Christian Szegedy

@ChrSzegedy · 23 nov. 2023

I am not sure what Richard means here exactly. Even if we consider supervised training only, it could still achieve superhuman breadth and consistency. If we allow for RL and self-play, then AlphaZero is a clear counterexample.

Richard Socher@RichardSocher· 23 nov. 2023

Proposed Theorem: it’s impossible to acquire super humanity skills when relying purely on data created by humanity. Why do I say super humanity and not super human? Because for example a translation algorithm is already better than any single human in terms of how many

Elon Musk

@elonmusk

Unfortunately, us humans are much dumber than we think we are

16:50 · 24 novembre 2023 · 66,5 k vues

188

970

X →