Twelve months ago, we didn’t have working antibody design models. Now we have models that can compress 18 month design timelines down to one month, and agents that can collapse weeks of work by expert protein designers to an afternoon. This is an exciting time.<br>Simon Kohl is a former Google DeepMind research scientist on the Nobel Prize–winning AlphaFold 2 team, and is now the founder and CEO of Latent Labs, a startup that is building a future where AI agents can design safe, effective and personalized drugs at the push of a button for every patient. With 3 frontier models released in just 9 months, this future is coming quickly.

The future of AI-powered personalized drugs is only a few years away - Latent Labs founder Simon Kohl

Free Radicals

# Full Transcript

**Podcast:** Free Radicals

## Transcript

Welcome to the Free Radicals Podcast, where we interview the scientists and builders working to dramatically extend human lifespan and bring about a sci-fi future where humanity has full control over biology. Today's guest is Simon Kohl, a former DeepMind research scientist on the Nobel Prize winning AlphaFold II team. He is now the founder and CEO of Leighton Labs, a startup that is building a future where AI agents can design safe, effective and personalized drugs with the push of a button for every patient.

With three frontier models released in just nine months, this future is coming quickly. Simon walks us through his company's rapid progress from models for designing binders to autonomous agents that can design drug-like antibodies from a text prompt. Twelve months ago we didn't have working antibody design models. Now we have models that can compress 18 month design timelines down to one month, and agents that can collapse weeks of work by expert protein designers to an afternoon.

This is an exciting time. I'm your host Daniel Shur and my co-host is Eric Dai. I hope you enjoy this episode of the Free Radicals Podcast. All right, well, Simon Kohl, welcome to the Free Radicals Podcast. Thank you for having me. Looking forward to the conversation. We had a lot of issues getting set up in this room. We're at the Symbio Beta Conference, but smooth sailing from here now. It will be.

All right. So we want to start off talking about Leventhal's paradox, because we're going to talk about protein folding a lot in this conversation and computation. So predicting a protein structure from its amino acid sequence has been an open question since the 1960s, and the nature of the question was really crisply articulated in the form of Leventhal's paradox.

And the idea is that when you think about all the degrees of combinations for a protein to fold into, if it had to randomly sample all those positions, it would take more than the length of the universe for a protein to figure out how it should fold. And yet proteins actually do it in nanoseconds. And of course, alpha fold, which you worked on, is also able to do it in an inference call. And so I think it would be super interesting to start with, what is your intuition for how AI systems are able to solve this problem? Yeah, it's a great question.

As you said, really, protein structure prediction has been this really longstanding problem, like decades since this paradox, Leventhal's paradox, was hypothesized. Actually, I believe at a Nobel Prize recipient speech, this was shared.

And so the community has sort of worked on this problem since ages with the idea that, you know, if you were able to predict it, then you could sort of forego what takes months to sometimes years to sometimes being an impossibility in the lab to determine a protein structure. And so at DeepMind, we set out to work on that problem, entered this competition called C...

The future of AI-powered personalized drugs is only a few years away - Latent Labs founder Simon Kohl

Transcript

Built for investment professionals

Every day you wait, your competitors are finding
edge in conversations you're not hearing.

The future of AI-powered personalized drugs is only a few years away - Latent Labs founder Simon Kohl

Transcript

Built for investment professionals

Every day you wait, your competitors are findingedge in conversations you're not hearing.

Every day you wait, your competitors are finding
edge in conversations you're not hearing.