Charlie Steiner comments on AI models inherently alter “human values.” So, alignment-based AI safety approaches must better account for value drift

[Charlie Steiner]

The real chad move is to put “TL;DR: See above^” for every section.