Mathematicians Express Mixed Reactions to OpenAI's Release of Solutions to Hundreds of Open Problems

OpenAI released solutions to numerous open problems in mathematics, including the non-amenability of Thompson's group F, the falsity of the Eilenberg–Ganea conjecture, and the resolution of the Boone–Higman conjecture. The release, which involved over 700 'papers' on GitHub, was described by some as a 'tsunami' that disrupts the mathematical community. Reactions were mixed: some mathematicians expressed sadness and anger, noting that their own research programs (such as those focused on the Boone-Higman conjecture or Katok’s entropy conjecture) were rendered obsolete or 'killed' by the AI's solutions. Critics highlighted that many of the papers were 'badly written,' 'slop,' and lacked transparency regarding authorship and methodology. Others noted that while the results are impressive, the models have limitations, solving only about 700 of 4,000 attempted problems and often providing weaker results than targeted (e.g., quasi-Riemann Hypothesis instead of the full RH). There is significant concern about the impact on junior researchers, the future of peer review, and the potential for AI to undermine the collaborative and understanding-driven nature of mathematical research. Some voices, however, expressed optimism about the potential for 'Big M' mathematics and the ability to tackle larger questions.