The Exchanges

Q the basis for the next generation of these types of models? Their performance? Where does it asymptote? How do you think about scalability? How do you think about the underlying silicon that drives it? Is it a data issue? Is it a compute issue? Like, I'm just really interested in how you think about more broadly these really large scale models since you folks are building many of them now.

A Well, the, the incredible thing about where we've got to at this point is that all of the progress, in my opinion, is a function of compounding exponentials, right? So over the last decade, the amount of compute that we've used, uh, to train the largest models in the world has, um, increased by an order of magnitude every single year. So I went back and And, and had a look at the Atari DQN paper that we published in. Um, and that used just two petaflops, right? And some of the biggest models that we're training today, uh, at inflection, uh, use ten billion petaflops. So like nine orders of magnitude in nine years is like just insane. So I feel like it's super important to stay humble and acknowledge that. There is this epic wave of exponentials, which is unfolding around us, which is actually shaping the industry. And so when it comes to predictions, you have to just like, look at the exponential. It's pretty clear what's going on. That's just on the amount of compute side, the data side, everyone's super familiar with. We're using vast amounts of data and that's continuing. But I think the other thing that people don't always appreciate is that the models are also getting much more efficient. So, um, you know, one of the big breakthroughs of last year, which got some attention, but probably didn't quite get as much given how many breakthroughs there were was the chinchilla pap…

AI assessment note: “all of the progress, in my opinion, is a function of compounding exponentials”

Q the level of society. You know, we were thinking about it in the context of just like, how do you get more people to use this thing, you know? And so I think it's really interesting that people then later realize the big ramifications of this. In terms of, you know, how that actually cascades in terms of social behaviors and other things. How did that lead to starting DeepMind?

A Well, it was clear to me from that moment on, like I left Copenhagen in 2009 thinking this is not the path to significant positive social change. It still needs to continue. And I support those processes obviously, but I'm just saying it is just not something that I feel I could continue to work on full time. And so my heart was set on technology at that point. So I reached out to Demis who, uh, was the brother of my best friend from when I was a kid. We got together, we had a coffee, we went, and actually, we played poker, um, at, at one of the casinos in London, because we both love games, both super competitive, uh, both good at poker, and on that night, I think we both got knocked out pretty early in the tournament, so we sat around drinking Diet Coke, uh, talking about ways to change the world, and we basically were, you know, having exactly this conversation, like, You know, is it going to be, I mean, obviously at that point I was mostly inspired by platforms and software and social apps and connectivity and so on. Whereas, um, you know, Demis was way more in the kind of robotics land and sci-fi land. I mean, he, he was, he was fully thinking that, you know, the way to manage the economy, the way to make economic decisions was to simulate the entire economy. Right. And, and he thought that he was very much obviously had just come off the back of his games like evil genius…

AI assessment note: “So I reached out to Demis who, uh, was the brother of my best friend”

Q I guess last question on sort of your time with Google and DeepMind and because I think there's a lot of really exciting things to talk about in the context of inflection and sort of the broader field and world. What are some of the things you were most excited to have the team create at DeepMind over the years or some of the breakthroughs that you're most proud of?

A Yeah. Well, I mean, in some ways we, we, we definitely sort of pioneered the deep reinforcement learning effort. And I think, um, you know, in principle, it's a very promising direction. I mean, you clearly want some mechanism by which you can learn from raw perceptual data, and that directly feeds into a reinforcement learning algorithm that can update And essentially iterate on that in real time with respect to some reward function, whether that's online or offline, like directly interacting with the real world in real time, or it's, you know, in, in a kind of batch simulation mode, um, you know, and, and that turned out to be very valuable for a specific type of problem, um, where a game-like environment had a very structured scalar reward, and we could Play that game many millions of times. Um, that's part of the reason why we started the Alpha Fold project, because it was actually my group that was, um, looking around for other applications of DQN-like, AlphaGo-like, uh, tools, and, uh, in a hackathon, um, that we did one week, um, someone stumbled across, across this problem. We'd actually looked at it back in 2013 when it was called Foldit. Which was, uh, a very small scale kind of version of this.

AI assessment note: “we definitely sort of pioneered the deep reinforcement learning effort”

Q Is that the working definition you use for intelligence today?

A Um, actually, no. I've changed. Um, I, I think that there's a more nuanced version of that. I think that's a good definition of intelligence, but I think in a weird way, it's over-rotated the entire field on one aspect of, of intelligence, which is generality, you know, and I think, um, OpenAI and, um, then subsequently Anthropic and others have taken up this default Sort of mantra that like it, all that matters is can a single agent do everything? You know, can it be multimodal? Can it do translation and speech generation recognition, et cetera, et cetera. I think there's another definition which is valuable, which is the ability to direct attention or processing power to the salient features of a, of, uh, an environment given some context, right? So, um, actually what you want is to be able to take your raw processing horsepower and direct it in the right way at the right time, because it may be that a certain Tone or style is more appropriate given a context. It may be that a certain expert model is more suitable, or it may be that you actually need to go and use a tool, right? And obviously we're starting to see this emerge. Um, and in fact, I think the key, and we can get into this obviously in a moment, but I, I think the key element that is going to really unlock this field is actually going to be the router. In the middle of a series of different systems which are speci…

AI assessment note: “actually, no. I've changed. Um, I, I think that there's a more nuanced”

Q Is that the working definition you use for intelligence today?

A Um, actually, no. I've changed. Um, I, I think that there's a more nuanced version of that. I think that's a good definition of intelligence, but I think in a weird way, it's over-rotated the entire field on one aspect of, of intelligence, which is generality, you know, and I think, um, OpenAI and, um, then subsequently Anthropic and others have taken up this default Sort of mantra that like it, all that matters is can a single agent do everything? You know, can it be multimodal? Can it do translation and speech generation recognition, et cetera, et cetera. I think there's another definition which is valuable, which is the ability to direct attention or processing power to the salient features of a, of, uh, an environment given some context, right? So, um, actually what you want is to be able to take your raw processing horsepower and direct it in the right way at the right time, because it may be that a certain Tone or style is more appropriate given a context. It may be that a certain expert model is more suitable, or it may be that you actually need to go and use a tool, right? And obviously we're starting to see this emerge. Um, and in fact, I think the key, and we can get into this obviously in a moment, but I, I think the key element that is going to really unlock this field is actually going to be the router. In the middle of a series of different systems which are speci…

AI assessment note: “Um, actually, no. I've changed. Um, I, I think that there's a more nuanced version”

Q it's super interesting, and I think we can talk more about that in the context of AI in a little bit. One other thing that you did is you also started a consultancy where you worked as a negotiator and facilitator, and I believe you worked with clients like the United Nations, the Dutch government, and others. Can you tell us a little bit more about that work as well?

A Yeah. I mean, I was always trying to figure out how to scale my impact and, you know, I quite quickly realized that delivering a sort of one-to-one service via a nonprofit was not going to scale a great deal, even though it had an amazing impact, um, you know, on a kind of human to human level. And so I was super interested in these like meta structures, like how does, you know, the UN actually influence, you know, behavior at, At the country level. Um, and you know, how could we run more efficient decision-making processes, um, where there's tension and disagreement? So we worked all over the world, actually in Israel, Palestine, and in, you know, um, in Cyprus between the Greeks and the Turks. Uh, my colleagues worked in South Africa, Columbia, Guatemala, and I think it really taught me that Learning to speak other people's social languages is actually an acquired skill. And you really can do it with a, with a little bit of attention to detail and some patience and care. It's kind of a superpower being able to deeply hear other people and make them feel heard such that they're better able to empathize with people that they disagree with. And that, that's been an important theme throughout my kind of career, something I've always been interested in. So I think, I think I Co-founded that, uh, and worked on it for, I think, three years, and soon realized the limitations of, like…

AI assessment note: “we worked all over the world, actually in Israel, Palestine, and in, you know”

Q Is that the working definition you use for intelligence today?

A Um, actually no. I've changed. Um, I, I think that there's a more nuanced version of that. I think that's a good definition of intelligence, but I think in a weird way, it's over rotated the entire field on one aspect of, of intelligence, which is generality, you know, and I think, um, open AI and, um, then subsequently anthropic and others have taken up this default sort of mantra that like it, all that matters is Can a single agent do everything? You know, can it be multimodal? Can it do translation and speech generation recognition, et cetera, et cetera. I think there's another definition which is valuable, which is the ability to direct attention or processing power to the salient features of a, of, uh, uh, an environment given some context, right? So, um, actually what you want is to be able to take your raw processing horsepower And direct it in the right way at the right time, because it may be that a certain tone or style is more appropriate given a context. It may be that a certain expert model is more suitable, or it may be that you actually need to go and use a tool, right? And obviously we're starting to see this emerge. Um, and in fact, I think the key, and we can get into this obviously in a moment, but I, I think the key element that is going to really unlock this field Is actually going to be the router in the middle of a series of different systems which are sp…

AI assessment note: “actually no. I've changed. Um, I, I think that there's a more nuanced version”

Q agents for representing different businesses or causes or political groups or the like. What do you think that means in terms of how the web exists and how it's structured? So to your point, the web is effectively really based on a lot of SEO and a lot of sort of Google is the access point. What happens to web pages or what happens to the structure of the internet?

A I think it's going to change fundamentally. I think that most computing is going to become a conversation, and a lot of that conversation is going to be facilitated by AIs of various kinds. So your pie is going to give you a summary of the news in the morning, right? It's going to help you keep learning about your favorite hobby, whether it's cactuses or, you know, like motorcycles, right? And So, you know, every couple days is going to send you new updates, new information in a summary snippet that really kind of suits exactly your reading style and your interests and your preference for consuming information. Whereas a website, you know, the traditional open internet just assumes that there's a fixed format and that everybody wants a single format and generative AI clearly shows us that we can make this dynamic and emergent and entirely personalized. So. You know, if I was Google, I would be pretty worried because that old school system does not look like it's going to be where we're at in 10 years time. It's not going to happen overnight. There's going to be a transition, but these kind of succinct, dynamic, personalized, interactive moments are clearly the future in my opinion.

AI assessment note: “I think it's going to change fundamentally. I think that most computing is going to become a conversation”

Q Let's talk about inflection. What was the motivation for starting another company?

A Um, well, I guess back in sort of, 2018, 2019, it wasn't clear that neural networks were going to have a significant impact in language. If you just think about it intuitively, um, for the, for the previous sort of five years, CNNs had been effective at learning structure locally, right? So pixel, in an image, in the input. So pixels in an image that were correlated in space Tended to produce, you know, sub features which were, you know, a good representation of what you were trying to predict. Maybe there were lines and edges and they grew into eyes and faces and scenes and so on. And that kind of hierarchy just intuitively seemed to make sense and seemed to apply to audio and other modalities, right? Whereas if you kind of think about it, a lot of the structure of predicting the next word or Letter or token in a sentence seems to exist in a very, very, very spread out, you know, far removed from the immediate next step of the prediction. Right. And so it didn't look like that was working. And then to be honest, like when GPT-III came out, that was like a big revelation. Um, I, I had seen the GPT-II work. And hadn't quite clicked for me that this was significant. It was really only when I started, saw the GPT-III paper that my eyes were wide open to this possibility. It's pretty amazing that you could attend to, you know, a very, very sparse, seemingly sparse representation an…

AI assessment note: “when GPT-III came out, that was like a big revelation.”

Q Yeah. Yeah. Well, we can all work on it. Um, so you, you describe Pi as like the first foray that you guys can get out into the world and, um, learn from and improve with. What does improvement mean? Like, how do you, are you measuring emotional intelligence? What is better?

A Yeah, yeah. We're certainly measuring emotional intelligence. We're measuring the fluidity of the conversation. We're measuring, you know, how respectful it is. We're measuring how even-handed it is. Um, you know, we've already had a couple of errors where it's made some, um, politically biased remarks, and we try super hard to make sure that it's even-handed. No matter how, you know, sort of racist, homophobic, or misogynist in any way, it's, it should never be Dismissive, disrespectful, or judgmental of you. Um, it's there to talk through issues and make you feel heard and, um, take feedback. Like, it tries very hard to take feedback. So yeah, that's, that's, we're measuring all of those kinds of things, but, but the next phase of obviously where we're headed is that, um, we really think that this is going to be your ultimate personal digital assistant. And, um, it is going to, as I said, interact with other AIs to make decisions, buy your groceries, and, you You know, manage your sort of domestic life and help you book vacations and, you know, find, you know, fun information and that kind of stuff. So it's going to get, you know, increasingly more, uh, you know, down that route route. And, um, you know, the other thing is that quite soon it will, um, have the ability to access real time content in the web. So it'll be able to, you know, sort of look up, uh, the weather and n…

AI assessment note: “We're certainly measuring emotional intelligence. We're measuring the fluidity of the conversation.”