you're more polite than me, but that's essentially the same response as what i have to people citing chatGPT. i just say "ChatGPT told me that's wrong".

if somebody thinks that unverified LLM output is relevant to a conversation, i don't want to have to defend why it shouldn't be part of the conversion, i want to put the responsibility for justifying it back onto them.

that's a great idea, i'll use it in future, thanks! not sure whether i should use your post as reference, though ;)

i have to admit that i already used it :)

Then they start texting it to you in Apple Messages and expecting you to treat it with as much weight as their own words.