I recall them saying they use models to write CUDA kernels and whatnot. Makes sense, and unsurprising that models are good at writing code.
replyBut I think calling this “automating AI research” is misleading. I’m not sure there’s evidence yet that they do creative research work. Even in mathematics, but they are finding counter-examples by intelligent brute-forcing. Not to downplay the results, as they are incredible, but this is one very specific kind of proof and not the most creative type, which arguably requires generalisation.
Quite the gamble.
reply> but I didn't think it'd solve maths problems
replyFinding counterexamples is low-hanging fruit, the automation of which isn't shocking.
It’s a bird! It’s a plane! It’s…AI skeptics moving the goalposts at light speed!!
replyWhat about finding the 1st known complex structure over S^6, proving Ehrhart’s volume conjecture, proving a sharp "density" bound on primitive sets conjectured by Erdos >60 years ago?
reply> Finding counterexamples is low-hanging fruit, the automation of which isn't shocking.
It's not good to be confidently wrong the way you're being.