OpenAI claimed to have solved the longstanding Navier-Stokes problem using massive computational resources, sparking controversy after it was revealed their approach closely mirrored work by mathematicians Tristan Buckmaster and Levent Alpagay, who had used OpenAI’s Codex in their research. The dispute raised ethical concerns about AI’s role in mathematical research, data usage, and the potential impact on scientific collaboration, highlighting broader issues of trust and transparency in AI-driven discoveries.
The video discusses a recent claim by OpenAI that they have solved the Navier-Stokes problem, one of the most challenging and longstanding problems in mathematics related to fluid dynamics. The Navier-Stokes equations, developed in the 1800s, describe how fluids and gases move and are crucial for applications like weather forecasting and engineering. However, mathematicians have long struggled to prove whether these equations always hold true or can break down, a question so difficult that it was designated a Millennium Prize problem by the Clay Mathematics Institute with a $1 million reward.
Tristan Buckmaster, a math professor at NYU and a leading expert on the Navier-Stokes equations, had been working on related problems with Levent Alpagay, a mathematician at Anthropic. Their progress accelerated after they began using OpenAI’s Codex to assist with their research. In mid-August, they achieved a breakthrough by finding a counterexample for a simpler version of the equations, the Euler equations, marking the closest anyone had come to solving the Navier-Stokes problem. Shortly after, OpenAI reportedly used a massive computational effort involving thousands of agents and millions of dollars in compute power to claim they had solved the full Navier-Stokes problem.
The situation became contentious when Buckmaster contacted OpenAI to clarify their use of his and Alpagay’s work. During a phone call, Buckmaster learned that OpenAI’s approach closely mirrored the novel method he and Alpagay had developed and that the AI model had been trained on their code sessions, though OpenAI denied accessing user data. OpenAI allegedly offered Buckmaster two options: publish his Euler results first and let OpenAI publish the Navier-Stokes proof crediting him, or write the Navier-Stokes paper himself but exclude Alpagay due to his affiliation with a rival company. Buckmaster rejected both options and threatened to go public, leading to a tense exchange.
Following the dispute, both Buckmaster and Alpagay, and OpenAI published their respective papers and statements. Buckmaster and Alpagay detailed their experience and concerns about OpenAI’s conduct, while OpenAI maintained that no user data was accessed improperly and that their proof was distinct. The controversy sparked broader discussions about the ethics of AI research, data usage, and the potential impact on the collaborative nature of mathematical research. Notably, mathematician Terence Tao cautioned that such competitive AI-driven research could discourage open sharing of ideas, undermining centuries of scientific progress.
The video concludes by highlighting the broader implications of this episode for trust in AI companies and the future of scientific collaboration. It also briefly mentions that OpenAI is reportedly close to solving another Millennium Prize problem, emphasizing the power of combining human ingenuity with massive computational resources. The narrator ends with a sponsor message and a reminder to be cautious about which companies to trust with sensitive research.
