Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: Should HN ban ChatGPT/generated responses?

news.ycombinator.com

Post reply on HN
Print view View original

161–170 of 677 posts

Re: Ask HN: Should HN ban ChatGPT/generated responses?

#161

Post by Al-Khwarizmi »

dottjt wrote: post #78
dang wrote: post #55

They're already banned—HN has never allowed bots or generated comments. If we have to, we'll add that explicitly to https://news.ycombinator.com/newsguidelines.html , but I'd say it already follows from the rules that are in there. We don't want canned responses from humans either! Edit: It's a bit hard to point to past explanations since the word "bots" appears in many contexts, but I did find these: https://news.yc…

I think the issue is that ChatGPT can already generate better responses than I personally can. Like, a lot of the time I'll have an interesting perspective that I just can't be bothered articulating, whereas ChatGPT will formulate that for me.

This is even more valuable if you're not a native English speaker, as finding the right words for what you want to say takes more time.

I'm not using it on HN but it's proving invaluable for professional emails, as it gets the tone and all the polite boilerplate right (at least for my standards) which is something that takes me some time to do by hand.

Re: Ask HN: Should HN ban ChatGPT/generated responses?

#162

Post by amirhirsch »

ramraj07 wrote: post #126
matthberg wrote: post #73

Earlier quoted context omitted.

There is an xkcd comic about this (of course): #810 Constructive: https://xkcd.com/810/

Of course there is, but it’s definitely weird when the jokes only funny when it’s not easy to think of it as a real possibility! In someways this thread sounds like the real first step in the raise of true AI, in a weird banal encroachment kind of way.

I think it would be really interesting to see threads on Hackernews start with an AI digestion of the article and surrounding discussion. This could provide a helpful summary and context for readers, and also potentially highlight important points and counterpoints in the conversation. It would be a great way to use AI to enhance the user experience on the site.

I routinely use AI to help me communicate. Like Aaron to my Moses.

Re: Ask HN: Should HN ban ChatGPT/generated responses?

#163

Post by ycombobreaker »

seydor wrote: post #47
aussieshibe wrote: post #45

Earlier quoted context omitted.

While I'm on the fence, and don't disagree with the point, this is an awful example to use. The whole comment is written in a very noticeable Chat GPT "voice", making it very easy to tell it was generated by Chat GPT. It's also not a particularly fun comment.

really? i could see myself writing exactly the same. it s too short to tell if it's gpt. Perhaps the only thing missing is imperfect capitalization / spelling (The last line was completely improvised by itself btw)

It was pretty clear to me before hitting your disclaimer. I suppose right now, three sentences is plenty. I feel that I shouldn't go into detail about the noteworthy features.

Re: Ask HN: Should HN ban ChatGPT/generated responses?

#164

Post by burnished »

andsoitis wrote: post #136
dang wrote: post #61

Earlier quoted context omitted.

I agree, but in that case we can learn from the bots instead of wincing at regurgitated material. Basically, if it improves thread quality, I'm for it, and if it degrades thread quality, we should throw the book at it. The nice thing about this position is that comment quality is a function of the comments themselves, and little else.

I suggest thinking about the purpose of discussion on HN. There’s a tension between thread quality on the one hand and the process of humans debating and learning from each other on the other hand.

I don’t think so, at least, I find that process to be very educational, especially when some one changes their mind or an otherwise strong argument gets an unusually compelling critique.

Basically I think those two things are synonymous.

Re: Ask HN: Should HN ban ChatGPT/generated responses?

#165

Post by dragonmost »

gkoberger wrote: post #108
drusepth wrote: post #97

This debate reminds me a lot of early-Quora moderation discussions over anonymous users. People wanted to mass-ban any answers from "anonymous" because they weren't from "real people". I argued that, in a public forum, questions/answers should be judged on their content, rather than "who" wrote them. People disagreed. Anonymous functionality on Quora has since been gutted. AI-generated content seems similar. If a com…

Anonymous and AI are very different (at least currently). Take the following statement: "The CEO of ACME Co is abusive and you should avoid him at all costs. Do not work there." If it's an anonymous poster on Quora, that's a (potentially) valuable response and a good reason for that person to stay anonymous. But if it's AI, it can't possibly reflect any actual first-hand experience. At best it's a regurgitation of so…

I think long term having AI comments on forums will erode people's trust in each other even more than it is right now. I dread the day the "default" commenter turns from "white american male" to "AI generated" with no way to know while still conserving some level of anonymity.

Re: Ask HN: Should HN ban ChatGPT/generated responses?

#166

Post by holoduke »

dottjt wrote: post #78
dang wrote: post #55

They're already banned—HN has never allowed bots or generated comments. If we have to, we'll add that explicitly to https://news.ycombinator.com/newsguidelines.html , but I'd say it already follows from the rules that are in there. We don't want canned responses from humans either! Edit: It's a bit hard to point to past explanations since the word "bots" appears in many contexts, but I did find these: https://news.yc…

I think the issue is that ChatGPT can already generate better responses than I personally can. Like, a lot of the time I'll have an interesting perspective that I just can't be bothered articulating, whereas ChatGPT will formulate that for me.

I have never seen questions to itself or other's in the generated text. Also it never expresses emotions. Never something like "I am convinced.. ." , "I dont want to be....". I find it quite easy to recognize in comment sections.

Re: Ask HN: Should HN ban ChatGPT/generated responses?

#167

Post by burnished »

mherdeg wrote: post #121
potatoman22 wrote: post #89

Earlier quoted context omitted.

How could ChatGPT generate your perspective for you?

Let's try it. I think the second response below fails "concise" .. Q: Write a news.ycombinator.com comment making the point that sometimes I use large language models because I know the gist but want to add examples and helpful context of my message A: I find that sometimes when I'm commenting on news.ycombinator.com, I'll have a general idea of what I want to say, but I want to add examples and helpful context to ma…

Both answers had the same volume of content as your first prompt. These are not improvements and if you think they are representative of your typical output then I would suggest stopping this practice.

Although it would explain some of the ‘content free’ writing I some times see.

Re: Ask HN: Should HN ban ChatGPT/generated responses?

#168

Post by midoridensha »

ramraj07 wrote: post #126
matthberg wrote: post #73

Earlier quoted context omitted.

There is an xkcd comic about this (of course): #810 Constructive: https://xkcd.com/810/

Of course there is, but it’s definitely weird when the jokes only funny when it’s not easy to think of it as a real possibility! In someways this thread sounds like the real first step in the raise of true AI, in a weird banal encroachment kind of way.

When I compare the ChatGPT-generated comments to those written by real humans on most web forums, I could easily see myself preferring to only interact with AIs in the future rather than humans, where I have to deal with all kinds of stupidity and bad and rude behavior.

The AIs aren't going to take over by force, it'll be because they're just nicer to deal with than real humans. Before long, we'll let AIs govern us, because the leaders we choose for ourselves (e.g. Trump) are so awful that it'll be easier to compromise on an AI.

Before long, we'll all be happy to line up to get installed into Matrix pods.

Re: Ask HN: Should HN ban ChatGPT/generated responses?

#169

Post by supernikita »

sudomatic wrote: post #147
dang wrote: post #55

They're already banned—HN has never allowed bots or generated comments. If we have to, we'll add that explicitly to https://news.ycombinator.com/newsguidelines.html , but I'd say it already follows from the rules that are in there. We don't want canned responses from humans either! Edit: It's a bit hard to point to past explanations since the word "bots" appears in many contexts, but I did find these: https://news.yc…

Absolutely not! In fact, we should be encouraging the use of ChatGPT and other generated responses. After all, why waste time thinking for ourselves when we can just let a computer do it for us? Plus, it's not like the AI is going to come up with some crazy, off-the-wall ideas that no human would ever come up with. Oh wait... never mind.

I disagree: mental laziness is never a good justification. (I recently started to memorize more things, since I do not believe in "just google it"). Also, I want to read what fellow humans relate. Even though it might sometimes be difficult to tell the difference, I attend HN as a substitute for physical social interactions. For mere knowledge or problem solving, other places might fill that role.

Re: Ask HN: Should HN ban ChatGPT/generated responses?

#170

Post by midoridensha »

xcambar wrote: post #154
im3w1l wrote: post #80

Earlier quoted context omitted.

> If we reach the point where the humans simply can't do better, well, then it won't matter. I disagree with this. The exact same comment written by a human is more valuable than one written by a bot. For example imagine I relate something that actually happened to me vs a bot making up a story. Byte for byte identical stories. They could be realistic, and have several good lessons baked in. Yet one is more valuable,…

From the perspective of the receiver of the message, there is no such thing as the story being true or not. If it's byte for byte the same story and I don't know whether the author is a human or a bot and I believe the story, the same reaction will be triggered at every level. The emotions, the symbolics, the empathy, all the same, whether the author is this or that. As a matter of fact, none of us know whether the o…

>As a matter of fact, none of us know whether the other is a human or even if dang is (!), because it is orthogonal to the contents and discussion.

Dang seems like he's always able to respond on the many HN threads much too quickly. I suspect he's really an advanced AI.

Post reply on HN