Comment by JohnFen
6 days ago
I find such replies to be worthless wastes of space on par with "let me google that for you" replies. If I want to know what genAI has to say about something, I can just ask it myself. I'm more interested in what the commenter has to say.
But I don't know that we need any sort of official ban against them. This community is pretty good about downvoting unhelpful comments, and there is a whole spectrum of unhelpful comments that have nothing to do with genAI. It seems impractical to overtly list them all.
There is friction to asking AI yourself. And a comment typically means that "I found the AI answer insightful enough to share".
Unfortunately it's easier to train an AI to be convincing than to be correct, so it can look insightful before it's true.
Like horoscopes, only they're not actually that bad so roll a D20 and on a set of numbers known only to the DM (and varying with domain and task length) you get a textbook answer and on the rest you get convincing nonsense.
> Unfortunately it's easier to train an AI to be convincing than to be correct, so it can look insightful before it's true.
This nails it. This is the fundamental problem with using AI material. You are outsourcing thinking in a way where the response is likely to look very correct without any actual logic or connection to truth.
The problem is that the AI answer could just be wrong, and there’s another step required to validate what it spit out. Sharing the conversation without fact checking it just adds noise.
"Friciton" in this case is just plain old laziness and I don't think that it should be encouraged.
Then state your understanding of what it said in your own words, maybe you’ll realize it’s bunk mid-sentence.
I'd rather you attribute your facts to an LLM vs. rephrase a hallucination that sounds right.