OpenAI Cracks a Million-Dollar Math Problem — and the Credit Fight Starts Immediately

The most significant AI story of the day is not the breakthrough itself, but the controversy surrounding the claim of ownership. OpenAI announced that an undisclosed internal model, described as significantly more capable than its recently launched flagship, had produced a proof solving the Navier-Stokes problem—one of mathematics' seven Millennium Prize problems.

The company claimed it utilized roughly 10,000 AI agents simultaneously, ran them for 88 hours, and spent millions of dollars in compute to achieve this result. Sam Altman hailed it as one of the most remarkable moments in the company's history. However, the victory was marred almost immediately by disputes. Mathematicians and researchers at competing institutions claimed they had independently been working on the same approach and had posted partial results online just before OpenAI's announcement.

The NYU mathematician stated that OpenAI only intensified its efforts after discovering this work and questioned whether the data he fed into OpenAI's tools directly contributed to the final model. OpenAI responded that they never viewed the specific work and that no particular user data was accessed, though they hesitated to rule out that general usage patterns could have influenced their models.

Despite the controversy, the story also highlights the significant gap between private and public AI capabilities, a gap that has been evident in recent model releases. This gap between research and public-facing models has also sparked concerns about the transparency and safety of these systems. Additionally, a new always-on personal AI agent launched by a major tech company is available for use, integrating seamlessly with various services and offering a user-friendly interface.

This development adds another layer to the ongoing competition among AI labs to push the boundaries of what these systems can do.