Ask HN: Should I answer “Ask HN” questions using ChatGPT?

[pedrodelfino]

[jstx1]

No, I made a thread about this when ChatGPT came out - https://news.ycombinator.com/item?id=33891538

[anonym29 → jstx1]

100% this. There's enough of this crap on the web as-is. If you've ever restricted your search engine results to Reddit, you understand why this is a problem.

Let's not s*** where we eat, folks.

[catchnear4321]

The user didn’t go to ChatGPT.

They didn’t title it “Ask ChatGPT.”

That being the case, what would you hope to accomplish in doing so?

[pedrodelfino → catchnear4321]

Some answers from ChatGPT just feel better than some HN comments.

[catchnear4321 → pedrodelfino]

wouldn’t it have saved you time to ask chatgpt instead?

if you would consider using it to answer others, why not to answer yourself?

your question puts your question into question.

[smoldesu]

No, unless the question itself is AI-generated or specifically calls for AI response.

[sharemywin → smoldesu]

or you're arguing about it's capabilities or lack there of.

[smoldesu → sharemywin]

It's frankly kinda annoying then, too. I don't doubt you can get the AI to model the text you want -- with enough prompt engineering, anything is possible. If your anecdote is about ChatGPT generating text, it's probably not as novel as you think it is.

[sharemywin → smoldesu]

so is that true for other technologies too? Somebody posts, Oh you can't do this on a mac only on a pc. people should just let it go?

[smoldesu → sharemywin]

If someone says something wrong, you can simply refute them. "This also works on Mac, you're wrong about that."

If they challenge you after that, then you should use evidence to support your claim. In the realm of ChatGPT, individual attestation doesn't really push the field forward.