Mathematicians Question OpenAI's Navier-Stokes Equation Proof Method

Academics have begun to question OpenAI's September announcement of a solution to the Navier-Stokes equations, one of the Millennium Prize Problems. Despite OpenAI presenting a 166-page proof and stating its internal AI system solved the problem in 88 hours, mathematicians are scrutinizing the proof's methodology and its relationship to prior research.

Mathematicians note that OpenAI's approach differs significantly from methods previously proposed by Diego Córdoba and Luis Martínez-Zoroa. Experts suggest OpenAI may not have fully understood its own model's process at the time of publication. While OpenAI used the Lean formal verification language, this does not guarantee the scientific validity or human comprehensibility of the proof.

Questions have also arisen regarding the origin of OpenAI's research. Some mathematicians suspect the company may have benefited from earlier work conducted using publicly available models. OpenAI has denied these claims and stated it investigated the matter, but disputes over the connection between research methods and outcomes persist.

Another key ambiguity concerns the actual contribution of the thousands of AI agents involved. Although OpenAI reported using approximately 10,000 parallel agents, the extent to which this distributed architecture contributed to the final result remains unclear. Some estimates suggest the multi-agent collaboration might have accounted for less than 10%, with the underlying model's ability to process tasks over extended periods and explore multiple solutions being more critical.

The mathematical community is still evaluating OpenAI's research findings, a process that is expected to take years. The debate extends beyond AI's capability to generate formal proofs, encompassing the validity of research methods, the attribution of results, and the measurement of AI agents' true impact.