gwern comments on [Linkpost] Applicability of scaling laws to vision encoding models

[gwern]

This is a paradigm I’ve dubbed “brain imitation learning” (links).