If you want to actually understand the arguments about LLMs and world models—and why LeCun, Bengio, @RichardSSutton and I have landed in similar places—read my essay “Generative AI’s crippling and widespread failure to induce robust models of the world”