The Failure of Language Models

The way I think of system two reasoning in human brain is that it is a semi-logical system operating on a fuzzy, not fully formal ontology. So one of the reasons why expert systems and logic programming has failed is because they really failed at making fuzzy Ontologies. This is what I think language models and general cognition engines do. They provide this latent, this common latent space which can be used for all kinds of different things. And as you become more high powered and like competent with structured thinking an awful lot of thinking becomes bottleneck through these bottlenecks. Not there's not complexity by their nature, which are very low dimensional - we're just talking about bits and pieces

Transcript

Play full episode

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.