OpenAI Reports AI Advances in Mathematics, Physics and Computer Science

OpenAI has announced that its artificial intelligence models have made progress on three complex scientific problems spanning mathematics, physics and computer science.

Among the results is a finding that the company says disproves the Exponential Time Hypothesis, known as ETH. The claim concerns a major assumption in theoretical computer science about the computational difficulty of certain problems.

ETH broadly holds that some problems cannot be solved in substantially less than exponential time as their size increases. A confirmed refutation would have significant implications for how researchers understand the limits of efficient computation.

The announcement also describes advances in mathematics and physics, although the summary provided does not identify the specific problems or explain how the models contributed to solving them.

The findings remain claims attributed to OpenAI in the available information. No details were provided about independent verification or peer review, leaving the scope and significance of the reported advances to be assessed.