Again, the more I know about it the more complex it seems to me, which is the opposite of what you're saying. So many ways to exert control over it between fine-tuning, RAG, and loop/graph based agentic flows.
I'm not convinced anyone is an expert on it yet relative to what AI dev baseline will be in a year or two.
It's gonna seem complex if you don't know the subject and don't know how to do the thing without an LLM.
But it's just trying and trying until you get what you want