gary marcus' objections have nothing to do with world models, I think they do. See: [https://garymarcus.substack.com/p/how-come-gpt-can-seem-so-brilliant](https://garymarcus.substack.com/p/how-come-gpt-can-seem-so-brilliant) . GPT and other LLMs don't are not grounded in the real world, so cannot form ... doesn't use the term there, but that's what he's getting at. Also, at [https://garymarcus.substack.com/p/how-new-are-yann-lecuns-new-ideas](https://garymarcus.substack.com/p/how-new-are-yann-lecuns-new-ideas) he says: >A large part of LeCun’s new manifesto is a well-motivated call

garymarcus.substack.com/p/is-microsoft-about-to-get-the-deal

Gary Marcus but he raises valid criticisms here in a very recent essay: [https://garymarcus.substack.com/p/how-not-to-test-gpt-3](https://garymarcus.substack.com/p/how-not-to-test-gpt-3) Certainly, there are even more impressive models to come. I believe firmly that, some day, human intelligence will