So there's a big ceremony today. Nobel Prize — I mean, two people at DeepMind, your founder and the chief scientist, got the Nobel Prize. And you. And you just flew in from basically from Stockholm. I mean, is it life changing?
Oh, it was surreal. It's such an exciting moment not only for Demis and John, but I think for the team that worked on it. I definitely feel excited about that in the broader organization. And actually, in many ways, for science, you know, this is the first time in Nobel history that the committee has awarded an organization for the application of AI. And I think the fact that AlphaFold hasn't been around for very long — I mean, DeepMind's 14 years old, AlphaFold we started hitting benchmarks about four years ago and a Nobel Prize already. So I think this is really showcasing that we can use AI as a new tool for discovery and we can have impact at digital speed. So I think this is an exciting and important moment. And I think, yeah, just surreal, exciting.
And this is really like what you do at DeepMind. It's basically real life application. Yes. For the theory of AI, where's it been most incredible to see it?
Oh, you know, we have such a diverse portfolio within Google DeepMind. So we work on generative AI, Gemini of course, and generative AI. We power Google products. Also, we still believe that not just scaling data, but there's still breakthroughs to be had in generative AI. How do we get it to be more factual, grounded, and think about reasoning and concepts? So there's memory. There's still quite a lot in order for AI to add value. We're in the very early days of generative AI, so we still do a lot of fundamental research, but then we also have this AI for science, and AlphaFold is really, I think, quite extraordinary. And yet we're doing work in weather forecasting. Last week, actually, being an American, a Californian in London, I'm so excited about GenCast, which came out last week, which does state-of-the-art weather prediction for 15 days. So hopefully next year airlines will adopt it and then my flights will run on time. But also a lot of the work around learning. I'm really excited about some of the learning science, material science, and discovering new materials. Last year we announced the discovery of about 400,000 new materials, which now scientists are validating and thinking about how could we make new batteries to power electric vehicles. So I think we're in the very early stages of how AI can help us solve and address some of humanity's biggest challenges.
How do you measure success and actually how do you prevent it from doing some of the bad things that we read about?
I think, you know, if I could wave a magic wand, I'd go back in time and say, like, let's actually rebrand this. I think artificial intelligence sometimes overshadows, I think, what the technology is and what it's meant for. You know, at Google DeepMind, we believe that, you know, our mission is to build it responsibly to benefit humanity. So the way I describe it is like, what do we do? We build AI, we build the foundational models. How do we do it responsibly, and why do we do it? To benefit humanity. So we kind of keep — in my job as chief operating officer is actually to make sure that we're doing that. And I think that if we can think about the application of AI and it starts to add value, then suddenly we don't call it AI anymore. If your phone battery lasts longer, that's adaptive battery, right? But there's AI charging, movie recommendations, song recommendations, maps, Google Translate. I mean, all of these are AI-powered tools, but we don't talk about them as being so. AI is around us everywhere. So I think if I take that and scale it, I feel like when we start talking more about the benefits and the features and how it's helping us, and we move away from just talking about risk, then I think that's what success looks like. But that takes earning people's trust, and that means building the technology with people and not having AI happen to them.
Yes. And in terms of disruption, is there really science that you would be most proud of in terms of disrupting or advancing it, making it quicker to find cures?
Yeah. So maybe would it be helpful to give a little story on AlphaFold? Feels kind of timely for today. Yeah, one minute. Okay. So Parkinson's, Alzheimer's, these are all protein-based diseases. There's a 50-year grand challenge in trying to understand the structure of a protein, how it folds. If you can do that, you can understand its function. If it misfolds, you can understand what's wrong. AlphaFold, we were able to fold all 250 million known proteins to humankind, and then provide it in an open database for scientists to use. There are 2.2 million researchers using this database. I didn't even know there were 2 million people watching proteins. And it used to take scientists four years of a PhD to do a single protein. So that's almost like advancing the field of biology about a billion years, if we do crude math. But why that's important is we can understand diseases, we can do better therapeutics, food security — why are some plants more resistant to disease than others? Plastic-eating enzymes, dealing with industrial waste. These are also protein-based problems. So there's a lot of work happening. So to your point, if we can use AI for scientific discovery, much like we use a microscope or a telescope to understand the world around us in a new way, that's how we're hoping we can now have scientific breakthroughs at digital speed, but do it with the scientific community, because our expertise is machine learning and we really need to take a collaborative approach. This isn't about any one company or any one country.
I mean, it's pretty incredible that you are so based in the UK. Yes. Has that changed your perspective on things?
Absolutely. You know, I've spent 30 years at the intersection of technology and social impact. In fact, I'm an electrical engineer by training and a semiconductor engineer, so I spent most of my career in the US, but also six years in Asia. And so when I first took this role, I'm like, why am I moving to London to work in the tech sector? I could stay here in Silicon Valley. But what I found is very unique about being in London is the fact that there's so much cultural diversity and ethnic diversity as well. And I think you bring that into work every single day. You know, even just the blue plaques with the history, you're just reminded of everything that has happened here. And even our conference rooms are named after philosophers and scientists, and it's a kind of paying respect to all the rich history around here. And I think, again, you just bring that into your thinking long term, and you can take ambitious goals like AlphaFold versus worrying about who got funded, who's moving to which new startup, and instead say, okay, we're really trying to benefit humanity. What are the biggest problems that we can work on and solve and apply the scientific method to do that?
A lot of the businesses will be thinking like, how do I use this? What are the pitfalls? What should I worry about? How are you seeing businesses experiment with AI?
