Are you able to help the audience understand / find guidance as to, from first principles, why these benchmarks were chosen for your index in general?
Curious to see how the AA team thinks of benchmarks, public and private, and their relevance to a model’s intelligence. E.g the correlation factor. maybe I’m missing a think piece
@JaredKubin $ant may have solved running inference on cheaper / easier to acquire gpus than what’s in spacex datacenters… my only thesis for them moving off, but they must have a plan based off their internal frontier models & time to run inference on random gpus
@AlexBores Does Bernie even know what ai stands for? Why does an 80 year olds credibility for anything technology related matter? He’s been a hater for his entire life, all he does is chat shit, never suggests anything economy reinforcing
Invalidates your post for me