David Sinclair: "AI helped us do 160 years of biology work in two months"

What's on with mathematics. Like just over the last couple weeks they've solved more than 50 programs or 50 mathematical equations that they have struggled for decade upon decade to try to solve. And AI is like, oh, I got it. Yeah, well, biology is a little bit more complicated, but already AI is doing a lot of the science that we used to do in my lab. In fact, what we just recently did it in a couple of months, would have taken about 160 years to complete. What? Only getting faster. Eventually someone can make a drug on their own. Two months instead of 160 years. What was it that you had to do? So this drink and the pill that we're looking for. That's not easy. Right? So we have some candidate molecules. We have a cocktail that works. But we want to single molecule. Are you taking this stuff? Not yet. You're thinking about it. Yeah. I might, well, that is true. But it'd be a bad look if I died. So maybe I'll wait. I don't think so. We've done a lot of safety tests and it looks really safe. When you're telling me that in two months, it did something that would have required 160 years of work. Like what do you exactly mean by that? Yeah. And now it would probably be impossible in thousands of years. What we do is we're looking for molecules. A single molecule that can replace this cocktail. So you need a molecule to do three or four different activities in the cell. Finding one activity is what pharmaceutical companies do. We do it inside the computer. We've teamed up with a group at St. Jude's hospital where they can take 10 billion molecules. Now it's a trillion molecules. And also non-drug known molecules. So that funnel is uh, full of a trillion molecules. And we have to come down to one. How do we do that? Well, we don't look for a trillion molecules at the bench. That would take 160 years. We do it inside the computer. We just use a lot of compute and we dock each of these molecules against the enzymes that we know, that we know work thanks to Google DeepMind, solved the problem of how all our proteins look. 20,000 of them. And there's a lot of compute. But out the other end, a couple months later, we have two, 200 molecules that we can now test in the lab within a few weeks, actually next week, we're testing these 200 molecules against human cells. And we've also got AI helping us to tell us if the human cells go from 92 years old back to 20. So, uh, that's the goal. We're trying to make people live longer and healthier lives.