ʝօռ ʍօʀֆɛ: Well, it does matter because... *glances around suspiciously* ...I wouldn't put it past the techbros to spread a bunch of fake bad answers which immediately get debunked so that they can claim *all* the bad answers are fake. So folks who want to end this nonsense need to not be spreading fakes.

@NBCNews: Social media has been buzzing with examples of Google’s new, “experimental” AI tool going awry. A Google spokesperson said the company believes users are deliberately attempting to trip up the technology with uncommon questions. https://t.co/Wj9iuKzzHk

Yeah, “how do I deal with depression?” is a really niche, obscure question, so you can see how the AI tool inevitably had little information to lead its response with beyond “One Reddit user suggests jumping off the Golden Gate Bridge.”

That one was actually faked, though. Still, there's a whole slew of perfectly normal questions it's fucking up.

It really doesn't matter if it's faked or not, a LLM will inevitably feed incorrect information if you start waking it for facts

Well, it does matter because... *glances around suspiciously* ...I wouldn't put it past the techbros to spread a bunch of fake bad answers which immediately get debunked so that they can claim *all* the bad answers are fake. So folks who want to end this nonsense need to not be spreading fakes.

Post