With AI making progress in settling mathematical conjectures, is a proof of P ≠ NP around the corner?
No, you need to respect the difficulty of P v NP.
Since today Bipartite Perfect Matching is in NC. The proof uses connections between coding theory and Hall's theorem. Presented at WACT 2026. Yay!!!
I dreamt that I proved something important using ChatGPT. This morning for some reason I can't find the proof in Chatty's history.
Some unknown theorem survives for another day.
P.S. Bill is being sarcastic. Or is he?
2/2
Bill posts about the sum-product conjecture, one of the last great conjectures disproven by humans.
blog.computationalco...
1/2
Conjectures inspire mathematicians and proofs can extinguish that. Budding mathematicians of my generation all dreamt of being the first to prove Fermat's Last Theorem. Not any more.
Though the best proofs lay a foundation that leads to new conjectures and inspirations.
American Express Traveler's Checks were the stablecoins of the 1970s: A store of fixed USD value, which required signatures at time of sale and use, and had a register of ownership to detect and prevent fraud.
I got by this "Given that you're interested in this problem, why won't you make an effort to solve it? We've seen some other open math problems gets off by AI surprising ways."
Didn't get a proof but at least it tried.
2/2
The 2026 Gödel Prize is awarded to Ilias Diakonikolas, Gautam Kamath, Daniel Kane, Jerry Li, Ankur Moitra, and Alistair Stewart for Robust Estimators in High Dimensions without the Computational Intractability
Paper: doi.org/10.1137/17M1...
Citation: www.sigact.org/prize...
