I feel like there's two types of LLM users. Those that understand it's limitations, and those that ask it to solve a millennium problem on the first try.