@Frances01896069@nihilunbounded I think Alpoege also depended on the QRH of OpenAI. OpenAI also realized the spirit of Langlands program behind QRH and they believed they can do something with it. That's why they hired an expert in Langlands. Unfortunately they were too slow.
https://t.co/ZEIDpRvmNU
I made this short video, in collaboration with Claude, about the work I have done with my colleagues on the dark sector of the universe. It may be the first concrete evidence for string theory!
Instead of just giving Fields medals to AI, how about we come up with a new award:
Best Button Pusher of the Year Award!
I want to compete for it too (not being sarcastic here).
@plain_simon@plumsirawit@ElliotGlazer The proof of Bloch's conjecture was too specific to the p_g = 0 case on surfaces, there is no chance to extend it to resolve generalized Bloch's conjecture I believe.
@plumsirawit@ElliotGlazer Well the situation of Hodge Conjecture is very different, as the only progress is Hodge Conjecture for CM Abelian varieties. The 700 slops also didn't add much advance progress on it, which is totally opposite to RH situation.
Here comes the salesmen from AI labs telling you intelligence is not something you should care about, step outside, take a walk down the street, touch some grass and you reach the OpenAI general intelligence store selling intelligence in monthly subscription rate of $100.
@tonyyueyu@OpenAI@axiommathai So the MMP for terminal and canonical 4-folds has been complete since 2003. What stayed open was termination of log flips for klt/lc pairs, with partial results by Alexeev–Hacon–Kawamata, Birkar, Shokurov, Han–Liu, Chen–Tsakanikas and Moraga (arXiv 2018). 4/4
@tonyyueyu@OpenAI@axiommathai Existence: Shokurov (2003) proved existence of 4-fold log flips, the subject of "Flips for 3-folds and 4-folds", which has been cited. Also Shokurov proved the termination of 3-folds flips before Mori proved the existence of flips in 1988. 3/4
@nihilunbounded Interesting that in the four statements that need to be addressed to claim "solving" Yang-Mills, OpenAI's construct gave (1) and (2), but has not finished (3) and no where near (4). This is much different than "only mass gap left" statement from various sources.
@exmuslim_ayesa@rand_longevity If you understand algebraic geometry, then CM Abelian varieties case couldn't even count as 0.0001% progress. On the other hand, the conjecture is widely open even for the statement to be true for every dimensional 4 smooth varieties case.