ParentFull threadroot_axis·None of those arguments claim "LLMs could not possibly be good models of some cognitive capacity"View on HN