Did AI just solve P vs NP? Not quite. But OpenAI and Anthropic both say their AI disproved the Exponential Time Hypothesis, a famous math conjecture, on the same day, both with machine-checked proofs. I read both papers, and honestly the most shocking part is the number that makes the whole thing useless in practice… Also: this ETH is not the crypto one. Do you think P vs NP falls to AI next?
⏱️ Jump to a part
0:00 Two AIs, same day, same conjecture
0:53 3-SAT: a switch puzzle
2:25 The Exponential Time Hypothesis: a floor under the base
4:15 October 16: the floor breaks
6:22 Lean proofs, and who actually decides
8:02 Does P = NP? Are passwords safe?
9:49 10 to the ten thousand
10:42 What really cracked: the foundation under 800 papers
11:55 Takeaways
🍿 Watch next
OpenAI’s “alien math”: what did AI actually solve?
Can AI crack Napoleon’s cipher?
Is AI improving itself? I let GPT-6 Astra do AI research
✨ You might also like
I Asked GPT-6 and Claude for a Stat
GPT-6.1 Terra Just Dropped
📷 Credits
Paper excerpts: OpenAI, “The Exponential Time Hypothesis is False”
Paper excerpts and Lean statement: Claude (Anthropic), “A 2^o(n) Time Algorithm for 3-SAT”
#ExponentialTimeHypothesis #PvsNP #AImath
💚 Follow Mia for AI, explained in plain English.
source
@adamrussell658 you seem to take it for granted that it wasn’t. What makes you think that?
If a proof cannot be verified by a human it is worthless and probably wrong. Anything an AI does should be verified by a human. This is not just a bias against machines, but any proof can have errors. The logic of the proof must be checked from start to finish.