OpenAI says AI solves century-old fluid dynamics puzzle

OpenAI announced that its top-secret artificial intelligence model had cracked a math puzzle that has puzzled experts for centuries. The company claimed the feat was achieved in just a few days and required millions of dollars in computing power. The puzzle, known as the Navier-Stokes problem, is one of the seven Millennium Prize Problems set by the Clay Mathematics Institute.

These problems are among the most difficult open questions in mathematics and have been awarded a $1 million prize, so far unsolved, for each. OpenAI's work revealed that the mathematical equations describing fluid movement can break down over time, answering a key aspect of the Navier-Stokes problem. Mark Chen, OpenAI's chief research officer, emphasized this breakthrough as a major milestone for AI research, hinting at its potential to solve even more complex problems in the future.

However, OpenAI refrained from claiming the $1 million prize, citing the institute's stringent evaluation process, which requires published solutions in peer-reviewed journals and a two-year waiting period to be considered by a Clay Mathematics Institute committee. The institute's President, Martin Bridson, highlighted the rigorous evaluation process to ensure the utmost credibility.

The AI-driven achievement has sparked debate about whether machine-generated solutions truly embody human comprehension. OpenAI's account of the journey began with the training of a more advanced model in late August, followed by the model being given the task of tackling the remaining six Millennium Prize Problems shortly after.

A notable development came when competitor researchers Tristan Buckmaster and Levent Alpoge from Anthropic released their own work on AI-assisted solutions to related equations.