@bremen79 @AlexShtf Sometimes I do this (tell people to look at new stuff). Just a way to shake things up a little. But looking at old stuff is also good.
@AlexShtf Well, he actually complained that papers are COLT, while being great on their own, they keep focusing on "old" topics (e.g., bandits, PAC learning) that are not opening any new frontier
No Digg Deeper questions have been answered for this story yet.
@bremen79 @EdwardRaffML @AlexShtf I always hated this. Complexity on its own is not a merit. (When I started my career if it was not in a Banach space, it was not worth doing it. Now, I am very content with good old fashioned Euclidean spaces. Plenty of puzzles left there, too.)
@EdwardRaffML @AlexShtf Yes they did. People discussed if this should change given that now LLMs are able to produce very complex proofs. In the short term, I doubt this will change, it is really part of the COLT DNA.