OpenAI publishes new AI-generated maths and computer science results for outside review
If the Exponential Time Hypothesis proves false, algorithms that rely on it as a limit could become faster than assumed.
OpenAI has disclosed a new set of AI-generated results spanning mathematics, computer science and physics, following an earlier release of manuscripts covering 372 math problems. For this round it is using three separate approaches: math results are published for independent researchers to examine, computer science results come with outside experts invited to write their own explanations, and physics material is being handled by a separate group that has not yet published anything. Among the manuscripts now public on GitHub is one establishing Sogge's local smoothing for waves in every spatial dimension, a conjecture about how solutions to the wave equation behave. A second manuscript argues that the Exponential Time Hypothesis, long assumed true and relied on by hundreds of papers in fields such as databases, computational biology and geometry, is false, which would open the door to faster algorithms if confirmed. OpenAI says it will not claim academic credit for these discoveries, with credit instead going to the people who write up the results, who are also responsible for ensuring the AI output properly acknowledges the researchers and ideas it drew on.
Why it matters
The announcement puts AI-generated results in front of independent researchers before they are accepted, and leaves credit with the people who write them up rather than with OpenAI. Claims of this size are only meaningful once outside experts have checked them, and that checking is still ahead. For readers who follow the field, the open question is whether the manuscripts hold up under scrutiny, since many papers in several fields depend on the hypothesis being true.
Matters or hype?
Does this story matter, or is it hype? Vote before you see what everyone else said.