Yeah, but it would have to reason about the thing it just halucinated. Or it would have to be somehow hard prompted. There will be more tools and code around LL

Yeah, but it would have to reason about the thing it just halucinated. Or it would have to be somehow hard prompted. There will be more tools and code around LLM, to make it behave like a human then people can imagine. They are trying to solve everything with LLMs. They have 0 agency.