When it comes to artificial intelligence, most of us think of chatbots or computer-generated images of six-fingered hands. But at the Institute for Protein Design at the University of Washington, Dr. David Baker has used AI to revolutionize computer programs that create brand-new molecules. Baker and his team build things never before seen in nature, designed specifically to fight disease or break down waste—and he just scored a Nobel Prize for it.

Raised in Seattle, Baker finds his escape climbing and skiing, but at the university he sees his primary role as connecting the young scientists around him. The institute hosts bagel breakfasts and happy hours, spurring conversation and collaboration. After the giant leap forward AI has prompted, Baker doesn’t presume to know what’s next. “Things are changing so fast, and things can come out of left field that you don’t expect,” he says. “I don’t really trust scientific visionaries…there’s so many things that you can’t predict.” 

On October 9, a few weeks after we spoke, Baker was awarded the 2024 Nobel Prize in Chemistry. He shares it with two London-based protein structure prediction researchers from Google DeepMind, Demis Hassabis and John M. Jumper. “I got a phone call, and my wife promptly started screaming,” Baker told the Nobel Prize interviewer about hearing the news. He will receive half of the million-dollar prize at the December 10 award ceremony in Sweden. Baker called his fellow awardees “inspirers” rather than competitors, keeping with his collaborative view of groundbreaking science. 

This is definitely the world center for protein design.

Compared to simple drug molecules, like aspirin or any of the drugs we take, or the simple chemicals around, proteins are thousands of times bigger. They’re absolutely enormous.

From a chemist’s point of view, what proteins do is pretty magic.

For a very long time, basically since proteins were discovered, all we could do was take proteins that existed in nature and make them in the lab, like insulin.

It wasn’t really thought possible to make new proteins until relatively recently.

Once we could make new proteins from scratch, my group has really focused on, Can we now make brand-new proteins that would, say, improve health or break down toxic compounds in the environment or make new kinds of material?

The scope is enormous.

We had been developing this method for 20-plus years, called Rosetta, and then within a very short time, basically the new [AI] methods we were developing…completely replaced the traditional methods.

They work very analogously to…DALL-E, the image generation program. It’s basically generative AI.

Suddenly, there’s this new kid on the block that’s even more powerful.

The most advanced medicines...they’re really blunt instruments. They go in and they bind to something, and they block an interaction. But the problem is you get systemic effects; chemotherapy is really bad.

We’re trying to make proteins that are just smarter, that only act at the right time and place in the body. So they kill cells in the tumor, but not anywhere else.

There’s a bit of the worry about what it could be used for that’s bad. But my feeling on that is not only that you can regulate it at the level of gene synthesis, but nature’s already perfected ways of doing death and destruction. Like the Spanish flu, we saw the coronavirus.

With protein design, you could solve so many different current problems in medicine. We’re working on proteins for carbon fixation now, for climate and sustainability, for technology.

Ever since I started, really, my main role is to connect people.

There’s really intense social engineering here. There’s free food every day of the week. All to maximize collaboration.

I have this communal brain concept where each researcher is like a neuron. Individual neurons can’t do much, but they’re all connected together.

I believe ideas come from really smart and energetic people talking together.

We’re definitely riding the wave, and there’s a lot more exciting stuff ahead.