alt Hacker NewsIs "loving math for maths sake" just about knowing the answers? I think one can love math for exactly the process and understanding that a several-thousand-line uncommented Lean proof denies. If a deity rearranged the stars to spell out "The Riemann hypothesis is false" for a night, would that be intellectually sufficient?
No that's not sufficient but that's not what's happening.
Why do we believe that we cannot train models which could explain the jargon in more human terms when current LLMs can perfectly explain the most complicated codebases?