OpenAI's blockbuster Navier-Stokes proof isn't what we thought it was
Rumors that OpenAI’s “Navier-Stokes” solution stole the work of two mathematicians may have been overblown
OpenAI's landmark 166-page proof of the Navier-Stokes problem, a "million-dollar" math puzzle, has been revealed, but experts are now questioning its authenticity. Initially, rumors surfaced that the proof adopted the approach of two Spanish mathematicians, Diego Córdoba and Luis Martínez-Zorúa. However, it now appears that OpenAI's solution diverges significantly from the Spanish program.
The Navier-Stokes equations, which describe fluid flow, were exploited by OpenAI through an "agent swarm" of 10,000 large language models. These models spent 88 hours computing time and 130 billion tokens to create a blow up, a scenario where the equations predict infinite fluid speed – a physically impossible situation. This proof builds on a loophole in the Navier-Stokes problem's framing, but its implementation is entirely unique.
Despite OpenAI's claims that its model was not influenced by the Spanish program, the superficial resemblance raises questions about the extent of influence. As top experts continue to decipher OpenAI's proof, the lack of transparency in the workings of LLMs makes it challenging to determine the swarm's exact reasoning. OpenAI's initial claim that their model was not influenced by the Spanish program proved to be misleading, highlighting the challenges in navigating the "black box" nature of AI models.
Written by urgent.news from Scientific American's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
