"Almost every drug developed from now on will have probably used AlphaFold."

On April 07, 2026, Google DeepMind CEO and Nobel Prize winner Demis Hassabis joined Cleo Abram on “Huge Conversations” for an expansive, visually engaging interview. Using a game of Jenga—where each block represented a different DeepMind project—to frame the discussion, Hassabis detailed the inside story of AlphaFold, the shift from quiet lab science to the rapid generative AI race, the creative leaps of AlphaGo, and his ultimate sci-fi vision for an AGI-powered future.

The AlphaFold Breakthrough: “Why Don’t We Just Do That?”

The conversation kicked off with the project that earned Hassabis a Nobel Prize: AlphaFold. For 50 years, the “protein folding problem”—predicting a protein’s 3D structure from its 1D amino acid sequence—stood as a grand challenge in biology. Hassabis recognized early on that cracking it could revolutionize drug discovery.

Abram highlighted a pivotal 2021 meeting where the team discussed setting up a server for scientists to submit proteins for folding. Hassabis recalled a sudden realization that changed everything:

“I was just sort of doing the back of envelope calculation like how many proteins are there known to science… 200 million. And then how many computers do we have… and if we folded one every 10 seconds. I sort of realized sort of in the middle of that meeting I was fiddling on my phone that it would be possible in a year. So why go to all the effort of building the servers and the databases… when we could just actually fold everything ourselves, everything anyone could ever request and ever want, and then put it on a database somewhere for free for all the scientists in the world to use.”

Today, over 3 million scientists use AlphaFold. Hassabis also detailed the next frontier with his spin-out company, Isomorphic Labs, which uses advanced models to speed up drug discovery from a grueling 10-year process into a highly efficient digital search. By searching millions of compounds in silico to check for efficacy and toxicity before testing them in a wet lab, they aim to radically improve human health.

The Generative AI Race: Dealing with the World as We Find It

Abram noted that Hassabis founded DeepMind to solve intelligence and use it to solve everything else, selling to Google for the freedom to explore science. But the release of ChatGPT changed the trajectory of the entire industry. Abram asked what was gained and lost in this shift.

Hassabis admitted he originally envisioned a slower, more deliberate path. “In my ideal world,” he said, scientists would collaborate on AGI “in a kind of CERN-like way effort… carefully considering each next step.”

However, language models progressed faster than anticipated. “It turned out transformers which my Google colleagues invented and some reinforcement learning as well on top was enough to crack things like language,” Hassabis explained. “We were sort of playing around with that so were the other leading labs but… fair play to OpenAI, they scaled it and then they put it out there.”

While he lamented the “ferocious commercial pressure race” and geopolitical pressures, he acknowledged the benefits: incredibly fast progress, the democratization of AI, and the necessity of real-world stress testing. “It’s not the way I dreamed about years ago where we would be sort of contemplating this philosophically,” Hassabis said. “But I’m a pragmatic engineer so… we have to deal with the world as we find it and make the best of that.”

Move 37 and True AI Creativity

To understand DeepMind’s unique approach to AI, Abram pointed to the historic 2016 AlphaGo match against Lee Sedol. Unlike expert systems like IBM’s Deep Blue—which Hassabis critiqued because “it can’t even play strictly simpler game like tic-tac-toe”—AlphaGo learned from experience.

In game two, the system played “Move 37,” a highly unusual, creative move that ultimately won the game. Hassabis described this as the turning point for the company: “Not only did it win the match but it was how it won and with these creative new ideas like move 37, and that for me was the signal that we were ready to turn it to scientific problems like AlphaFold.”

He also highlighted AlphaZero, which removed human heuristics entirely, learning chess and Go strictly from self-play. Hassabis believes these underlying techniques are the key to the future: “I think we need these types of ideas back here now with our foundation models… we still need this ability to search and think and reason on top of those models.” This methodology is already driving breakthroughs in algorithmic efficiency (AlphaTensor) and computer chip design (AlphaChip).

Navigating Government Use and AI “Going Rogue”

As AI becomes central to geopolitics, Abram asked how Hassabis hopes governments will use it. He pointed to public health, education, and optimizing energy grids, noting DeepMind had already used AI to save 30% of the energy in Google’s data center cooling systems.

But he didn’t shy away from the immense risks, highlighting two massive concerns that he feels the world isn’t taking seriously enough:

1. Bad Actors: “Repurposing these technologies that we’re trying to build for good… for harmful ends.”
2. AI Going Rogue: As systems enter the “agentic era” and act autonomously over days or weeks, control becomes paramount. “How do we make sure… the guardrails are put in place that we can ensure that they do exactly what they’ve been told to do… and there’s no way of them circumventing that or accidentally breaching those guardrails? That’s an incredibly hard technical challenge.”

While DeepMind actively deploys immediate safety measures like SynthID for watermarking deepfakes, Hassabis stressed that long-term AGI alignment requires unprecedented international cooperation.

A Sci-Fi Vision of the Future

Despite the heavy responsibilities, Hassabis remains fundamentally driven by curiosity and optimism. “I want to use AI as a tool to help us understand the nature of reality around us,” he stated, noting that AGI could solve the “root node problems” of science.

When asked to play out the sci-fi movie in his head, Hassabis envisioned a post-AGI world within the next 50 years where the energy crisis is solved through AI-accelerated fusion or advanced solar. “If that’s sort of zero somehow because we can just make infinite rocket fuel out of seawater because we’ve cracked fusion… that really unlocks space,” he said. He imagined mining asteroids, building Dyson spheres around the sun, and curing terrible diseases.

“That should hopefully lead to maximum human flourishing,” Hassabis concluded. “And traveling to the stars bringing consciousness to the rest of the galaxy—that would be I think an amazing outcome.”