Mathematician at Cambridge University, working on automatic theorem proving, in which computers are trained to find proofs by mirroring how humans do so. His lab is exploring whether AI models can make non-obvious connections between mathematical subfields. As of April 2026 Gowers finds that large language models struggle to apply what they have learned solving one problem when tackling another. He also notes that human mathematicians develop an "aesthetic sense", prompting them to look for neater proofs that can sometimes yield surprising results—a combination of aesthetic sense and deep knowledge that seems difficult for AI to emulate. On that front, he says, "humans still have the edge."
I'm prepared for all emergencies but totally unprepared for everyday life.