society? How important culturally and I know this is a little bit philosophical, but it’s interesting to me which is, are we still gonna have storytelling where we have one story that we all share because someone made it?
Speaker: 2
23:48
Yeah. Or we each gonna start to develop and pull on our own kind of virtual I I actually foresee
Speaker: 1
23:53
a a world, and I think a lot about this having started in the games industry as a game designer and ram, is the, in the nineties is the, you know, I think the future of entertainment this is what we’re seeing is the beginning of the future of entertainment. Maybe some new genre or new art form. And where there’s a bit of cocreation, I still think that you’ll have the top creative visionaries.
Speaker: 1
24:12
They will be creating these compelling experiences and dynamic story lines, and they’ll be of higher quality even if they’re using the same tools than the everyday person can do. But also sana so millions of people will, potentially dive into those worlds, but maybe they’ll also be able to create, co create certain parts of those worlds and perhaps that, you know, the the the main creative, person is almost an editor of that world.
Speaker: 1
24:36
Right. So that’s the kind of things I’m foreseeing in the next few years, and I’d actually like to explore ourselves with with with with, technologies like Genie.
Speaker: 2
24:44
Right. Incredible. And how are you spending your time? Are you at ice maybe you can ai isomorphic. Of course. What isomorphic is, and are you spending a lot of your time there?
Speaker: 1
24:52
I am. So so I also run isomorphic, which is our spin out company, to revolutionize drug discovery, building on our alpha fold breakthrough in in protein folding. And, of course, knowing the structure of a protein is only one step in the drug discovery process. So isomorphic, you can think of it as building many, adjacent alpha folds to help with things like designing chemical compounds that don’t have any side effects, but bind to the right place on the protein.
Speaker: 1
25:20
And, I think we could reduce down drug discovery from taking years, sometimes a decade to do, down to maybe weeks or even days, over the next ten years.
Speaker: 2
25:30
That’s incredible. Do you think that’s in clinic soon or is that still in the discovery phase and
Speaker: 1
25:35
We’re building up the platform right now and it’s, we have great partnerships with Eli Lilly. I think you had, the CEO speaking earlier and and Novartis, which are fantastic, and our own internal drug programs. Ai I think we’ll be entering sort of preclinical phase, ai next year.
Speaker: 2
25:48
So candidates get handed over to the pharma company, and they then take them forward.
Speaker: 1
25:51
That’s right. And we’re working on cancers and immunology and oncology, and we’re working with, places like MD Anderson.
Speaker: 2
25:59
How much of this requires and I just sana go back to your point about AGI as it relates to what you just said. Models can be probabilistic or deterministic and tell me if I’m reducing this down too simplistically that the model takes an input and it outputs something very specific ai, it’s got a logical algorithm and it outputs the same thing every time and it could be probabilistic where it can change things and make selections.
Speaker: 2
26:21
The probability is 80% I’ll select this letter, 90% I’ll select this letter, next etcetera. How much do we have to kind of develop deterministic models that sync up with, for example, the the the physics or the chemistry underlying the molecular interactions as you do your drug discovery modeling?
Speaker: 2
26:39
How much are you building novel deterministic models that work with the models that are probabilistic trained on data?
Speaker: 1
26:45
Yeah. It’s a great question. Actually, we for the moment, and I think probably for the next five years or so, we’re building what maybe you could call hybrid models. So AlphaFold itself is a hybrid model where you have the learning component, this probabilistic component you’re talking about, which is, you know, based on neural networks and transformers and things.
Speaker: 1
27:02
And that’s learning from the data you give it, you know, any data you have available. But also to in a lot of cases with biology and chemistry, there isn’t enough data to learn from. So you also have to build in some of the rules about chemistry and physics that you already know about.
Speaker: 1
27:17
So for example, with AlphaFold, the angle of bonds between atoms. So and make sure that the the AlphaFold understood you couldn’t have atoms overlapping with each other and things like that. Now in theory, it could learn that, but it would waste a lot of the learning capacity.
Speaker: 1
27:31
So actually, it’s better to kind of have
Speaker: 2
27:33
that A constraint?
Speaker: 1
27:34
As a yeah. As a as a constraint in there. Now the trick is with all hybrid systems is and and AlphaGo was another hybrid system where there’s a neural network learning about the game of Go and what’s what kind of patterns are good. And then we had Monte Carlo Tree Search on top, which was doing the planning.
Speaker: 1
27:48
And so the trick is, how do you marry up a learning system with a a more handcrafted system, bespoke system, and actually have them work well together? And that’s, that’s pretty tricky to do.
Speaker: 2
28:00
Does that sort of architecture ultimately lead to the breakthroughs needed for AGI, do you think? Are there deterministic components that need to be ai?
Speaker: 1
28:07
Ultimately, sana you figure out something where this one of these hybrid systems, what you what you ultimately wanna do is upstream it into the learning component. So it’s always better if you can do end to end learning and and and directly predict the thing that you’re after from the data that you you’re you’re given.
Speaker: 1
28:23
So, so once you figured out something, using one of these hybrid systems, you then try and go back and reverse engineer what you’ve done and see if you can incorporate that learning, that that that that information into the learning system. And this is sort of what did with AlphaZero, the more general form of AlphaGo. So AlphaGo had some, Go specific knowledge in it.
Speaker: 1
28:44
But then with AlphaZero, we we got rid of that, including the human data, human games that we learned ram, and actually just did self learning from scratch. And of course, then it was able to learn any game, not just Go.
Speaker: 2
28:56
A lot of hype and hoopla has been made about the demand for energy arising from AI. This is a big part of the AI summit we held in Washington DC a few weeks ago. And it seems to be the number one topic everyone talks about in tech nowadays. Where’s all this power gonna come from?
Speaker: 2
29:10
But I ai the question of you, are there changes in the architecture of the models or the hardware or the relationship between the models and the hardware that brings down the energy per token of output or the cost per token of output vatsal ultimately, maybe sai, mutes the energy demand curve that’s in front of us or do you not think that that’s the case and we’re still gonna have a pretty kind of geometric energy demand curve?
Speaker: 1
29:35
Well, look, interestingly, again, I think both cases are true in the sense that, especially us at Google and at DeepMind, we we focus a lot on very efficient models, that are powerful because we have our own internal use cases, of course, where we need to serve, say, AI overviews to billions of users, every day. And it has to be extremely efficient, extremely low latency, and very cheap to serve.
Speaker: 1
29:56
And and so we’ve we’ve kind of pioneered meh, techniques that allow us to do that, like distillation where you sort of have a bigger model internally that trains the smaller model. Right? So you train the smaller model to mimic the bigger model. And over time, if you look at the progress of the last two years, the model efficiencies are are ai 10 x, you know, even a 100 x better, for the same performance.
Speaker: 1
30:16
Now the the reason that that isn’t reducing demand is because we’re still not got to AGI yet. So also the Frontier models, you keep wanting to train and experiment with, new ideas at larger and larger scale, whilst at the same tyler, at the serving side, things are getting more and more efficient.
Speaker: 1
30:33
So both things are true. And I and in in the end, I think that ram the energy perspective, Ai think AI systems will give back a lot more to energy, and climate change and these kind of things than they take in terms of efficiency of of of grid systems and electrical systems, material design, new types of properties, new energy sources.
Speaker: 1
30:54
I think AI, will help with all of that over the next ten years that will far outweigh, the energy that it uses today.
Speaker: 2
31:00
As the last question, describe the world ten years from now.
Speaker: 1
31:05
Wow. Okay. Well, I mean, you you know, ten years, even even ten weeks is is a is a hard time in AI. So, but
Speaker: 2
31:12
The ai field of ten years
Speaker: 1
31:14
for you. But Sai do feel ai if we will have AGI in the next ten years, you know, full AGI, and, I think that will usher in a new golden era of science. So a kind of new renaissance. And I think we’ll see the benefits of that right across from from energy to to human health.
Speaker: 2
31:32
Amazing. Please join me in thanking Nobel laureate, Denis.
Speaker: 1
31:36
Thank you.