Nifty, probably one of the simpler (and more skill appropriate) LLM uses, but I do wonder if there’ll be enough transparency to watch for code fuckups. Could definitely see a string of research papers based off of faulty LLM code in the future.