> I'm quite sure that LLM style things will be smart enough to generate things like complicated decision trees and state machines from human explanations, plus

> I'm quite sure that LLM style things will be smart enough to generate things like complicated decision trees and state machines from human explanations, plus test suites and explaining which cases have ambiguity and/or missing. I'm not.