logoalt Hacker News

throw7today at 12:13 AM0 repliesview on HN

"...the model may give different advice even when the underlying question is the same."

Isn't this the point of LLMs? If not it would be deterministic and that's not "new" and/or "exciting".