OpenAI shakes up mathematics again, debuting dozens of new math solutions

OpenAI's latest release of mathematical solutions continues a pattern of the company sharing results from proprietary models without full transparency regarding methods or access. The firm acknowledges potential expert scrutiny, indicating a commitment to improving paper quality and presentation for future releases. This suggests an awareness of past criticisms regarding methodological clarity and peer review, as the company implied it expects experts to take issue with the new solutions.

The company's previous announcement about solving a portion of the Navier-Stokes equation was deemed by academics as using a 'mathematical loophole,' leading to questions about the validity and scope of the solution. This history informs the current skepticism from experts, who are still awaiting access to the model reportedly used for the latest solutions, particularly given the claim of a single prompt to a single AI agent.

The Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study stated it did not endorse AI developers using proprietary models without the scientific community's involvement. While OpenAI announced a new advisory group to provide recommendations, the current release appears to use the same unreleased model, highlighting an ongoing tension between rapid AI development and the scientific community's call for open validation.