Just fantastic technology all around. Absolutely no worry where this is all going to go.

Is it just me? Am I using this wrong or am I asking questions that are too hard?

Here’s an example of a hallucination that happened while explaining away another hallucination I called it out on. I rarely have experiences other than these.

@mwichary fwiw, "calling out" hallucinations is useless. It's just going to generate a typical-looking apology for hallucinating and then hallucinate something else. This is another cognitive trap of the chatbot metaphor -- you can't catch an LLM in a lie like Susan Calvin short-circuiting a positronic brain, because there's nothing in there.
@aburka Yeah, but it gives me a bit of satisfaction. 😂
@mwichary can't argue with that