“Lak Ananth asserts that current Large Language Models (LLMs) do not "think" but are sophisticated probabilistic machines that generate legitimate-sounding answers based on their training data.”