After using millions of dollars worth of compute, OpenAI said it may have cracked a centuries-old math problem about how fluids such as air and water behave. The result could amount to a major scientific advance if it stands up.
However, the announcement was not without controversy.
Prominent mathematicians said OpenAI piggybacked off their research, creating a dispute over the proof, data use, and credit.
Here's what to know
As Wired reported, OpenAI said an internal AI effort produced a result connected to the Navier-Stokes equations, a famously difficult set of equations that is also one of the Clay Millennium Prize problems, which carry a $1 million reward apiece.
Sebastien Bubeck, a mathematician and AI researcher at OpenAI, said the company increased its focus on the problem after hearing Anthropic might be making progress as well. He said OpenAI deployed more than 1,000 agents for over 50 hours, with the agent total rising to 10,000.
"I thought there must be a mistake somewhere," Bubeck said, per Wired. "And on Sunday morning we had the final solution, Lean-formalized and everything."
TCD Picks » Quince Spotlight
💡These best-sellers from Quince deliver affordable, sustainable luxury for all
Tristan Buckmaster, a mathematician at New York University, and Levent Alpöge, an Anthropic researcher, pushed back. They posted documents describing advances they had made in a related area and said AI tools, including Claude, had been part of their work.
Buckmaster said OpenAI ramped up its effort after learning about the progress the team had made and questioned whether his usage of OpenAI's products may have trained OpenAI models. OpenAI executives denied directly accessing the work, according to Wired.
More background
If artificial intelligence models can generate or verify proofs, research in fields such as mathematics, physics, and engineering could move much faster. At the same time, the situation raises questions about transparency, authorship, ownership, and ethics.
The total cost was more than money, as AI is closely tied to the energy grid. Training and operating large models can require vast amounts of electricity and water for data center cooling, which can put pressure on infrastructure and even drive up energy costs. However, AI also has the potential to support grid management, improve forecasting, and optimize clean energy systems.
"While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models," OpenAI wrote in a blog post, as Wired reported.
Ven Chandrasekaran, a mathematician at OpenAI, also said the company's solution was meaningfully different from the approach taken by the other researchers.
What's being done?
The main response has been intense scrutiny.
Bubeck and other OpenAI leaders said neither researchers nor agents reviewed Buckmaster and Alpöge's prompts or unpublished proof. "We, whether it's the researchers or the agents, did not see any of their work until it was released publicly last night," he said, per Wired.
That leaves several questions unresolved: whether the proof will hold up under peer review, how credit should be assigned, and what standards should be followed when proprietary data is used in interactions that may improve other models.
The dispute also raises a more practical question about how much trust society should place in powerful AI systems as they move into high-stakes fields.
OpenAI's announcement may be remembered as a milestone in machine-assisted science or as an early sign of how complicated the future of business and scientific attribution could be.
"I want to be extremely clear that we recognize the priority of Levent Alpöge and Tristan Buckmaster's work on unforced Euler, and we have nothing but congratulations to them on this monumental achievement that they have made. To be clear, we did not use their prompt or proof to prompt our models or direct our agents," Bubeck said, according to Wired.
Where can I learn more?
These stories look at the bigger questions behind AI use, from who gets to set the rules to the areas where the technology could do public good.
• In Congress, a controversial plan quietly advancing could limit AI oversight despite climate and health warnings.
• At Research Park, engineers are testing a futuristic cooling system to reduce AI's colossal energy burden.
• At MIT, Priya Donti is building an AI system to optimize renewable grids more responsibly.
They show why arguments over AI breakthroughs touch policy, the power grid, and accountability.
Get TCD's free newsletters for easy tips, smart advice, and a chance to earn $5,000 toward home upgrades. To see more stories like this one, change your Google preferences here.
