[noodlesUK]

I think that for low-risk classification tasks and similar, something like an LLM is a great tool, and I can absolutely see it being extremely useful for intelligence work where sifting through stuff is very hard. However, I would not at all trust AI to make actually important decisions independently.