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

Don't post generated/AI-edited comments. HN is for conversation between humans

news.ycombinator.com

Post reply on HN
Print view View original

521–530 of 1001 posts

Re: Don't post generated/AI-edited comments. HN is for conversation between humans

#521

Post by lurkshark »

dom96 wrote: post #489
dang wrote: post #386

The rule has been around for years, but only in case law, i.e. moderation comments ( https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que... ). What's new is that we promoted it to the guidelines. Fortunately I found some things we could cut as well, so https://news.ycombinator.com/newsguidelines.html actually got shorter. --- Edit: here are the bits I cut: Videos of pratfalls or disasters, or cute animal pic…

I’m really curious how this will go. I have a suspicion that we will see more and more accounts all over the internet being controlled by AI agents and no amount of moderation will be able to stop it.

I assume we’ll end up with proof-of-identity attestation as a part of public posting (e.g. Worldcoin) which doesn’t necessarily solve the issue but will at least identify patterns more likely to be LLMs (e.g. a firehose of posts at all hours of the day from one identity). Then we’ll enter the dystopia of mandated real identity on the internet

Re: Don't post generated/AI-edited comments. HN is for conversation between humans

#522

Post by SchemaLoad »

semiquaver wrote: post #476
jedberg wrote: post #359

I'm absolutely 100% for this policy. My only caution is that good writers and LLMs look very similar, because LLMs were trained on a corpus of good writers. Good writers use semicolons and em-dashes. Sometimes we used bulleted lists or Oxford commas. So we should make sure to follow that other HN rule, and assume the person on the other end is a good faith actor, and be cautious about accusing someone of using AI. (I…

Good writers are often good in recognizably unique ways. To the extent that LLMs produce “good writing,” which I happen to think they mostly do, they tend to overuse specific devices which give their writing a quality that most people are already sick of.

You can tell good writers from LLMs because good writers post comments that mean something, that add to the conversation, that bring in personal experiences. While LLM comments just summarize the article and end with some engagement call to action like "Curious to hear what others think"

Re: Don't post generated/AI-edited comments. HN is for conversation between humans

#523

Post by loeg »

abtinf wrote: post #67

Good. This helps establish it in the HN culture. That’s the purpose of guidelines. 99% of rule enforcement, both IRL and online, comes down to individuals accepting the culture. Rules aren’t really for adversaries, they are for ordinary situations. Adversaries are dealt with differently.

I mostly agree, although we've seen big shifts in the culture towards rule-deviating norms over time. Look at the guidelines for ideological battles or throwaway accounts, for example. And, as always:

> Comments should get more thoughtful and substantive, not less, as a topic gets more divisive.

Re: Don't post generated/AI-edited comments. HN is for conversation between humans

#524

Post by alexjplant »

jedberg wrote: post #359

I'm absolutely 100% for this policy. My only caution is that good writers and LLMs look very similar, because LLMs were trained on a corpus of good writers. Good writers use semicolons and em-dashes. Sometimes we used bulleted lists or Oxford commas. So we should make sure to follow that other HN rule, and assume the person on the other end is a good faith actor, and be cautious about accusing someone of using AI. (I…

> Good writers use semicolons and em-dashes

I use semicolons a lot. If this is the nouveau tell du jour for LLMs then I'm in trouble.

Re: Don't post generated/AI-edited comments. HN is for conversation between humans

#525

Post by zahlman »

lisp2240 wrote: post #76

I want a social network that goes beyond banning bots and also bans the half of the population that doesn’t have an inner monologue.

Such a ban is impractical, but we can maintain an environment where such people are simply not interested in participating.

To my understanding, that has a lot to do with why the site remains so low-tech (and avoids, in large part, the appearance of a "social network").

Re: Don't post generated/AI-edited comments. HN is for conversation between humans

#526

Post by jjk166 »

djeastm wrote: post #493
jedberg wrote: post #359

I'm absolutely 100% for this policy. My only caution is that good writers and LLMs look very similar, because LLMs were trained on a corpus of good writers. Good writers use semicolons and em-dashes. Sometimes we used bulleted lists or Oxford commas. So we should make sure to follow that other HN rule, and assume the person on the other end is a good faith actor, and be cautious about accusing someone of using AI. (I…

>(I've been accused multiple times of being an AI after writing long well written comments 100% by hand) Perhaps always be sure to say something especially timely, original or insightful that an LLM can't have come up with.

Nah, just write not good like rest of we

Re: Don't post generated/AI-edited comments. HN is for conversation between humans

#527

Post by wasmitnetzen »

iammjm wrote: post #15

I believe the issue of proving who is and who isn't really human on the Internet will be a really important issue in the coming years, especially without sacrificing people's right to privacy and anonymity in the process.

We will just have to fucking swear all the time. The corporate-speak LLM won't do that.

Re: Don't post generated/AI-edited comments. HN is for conversation between humans

#528

Post by ddtaylor »

This is a welcome change and do will update Ethos [1] in the future with an AI sentiment score. I created a separate project called LLaMaudit [2] that attempts to detect if an LLM was used to generate text, but it needs to be improved.

[1]: https//ethos.devrupt.io [2]: https://github.com/devrupt-io/LLaMAudit

Re: Don't post generated/AI-edited comments. HN is for conversation between humans

#529

Post by rlt »

safog wrote: post #80
iammjm wrote: post #15

I believe the issue of proving who is and who isn't really human on the Internet will be a really important issue in the coming years, especially without sacrificing people's right to privacy and anonymity in the process.

I hope I'm wrong but I don't think a privacy friendly alternative is going to exist. It's going to go the way of show me your drivers license to use my site.

I feel like we need a distributed system/protocol that allows people to have pseudonyms not linked to their real identity, but with a shared reputation/trust score, so if you’re a bad actor using a pseudonym your real identity and all your other sock puppets are penalized too.

I know very little about this but sense that some combination of buzzwords like homomorphic encryption, zk-snarks, and yes, blockchains could be useful.

Of course this would present problems if any of your identities were ever compromised and your reputation destroyed.

Re: Don't post generated/AI-edited comments. HN is for conversation between humans

#530

Post by jmull »

GMoromisato wrote: post #272

I'm here to read what actual humans think. If I wanted to read what an LLM thinks, I could just ask it. But here's where it gets tricky: Do I prefer low-effort, off-the-top-of-my-head reactions, as long as it is human? Or do I want an insightful, well-thought-out response, even if it is LLM-enhanced? Am I here to read authentic humans because I value authenticity for its own sake (like preferring Champagne instead of…

If the goal is to read what actual humans think, it's hard to see how an LLM filter can do anything but obscure and degrade the content.

LLMs, as we know them, express things using the patterns they've been developed to prefer. There's a flattening, genericizing effect built in.

If there are people who find an LLM filter to be an enhancement, they can run everything through their favorite LLM themselves.

Post reply on HN