Large Language Models and the Rationalist Empiricist Debate