I didn't sign the Fields medallists' letter

Timothy Gowers explains why he refused to sign a letter from 25 Fields medallists about AI’s threat to mathematics. He opens with a childhood anecdote about attacking Fermat’s Last Theorem and discovering difference sequences, using it to illustrate how struggling with problems builds deep understanding. He agrees a crisis exists but rejects the letter’s implied hierarchy that conceptual understanding must always trump problem-solving. He clarifies his ties to OpenAI - early access but no payment - and his leadership of a Cambridge group on automatic theorem proving, arguing that these connections do not determine his stance. He stresses that mathematicians span a spectrum from problem-solvers to concept-seekers and objects to any message that devalues one temperament over the other.

He frames two outcomes: one in which powerful models are publicly available and quickly solve many problems, producing a flood of results faster than humans can absorb; and a second in which releases are restricted and companies only solve problems when the community deems it worthwhile. He argues the former will expand available results and tools but will erode the hard-won cognitive skills developed by sustained struggle, impair collective absorption of knowledge, and risk sidelining mathematicians motivated by problem-solving. He urges a nuanced response that balances rapid automated progress with preservation of human-driven understanding.