“of per flop as samples per flop times bits per sample. And…” — Dwarkesh Podcast at 2:13:11
“of per flop as samples per flop times bits per sample. And what I just mentioned a second ago is that the samples per flop go down as RL becomes more and more long horizon. But at least this kind of naive RL is also terrible from a bits per sample perspective. And here's what I mean, at least”
0:00
2:37:29