[keithnz]

I think for LLM like Open AI, it wouldn't be about hitting the target but target selection. Target selection is probably the most likely thing that won't be accurate