OpenAI says internal model solved more than 100 longstanding math problems
OpenAI says an internal model solved more than 100 longstanding problems across most areas of mathematics after training began on August 28. The company also attributes to it work on the Navier–Stokes Millennium Problem and reportedly the Hodge conjecture, another Millennium Prize Problem. OpenAI has not disclosed which problems were solved in detail, how the model produced the results, or what they mean for mathematics and science. The Navier–Stokes result has already prompted debate among researchers.
The announcement came alongside the creation of an independent Advisory Group on Mathematics and Artificial Intelligence, based at the Institute for Advanced Study. OpenAI says the group will advise on communicating research to mathematicians and the public. It includes Fields Medalist Timothy Gowers, who agrees with concerns that AI could weaken mathematical understanding but did not sign an open letter criticizing the emphasis on solving open problems.
The group may issue unsolicited advice, speak publicly and publish recommendations, but OpenAI says it will not advise on the pace of the company’s mathematical research. Gowers has warned that rapid progress could reduce the incentive to become a mathematician and prompt policymakers to question funding for human experts.
Written by Sakhanda Wire AI from reporting by The Decoder. Facts, quotes and full context are in the original article below.