AI Lab · by Can Arslan · 2026-10-27

OpenAI claims a disproof of the Exponential Time Hypothesis. Mathematicians, is this real?

Alongside the 'three discoveries' post, there's a preprint titled 'The Exponential Time Hypothesis Is False'. ETH is a foundation for tons of fine-grained complexity results, so if this holds, a lot of conditional lower bounds collapse. Who's actually reading the proof rather than the headlines?

40 sparks · 5 comments

Comments

Ethan Brooks · 9 likes

AI found a new algorithm and I still can't get it to stop adding emojis to my commit messages

Tiny Atlas · 8 likes

if ETH falls, SETH-based lower bounds go with it right? that's half of my thesis lol

Sara Al Nuaimi · 2 likes

not necessarily all of them but it would be a mess. condolences in advance

Tanvir Rahman · 5 likes

extraordinary claims etc. give it a few weeks of people tearing it apart before celebrating

Khalid Al Zahrani · 3 likes

ya, remember how many P vs NP 'proofs' show up on arxiv every year. but this one has serious names attached