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

This thread made me think of education (as in schools). To paraphrase:

“Don’t post generated/AI-edited assignments. School is for conversation between humans”

AI can be a great tool for learning, but also can pollute or completely hijack the medium for human interaction and learning.

Having HN flooded with AI generated content will be sad as I like reading it, but losing that same fight at schools will be detrimental.

For quite a while, I like use LLM to refine and fix my grammar issue, but my colleagues and professors reminds me that it was way too obvious. They said they can tolerate some mistakes in my words, but no tolerance for AI generated content.

We've all heard the phrase "the sum of all human knowledge".

I've been feeling more and more that generative AI represents the average of all human knowledge. Which has its place. But a future in which all thought and creativity is averaged away is a bleak one. It's the heat death of thought.

I am 100% behind this. I've been browsing hackernews since I started in tech, it is the only forum i regularly browse, and partake in. Simply because the quality of submissions and conversations are so high. There has been more AI related articles this part year, and it only seems ramping. I personally haven't found the AI part of the comments as big of a deal but dang and tom might be doing more than I realize on that front.

Though I do wish we'd see less AI related posts on the front page, they simply aren't sparking curiosity, it is the same wrapped in a different format, a different person commenting on our struggles and wins with AI, the 10th software "rewritten" by an AI.

At this point there nearly should be a "tax" on category, as of this moment I count 8-10 related posts on the front page related to AI / LLMs. It is a hot field, but I come to hackernews, to partake in discussions about things that are interesting, and many of those just doesn't cut it, in my opinion.

I feel a little bit of irony in this post of a company/forum that is asking its users to not use AI while simultaneously trying to fund countless companies that are responsible for ruining the internet as we speak.

What a welcome post. The whole reason I come here is to get thoughtful input from smart people, and not what I could get myself from an LLM. While we are at it; Think your own thoughts as well :) I know how easy it is to "let it come up with a first draft" and not spend the real effort of thinking for yourself on questions, but you'll find it's a road to perdition if you let yourself slip into the habit. Thanks to all the humans still here!!

There should be a "flag as AI" link in addition to "flag" and then a setting for people to show flagged as AI. Once the flagged as AI reaches a certain threshold then it disappears unless you enable "Show AI".

Maybe once enough posts have been flagged like that then that corpus could be used to train an AI to automatically detect content generated by AI.

That would be cool.

Maybe the HN site wouldn't add this feature but if someone wrote a client then maybe it could be added there.

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've been accused multiple times of being an AI after writing long well written comments 100% by hand)

I use AI for the elements I feel are weak or unclear in the transcription. Sometimes I copy-paste a paragraph into ChatGPT or whatever, to ensure my (aging) thoughts are being communicated in a crystal clear manner. I cannot always point out why I think they are unclear or jumbled.

I don't feel this is an imposition on others. I think it's the opposite. It enhances signal by reducing nitpicking, spelling/grammar errors that might muddle intent, and reminds me of proper sentence structure.

Many of us are guilty of run-ons, fragments, overly large blocks of text[1] because it's closer to how people often converse, verbally. Posts on the internet are not casual conversation between humans. They are exchanges of ideas.

[1] This is a classic example where I had to go back and edit it to ensure it was readable. As you do self-review with any commit ^^

How about comments that include AI output if labeled?

Earlier today I remembered that there was a Supreme Court case I'd heard about 35 years ago that was relevant to on an ongoing HN discussion, but I could not remember the name of the case nor could I find it by Googling (Google kept finding later cases involving similar issues that were not relevant to what I was looking for).

I asked Perplexity and given my recollection and when I heard about the case it suggested a candidate and gave a summary. The summary matched my recollection and a quick look at the decision itself verified it had found the right case and did a good job summarizing it--probably better than I would have done.

I posted a cite to the case and a link to decision. I normally would have also linked to the Wikipedia article on the case since those usually have a good summary but there was no Wikipedia article for this one.

I though of pasting in Perplexity's summary, saying it was from Perplexity but that I had checked and it was a good summary.

Would that be OK or would that count as an AI written comment?

I have also considered, but not yet actually tried, running some of my comments through an AI for suggested improvements. I've noticed I have a tendency to do three things that I probably should do less of:

1. Run on sentences. (Maybe that's why of all the people in the 11th-100th spot on the karma list I have the highest ratio of words/karma, with 42+ words per karma point [1]).

2. Use too many commas.

3. Write "server" when I mean "serve". I think I add "r" to some other words ending in "e" too.

I was thinking those would be something an AI might be good at catching and suggesting minimal fixes for.

[1] https://news.ycombinator.com/item?id=46867167

I don't think I'm going to spend the time to paraphrase my worthwhile AI-applied work for such hypocritical rules.

So develop and fund and use AI but manually paraphrase things and don't cite AI?

It is best to cite a source and/or a method.

Do you think it is better to paraphrase and not cite AI?

I don't recall encountering posts on HN that I've wanted to flag as AI.

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.

This is a wonderful rule.

It also points out the need for AI writing tools that very strictly just:

1. Point out misspellings and typos.

2. Point our grammar mistakes, if they confuse the point.

3. Point out weaknesses of argument, without injecting their own reasoning.

I.e. help "prompt" humans to improve their writing, without doing the improvement for them.

In fact, I would like a reliable version of that approach for many types of tasks where my creativity or thought processes are the point, and quality-control feedback (but not assistance), is helpful.

This is a mode where models could push humans to work harder, think deeper, without enabling us to slack off.

Don’t be afraid to make grammar mistakes or misspell stuff. Others will understand. You’re a human after all. That’s okay to make mistakes and feel uncomfortable with that.

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 pictures.

It's implicit in submitting something that you think it's important.

Please don't complain that a submission is inappropriate. If a story is spam or off-topic, flag it.

If you flag, please don't also comment that you did.

I hate cutting any of pg's original language, which to me is classic, but as an editor he himself is relentless, and all of those bits—while still rules—no longer reflect risks to the site. I don't think we have to worry about cute animal pictures taking over HN.

The most telling sign of a human commenter is brevity.

Consequently, I hardly ever spend the time to write out long and detailed HN comments like I used to in the pre-LLM era. People nowadays have a much harder time believing that an Internet stranger is meticulously crafting a detailed and grammatically-airtight message to another Internet stranger without AI assistance.

Honest question, why were folks posting AI generated comments in the first place? There's such a high inertia to comment. I only comment when I have something to contribute OR find something incredibly interesting.

So I'm just baffled, why anyone was using AI to generate comments. Like what was the incentive driving the behavior?

I've written tens of thousands of lines of code, autogenerated documentation with LLMs and use AI Agents daily.

But when I argue on the internet, it's always a 100% me.

And if I get a wiff of LLM-speak from whoever I'm wrestling in the mud with at the moment, they'll instantly get an entry in my plonk-file. I can talk with ChatGPT on my own thank you very much, I don't need a human in between.

"But my <language> is bad... that's why I use LLMs"

So was mine when I started arguing with strangers on the internet. It's better now. Now I can argue in 3 different languages, almost 4 =)

Good addition but to be fair HN guidelines have become so quaint particularly as they are now rarely enforced or even acknowledged. E.g. "Eschew flamebait. Avoid generic tangents. Omit internet tropes. " And " Off-Topic: Most stories about politics, or crime, or sports, or celebrities, unless they're evidence of some interesting new phenomenon. If they'd cover it on TV news, it's probably off-topic. " These are violated every day without consequence.

No way to verify. Relying on the humans here to self censor has never worked in the history of man. But the idea in itself is good. HN is for human to human conversation.

I'm tickled pink to read this! I very much support this move. HN is one of the few internet forums I use. It'd be awful to see this riddled by bot spew.

This rule will atleast partly stem the danger of HN getting turned into what dang calls a "scorched earth" situation.

Now that it's in the rules, I hope we also see less of "your comment was obviously AI generated so I won't respond" (ironically, in a response comment).

If you suspect it to be a bot, flag it and move on! If it is indeed a bot and you comment that it's a bot, it doesn't care! If it is not a bot and you call it a bot, you may have offended someone. If it's a human using AI, I don't think a comment will make them change their ways. In any case though, I think it's a useless comment.

I also feel the frustration of the llm reverse-compression - when a whole article is generated from a single sentence. But when I post something edited by AI it is usually a result of a long back and forth of editing and revising. I guess I could post the whole conversation thread - but it would be very long.

Personally I would just like to read the best comments.

I have the feeling my gramatical errors from being ESL appear to be "tolerated" a lot more than a few years ago. By that I mean that it doesn't get called out as much as it used to be.

TIL: definition fulminate

fulminated, fulminating to explode with a loud noise; detonate. to issue denunciations or the like (usually followed byagainst ).

(Because “don’t fulminate” is the rule that follows the referenced one :) )

Agree, AI generated articles & comments provide little to none value other than the original prompt. Please just post the original prompt instead.

All the weak excuses posted here are just making me lean more towards a hardline policy. No I don't want to read a human-generated summary of your llm brainstorming session. No I don't want to read human-written text with wording changes suggested by an llm. No I don't want to read an excerpt from llm output even if you correctly attribute it.

I acknowledge this is partly just my personal bias, in some cases really not fair, and unenforceable anyway, but someone relying on llms just makes me feel like they have... bad taste in information curation, or something, and I'd rather just not interact with them at all.

I do too care about this but I say this in the reality in which we are. This reminds me of those signs "no shirt, no shoes, no service" except it's much worse, only sentient beings will actually care about it, while non-sentients will simply trample over the sign while token predicted laughter erupts from their token predicted sense of humor artifact.

Elon said it well, there must be some disincentive to do this.

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 sparkling wine)? Or do I value authentic human output because I expect it to be of higher quality?

I confess that it is a little of both. But it wouldn't surprise me if someday LLM-enhanced output becomes sufficiently superior to average human output that the choice to stick with authentic human output will be more painful.

"AI-edited comments" is a very interesting one. Where is the line between a spelling/grammar/tone checker like Grammarly, that at minimum use N-Grams behind the scenes, and something that is "AI" edited? What I am asking is, is "AI" in this context fully featured LLMs, or anything that improves communication via an automated system. I think many people have used these "advanced" spellcheckers for years before Chatgpt et al came on the scene.

I think "generated comments" is a pretty hard line in the sand, but "AI-edited" is anything but clear-cut.

PS - I think the idea behind these policies is positive and needed. I'm simply clarifying where it begins and ends.

The only question is is the entity interesting and/or correct. Those properties are in the eye of the beholder. If they're human or not is beside the point.

After all, no one knows I'm a dog.

My words:

This feels like don't buy at Walmart, support the local small shop. We passed the no return sign miles ago.

Gemini's:

This is like advocating for artisanal blacksmithing in the age of industrial steel. It sounds great in theory, but we passed the point of no return miles back.

Yeah, we can tell the difference :)

This rule is very important. Like many of the other rules, it is open to interpretation, but it is a line in the sand that defines allowable behaviour and disallowable behaviour.

This rule will have an effect on the behaviour of the 'good players', and make the 'bad players' a lot easier to spot. Moderation needs this. I see this as stopping a race-to-the-bottom on value extraction from HN as a platform.

Absolutely love this. If people are relying on AI for a 30-45 word comment, I don’t want to waste my time reading it. And everyone using AI for discussions will end up coming to the same conclusion. Use your own ideas !

Since we now face a threat of large-scale de-anonymization, a reasonable countermeasure might be using AI to make one's writing style less personally identifying, in order to try and retain some pseudonymity.

    https://simonlermen.substack.com/p/large-scale-online-deanonymization
    https://news.ycombinator.com/item?id=47139716

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.

While I understand the sentiment, it ignores many people have English as a second language, many people are dyslexic and have dysgraphia. AI is a great assistant. A good approach will be to encourage people to develop their thinking than use the AI tools.

Does that extend to generated/AI-edited articles? I don't see why the same rationale wouldn't apply.

It’s almost certain that this exact thread is currently being used to train comment bots.

Not sure I agree with the AI edited comments. Using AI to improve the readability and clarity is fine. Sometimes a well structured comment is much better than a braindump that reads like ramblings. And AI is quite good at it (and probably will get better). To make the point, here is how this comment would have looked if edited:

"I don't fully agree with banning AI-edited comments. Using AI to improve readability and clarity is a reasonable thing to do. A well-structured comment is often much better than a braindump that reads like rambling. AI is quite good at this, and it will probably get better. To illustrate the point, here is how this comment would have looked if edited"

Translation is a form of AI-edition.

Language translation is the origin of (the current wave of) AI and its killer app. English is not the main language of the world, and translation opens us up to a huge pool of interesting thinkers.

I'm a native speaker in a foreign language, but out of practice except of a weekly family call. I recently had to write a somewhat technical email to my family, and found it easier to write it in (my more practiced) english and have AI translate it, than write it in the target language myself. Of course, in my case I was able to verify that the output conveyed the meaning I intended, because I am fluent in the target language.

As with the rise of GenAI, I've also noticed a rise of translated messages. It's usually hard to tell the difference, except by looking at the commenter's history (on other subreddits, impossible on HN).

I understand the original frustration with GenAI comments and reactionary response. I'm sorry that we're excluding what could be a large pool of interesting people because we can't tell the difference.

My question is, and this is genuinely a question: Do you think YC-backed companies would have respected this guideline if it was posted on some other website they wanted to operate in?

I wish more people would filter their comments through AI. It has so many benefits. If you're being emotional, it can detect that and rewrite your comment to be less confrontational and more constructive. If you're positing a position out of ignorance or as an armchair expert, it can verify your claims before posting. Most of the mod's problems would be solved if every comment were filtered through the HN guidelines before posting.

AI is a tool. You can use it constructively, like Grammarly, or spellcheck. You don't need to be afraid of it.

I like this guideline, at least in principle.

But I have some concerns about suppression of comments from non-native English writers. More selfishly, my personal writing style has significant overlap with so-called "tells" for AI generated prose: things like "it's not X, it's Y", use of em-dashes, a fairly deep vocabulary, and a tendency toward verbosity (which I'm striving to curb). It'd be ironic if I start getting flagged as a bot, given I don't even use a spell-checker. Time will tell.

Yes! This is really great feature, at the very least there being some proper Hackernews guidelines about it.

In my observation, recently there are quite many new AI generated comments in general. Like not even trying to hide with full em-dashes and everything.

I do feel like people are gonna get sneaky in future but there are going to be multiple discussions about that within this thread.

But I find it pretty cool that HN takes a stance about it. HN rules essentially saying Bots need not comment is pretty great imo.

It's a bit of a cat and a mouse problem but so is buying upvotes in places like reddit but HN with its track record of decades might have one or two suspicious or actions but long term, it feels robust. I hope the same robustness applies in this case well hopefully.

Wishing moderation luck that bad actors don't try to take it as a challenge and leave our human community to ourselves :]

Another point I'd like to say is that, if successful, then we can also stop saying, "did you write your comment by LLM" and the remarks as well which I also say time to time when I see someone clearly using AI but it seems that some false-positives happen as well (they have happened sometimes with me and see it happen with others as well) and they also de-rail the discussion. So HN being a place for humans, by humans can fix that issue too.

Knowing dang and tomhow, I feel somewhat optimistic!

How can HN be so pro-AI for the rest of the world, but anti-AI on HN?

Do we not think that other people want to see words, pictures, software, and videos created by humans too?

"If people cannot write well, they cannot think well, and if they cannot think well, others will do their thinking for them." - George Orwell

I don't think it is a moral failing to use AI to generate writing or to use it to brainstorm ideas and crystalize them, but c'mon isn't it weird to insist that you need them to write _comments_ on the internet? What happens when the AI decides you're wrongthinking?

Some basic things to do while thinking about longer-term bot detection:

1. Prevent any account from submitting an actual link until it reaches X months old and Y karma (not just one or the other.)

2. Don't auto-link any URLs from said accounts until both thresholds in #1 are met, so they can't post their sites as clickable links in comments to get around it. Make it un-clickable or even [link removed] but keep the rest of the comment.

3. If an account is aged over X months/years old with 0 activity and starts posting > 2 times in < 24 hrs, flag for manual review. Not saying they're bots, but an MO is to use old/inactive accounts and suddenly start posting from them. I've seen plenty here registered in 2019-2021 and just start posting. Don't ban them right away, but flag for review so they don't post 20 times and then someone finally figures it out and emails hn@.

4. When submitting a comment, check last comment timestamp and compare. Many bots make the mistake of commenting multiple detailed times within sixty seconds or less. If somebody is submitting a comment with 30 words and just submitted a comment 30 seconds ago in an entirely different thread with 300 words, they might be Superman. Obviously a bot.

5. Add a dedicated "[flag bot]" button to users that meet certain requirements so they don't need to email hn@ manually every time. Or enable it to people that have shown they can point out bots to you via email already. Emailing dozens of times a day is going to get very annoying for those that care about the website and want to make sure it doesn't get overrun by bots.

I've got some reflecting to do because the first thing I did after reading the headline, before even clicking to the actual post, is look for ai comments.

I miss pre 2010 internet. As soon as the advice animal memes started appearing on Facebook it was a quick decline.

What if English is my second language? Undoubtedly being well spoken is associated with higher class. Your arguments will come of as stronger to the reader.