Advisory Group on Mathematics and Artificial Intelligence

Commenters debated OpenAI’s announcement that its models have resolved hundreds of long‑standing mathematical problems, with several specific worries raised. Kingstnap called out the scope of the claims and worried about how results would be disseminated; mwkaufma and HarHarVeryFunny pointed to OpenAI’s explicit statement that the advisory group will not advise on pacing internal progress; piker noted Timothy Gowers’ refusal to sign; robotpepi and others called the publicity a stunt and criticized weak citations in the Navier-Stokes paper. Gaigalas warned that automated proofs without human understanding could flood mathematics with uninterpretable results, SirMaster and kingstnap named P vs NP as the true benchmark, and franticgecko3 and optimalsolver read the episode as the first major field to be overtaken by AI.

Commenters also split on implications and responses. Certhas argued for urgent regulation or taxation and UBI if labor impacts arrive; OtherShrezzing and HarHarVeryFunny dismissed advisory boards as toothless PR; laichzeit0 and others doubted non‑Western competitors would care about Western advisory etiquette. Several defended the practical value of mathematics (vablings, sigbottle) against claims that research is largely useless or grift (jfewhfuehg, dgellow). mturmon questioned the choice of IAS as host. Overall, voices divided between alarm about pace, loss of understanding, and labor effects, and skepticism about PR, incentives, and the advisory group’s efficacy.