Christian Szegedy
@ChrSzegedy
#deeplearning#, #ai# research scientist. Opinions are mine.
3.2K Following 47.4K Followers
Up until last year, all these results would be papers separated by years, waiting for peer review for months, communicated in arxiv papers first, published after a year in some journal.
Now, all that research using AI and formal tools can be compressed into hours.
Show more
I wonder what @GaryMarcus thinks of this bet now. It was in a thread originated by @MCHammer that seems to be deleted, but I saved the screenshots back then.
Mathematical reasoning will transfer to any area that requires reasoning and building intuition over a complex domain. It might be that AI will not have all the necessary data to make the right conclusions in some cases, but it will be filling in all the necessary holes in data with high sample efficiency, just like humans.
Superhuman intelligence is here, and the same model that solved all these problems will be great at all other scientific and reasoning-bound tasks.
Show more
Reading all those recent beautiful essays by Terry Tao, Kevin Buzzard and @wtgowers, I could not resist, but write up my deeply personal perspective on how I perceive and see things:
Show more
AI, as a whole, has reached a crossover point by April/May 2026, meaning that it has reached a level of 90th percentile of the reasearch level of working mathematicians.
The only thing that keeps us from reaching the level of *any* mathematicians is the reluctance of using formal methods in RL loops by the big labs.
At the latest, we will reach super-human level in mathematics by mid-next year.
Show more
