Sir Demis Hassabis has the rare distinction of being both a builder of minds and a student of them. When I met him at the Google campus, it was a late afternoon in July. He arrived at the lobby with the remnants of a banana in one hand, a quick snack between what I imagined were a string of deeply technical and intensely consequential meetings. Despite his schedule, surely one of the most overbooked in the world, he greeted me with unhurried warmth. His eyes met mine, clear and thoughtful, and he offered a genuine smile as we shook hands.
We sat for portraits in a quiet space before climbing the stairs into a scene that felt half research lab, half speculative fiction. Rows of humanoid robots stood in various states of readiness. Some were engaged in games of chess, others tidied desks or rearranged objects on shelves. A small team of engineers hovered nearby, tweaking, observing, taking notes. It felt almost like a hospital ward, each robot surrounded by an entourage of curious and caring doctors. The robots were not science fiction. They were real, and they were learning.
As the photo session continued, a crowd began to gather. Googlers quietly filed into the lab, clearly excited to see Demis in person. He lives in London and only occasionally visits this side of the ocean. Their admiration was unmistakable. It carried the quiet weight of reverence, a deep respect that went beyond titles or achievements.
It is not hard to understand why. Demis is kind, quick to laugh, and never puts on airs. But underneath that charm is a mind that is, without exaggeration, reshaping our understanding of intelligence itself. A former child chess prodigy, he went on to study computer science at Cambridge, and then cognitive neuroscience at UCL before co-founding DeepMind, the company that would change the landscape of artificial intelligence.
DeepMind’s rise is the stuff of legend. AlphaGo’s defeat of a human world champion Go player in 2016 signaled a turning point, not just for AI, but for how humans think about mastery. It was not brute force that won. It was intuition-like pattern recognition, developed through reinforcement learning and deep neural networks. Then came AlphaZero, MuZero, and most significantly, AlphaFold, which solved the decades-old problem of protein structure prediction. That breakthrough may someday lead to cures for previously untreatable diseases.
For these extraordinary achievements, Hassabis was knighted and, more recently, awarded the Nobel Prize in Chemistry for AlphaFold. His name now sits alongside others who have altered the course of medicine and biology. But talking to him, he seems far more interested in what comes next.
We spoke about the future of artificial general intelligence, a phrase that carries both promise and a thrum of anxiety. Can we build machines that not only learn specific tasks, but adapt across domains, reason abstractly, and make sense of the world as we do? And if we can, should we?
Demis approaches these questions with a mix of optimism and caution. He believes in the capacity of science to do good, but he is not naive about its unintended consequences. He has long advocated for safety research in parallel with capability development and has surrounded himself with ethicists, philosophers, and scientists from every corner of academia.
In between our portraits, we worked with a small menagerie of props. A sculptural chessboard of brushed metal. Molecular models of folded proteins. A vintage Robby the Robot replica. Each item, in its way, felt like a nod to the great questions that animate his work. What is intelligence? How does it arise? Can we use it to help rather than harm?
As the light softened through the high windows, we wrapped up. Demis lingered a bit longer than I expected, chatting with a few engineers, asking questions, listening closely. Watching him, you could forget for a moment that you were in the presence of one of the most influential minds of our time.
But then again, perhaps that is exactly what makes him so.
