General Discussion

snot

(11,986 posts)

11. Joe, can we at least agree that

Wed Jul 15, 2026, 02:19 PM

19 hrs ago

(1) LLMs don't really "understand" queries or responses in the way a human does, and

(2) their responses are often erroneous or misleading?

Just last night I spent an hour on the phone with a sales rep who wanted to answer my question about a product based on a response from Google Gemini. While the AI's answer sounded plausible, and the AI did not indicate any uncertainty about it, when we tried to identify a source for the response, neither the AI nor we could find anything to verify it, even on the manufacturer's website.

Edit history

Please sign in to view edit histories.

Recommendations

0 members have recommended this reply (displayed in chronological order):

11 replies

= new reply since forum marked as read

Highlight:

AI Error Rates 2026: How Often AI Gets Facts Wrong -- And What It Costs [View all] snot Sunday OP

"AI" does not " hallucinate"... Ohio Joe Sunday #1

I get you-- snot Sunday #4

A few more things... Ohio Joe Sunday #6

I believe LLMs are a subset of AIs. There are also AIs designed for non-linguistic purposes, snot Sunday #7

You have zero idea of what you are talking about... Ohio Joe Sunday #8

Perhaps you could direct me to what you'd consider to be a better description snot Sunday #9

Let me tell you a story... Ohio Joe Sunday #10

Joe, can we at least agree that snot 19 hrs ago #11

That's a lot of errors questionseverything Sunday #2

Another way to look at it snot Sunday #5

I use AI with medical things, but I always doublecheck everything it says, since can give GPV Sunday #3