Terence Tao says AI makes proofs abundant and warns against solving problems alone

Fields Medal winner Terence Tao says AI is making mathematical proofs abundant, ending what he calls the math 1.0 era of scarce proofs, according to a QbitAI report.

He argues that solving open problems does not by itself add human understanding, and calls for AI results to be reported together with failures and compute costs.

In a Caltech lecture titled "Math 2.0," as quoted in an X post by Rohan Paul, Tao says blind optimization of problem-solving with automated tools is now actively harmful to the long-term health of mathematics. He says open problems serve as guides rather than destinations, and that reaching them prematurely can sterilize the surrounding field. The QbitAI report also describes his proposed five-stage path from proof generation to integration into standard mathematical theory.