I think many cognitive scientists think language is a core part of human cognition, and that LLMs were a major advance in AI/cognitive science due to connecting AI algorithms with language production, but LLMs aren’t considered (or even aiming) to be accurate models of human cognition