Now, what this tells me is that LLM AIs can supply the knowledge of the moment for the prepared human person, and the person does not need to learn that knowledge, for a extremely large class of problems. Which is fantastic for capitalism and business in general, and very very bad for humanity in general.