LLMs are not thinking for themselves, they are just allowing for the interrogation of training data based on a natural language query model.