Frontier labs say their models refuted the Exponential Time Hypothesis; TechCrunch asks an MIT scientist what that means

TechCrunch published a report on October 27 examining claims from frontier AI labs that their models have "refuted" the Exponential Time Hypothesis, a central conjecture in computational complexity theory. The piece is built around an interview with an MIT computer scientist asked to explain what OpenAI's latest discovery actually means.

The Exponential Time Hypothesis holds that general satisfiability problems cannot be solved in sub-exponential time. It matters because the perceived difficulty of a large class of problems is measured against it: if the hypothesis fails, problems long treated as inherently hard may turn out to have far faster algorithmic routes than assumed.

The wording of the coverage is worth noticing. The claim is attributed to frontier labs, not presented as a verified proof. TechCrunch chose the explainer format, which itself signals that the result is still being absorbed rather than settled — the expert is on stage precisely because the conclusion is not yet common ground.

For the AI industry, the weight of this is not another benchmark being topped. It is that models are being used to produce new results in mathematics and theoretical computer science. That differs from having a model write code or retrieve facts: the output is meant to be checkable, and it challenges an existing body of knowledge, which makes verification far more expensive.

Before treating it as evidence that AI is doing science, several steps are missing. Can the argument be formalized and machine-checked line by line? Have human experts reviewed the key steps? And did the model propose the conjecture, construct the proof, or search and reassemble known techniques? Those three possibilities carry very different meanings.

What to watch next is whether the labs or the researchers publish proof details that others can check, and how the complexity-theory community responds. Until then, the accurate description is narrower: a frontier lab says its models produced a result that challenges a mainstream assumption, and the expert explanation of it has only just begun.