This is Sir Demis Hassabis’s talk “Accelerating Scientific Discovery with AI,” given during a March 2025 visit to the University of Cambridge and posted on the official Cambridge Department of Computer Science and Technology (Computer Lab) YouTube channel. Hassabis is co-founder and CEO of Google DeepMind and shared the 2024 Nobel Prize in Chemistry with John Jumper for AlphaFold.
The lecture lays out his long-running thesis: that the way to make broad scientific progress is first to build general-purpose AI - to solve intelligence - and then turn it on hard scientific problems. He traces the arc of DeepMind’s work from game-playing systems such as AlphaGo, through AlphaFold’s prediction of protein structures, to the wider ambition of using AI as a tool for discovery in biology, chemistry, and beyond, including the drug-design work spun out into Isomorphic Labs.
This is a practitioner-level talk that is accessible without deep technical background. It is distinct from his shorter, formal Nobel lecture, offering a fuller, firsthand account of how one of the field’s leading figures connects his game-playing roots to his scientific ambitions.