Comment by gilgamesh-OG
5 days ago
I don't know if you can "fix" hallucinations without changing the fundamental architecture. The other factor in this article is that prior to the AI summary at the top, Google could simply state that it was an error on the part of the website owner. Now it is being held liable for whatever the summary states - even if it's more accurate, it can still be wrong enough times to be expensive.
> I don't know if you can "fix" hallucinations without changing the fundamental architecture.
Exactly. "hallucinations" are not some special case. They are the LLM working as designed.
They don’t need to be eliminated or “fixed”, they just need to be less frequent than human error.