@thedansimonson @inthehands But the problem there is conflating "generative AI" with all of machine learning, no? It is quite possible to build reliable (safety critical) software systems that solve hard problems using machine learning AND do not "hallucinate" anything. But there is no known way to do it cheaply.