Generative AI as a Mirror of Human Cognition: Psychological Primitives in Language Model Emergence — LessWrong