The inside story behind OpenAI’s million-dollar math proof
AI’s solution to the Navier-Stokes problem, and the details of how it happened, preview a chaotic future for math—and the world
On September 8, OpenAI announced they had solved one of mathematics' "million-dollar problems." This announcement came shortly after N.Y.U. mathematician Tristan Buckmaster publicly criticized the company for potentially stealing his work. Buckmaster's lecture at New York University previewed a chaotic future for math and the world.
His talk began with a packed room of attendees eager to hear about the groundbreaking math proof. The problem, known as Navier-Stokes, involves predicting how fluids flow. Mathematicians had long suspected the equations sometimes failed, causing fluids to "blow up" with infinite velocities. Solving this problem would earn the mathematician a $1 million prize from the Clay Mathematics Institute.
Buckmaster had previously collaborated with Google's DeepMind team on this goal. He accepted Levent Alpöge's proposal to work on a potential loophole in the Clay Institute's problem statement. Alpöge, an Anthropic employee, suggested that a finely-tuned force could manipulate the fluid to blow up. Buckmaster instructed LLMs from Anthropic and OpenAI to carry out this program.
They successfully applied the program to simpler equations describing idealized fluids. Alpöge's work with LLMs has been controversial, as he sometimes posts cryptic results on social media, leaving mathematicians unsure if the work represents his employer. Meanwhile, OpenAI's approach remains unclear.
Written by urgent.news from Scientific American's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
