DanielFilan comments on On the Dwarkesh/​Chollet Podcast, and the cruxes of scaling to AGI

[DanielFilan]

Since nobody else has responded, my best guess would be “conditional probability table”.