Hacker News Reader

Comment by nomilk | Hacker News Reader

[flagged]

Honest question, genuinely curious, I'm not picking on you; what's the point of pasting a ChatGPT response to a question? I've seen colleagues do this on internal mailing lists, even some of my friends do this on our group chats.

I really don't get it. If I wanted the opinion of a language model, I'd ask a language model (and I do semi-frequently for certain questions). Getting ChatGPT output from an human rubs me the wrong way, like getting a LMGTFY link.

But I assume both in your case and my friend's that the intent isn't to be passive aggressive. Help me understand?

I've seen this ~10 times today - They are all phrased the same way too. This person seems like an actual person and not a bot, so IDK. It's really, really annoying. If we wanted to, we could ask chatgpt ourselves.

Clearly they all want to welcome their AI overlords.

Reason for posting: I was surprised at how economical (in terms of words and, most importantly, reader time) GPT was compared to a human in answering the same question, without sacrificing accuracy.

Thanks for pointing out that it's distracting; I hadn't considered that before posting.

the output of LLMs is not novel to this audience

Well, the original post isn't about how particularly efficient ChatGPT is at stripping all of the character and large amounts of information from a perfectly good source. The original post is about railway ballast, and as such, this comment is completely and utterly irrelevant.

I assume they (and your co-worker) are just sharing something they thought was interesting and wanted to save people time from asking themselves. Like finding a source and pasting it even though everyone could google and find it themselves.

I know you might have done this for the first time, but there’s a response like this in every HN thread now. Please stop doing it.

It’s fine to discuss GPT in GPT threads (there’s no shortage of those). But this adds nothing to the discussion of the actual topic: The fact that it may or may not be hallucinated means that others would have to fact check the post, which means that in terms of net time, this adds negative value.

I’m not trying to be harsh, it’s not personally about you. I’m just hoping that a trend that significantly degrades the information value of HN doesn’t continue.

Please don’t start posting LLM summaries. If anyone wants this summary they can query chatGPT themselves. It’s just noise.

If you consider this article lengthy you should read a few books to recalibrate your attention span.

Announcing to everyone that you didn't bother reading the article because you don't have time, but then wasting your own time playing around with ChatGTP, then wasting everyone else's time by posting something everyone who actually did bother reading the article already knows, is hypocritical and disrespectful of everyone's time, and adds nothing to the conversation.

If you didn't bother reading the article, then simply don't post anything, especially not an announcement you didn't read the article.

I’m pretty sure the article is GPT generated too.

I asked ChatGPT to write an HN comment thread on this topic and posted the whole thing here. - Nick

Even GPT wouldn't use "extremely" that many times.

Reply on news.ycombinator.com