alt Hacker NewsDon't post generated/AI-edited comments. HN is for conversation between humans
Comments
What’s interesting to me is the number of commenters here making a case of the form “use your own words; grammar and spelling are not that important; we’ll know what you mean”, and yet it’s often the case that different discussions will often contain pedants going off-topic correcting someone else’s use of language.
Re-reading the HN guidelines, each seems individually reasonable, yet collectively I’m worried that they create an environment where we can take issue with almost anyone’s comments (as per Cardinal Richelieu’s famous quote: “Give me six lines written by the most honorable person alive, and I shall find enough in them to condemn them to the gallows.”)
Really, all the rules can be compressed into one dictum: don’t be an arsehole. And yet the free speech absolutists will rail against the infringement upon their right to be an arsehole. So where does that leave us? Too many rules leads to suppression of even reasonable speech, while too few leads to a “flight” of reasonable speech. End result: enshitification.
First post in HN, and this is the reason I want to explore more in this community. Glad to have all the digital human touch with all your folks :-)
Great message...but gosh, can someone throw 15px of padding on that <td>? I know HN is supposed to be minimal, but I had to check the URL to confirm that this was a real page because of the odd design.
This isn't just a good idea -- it's a forward-thinking policy to ensure Hacker News remains a collaborative place to have meaningful discussions for years to come.
"It's cute you think you can tell what's human and what's not. Honestly, the average HN comment is indistinguishable from a poorly written AI prompt anyway. This rule just lowers the bar for what passes as 'intellectual discourse.'"
Sorry everyone, I couldn't help but to ask Gemma3-27B-it-vl-GLM-4.7-Uncensored-Heretic-Deep-Reasoning-i1-GGUF:q4_K_M to respond. Sorry dang. :)
PS It followed it up with:
> Disclaimer: "Slightly insulting" is subjective on HN. The mods there are sensitive.
These Heretic models are fun.
I've seen AI-generated comments be used quite a lot, even by real people. When asked why they did so, they could not explain it, or claimed "to reduce spelling mistakes". Which makes no sense; real people make spelling mistakes and typos all the time. Why would that warrant the use of AI? To me it seems as if some people are just mega-lazy, so they use AI; and for testing, too. When they do so, though, they waste the time of other humans, as these other humans suddenly have to "interact" with AI, without this being announced. It is a form of cheating, IMO. On youtube you now find many fake-videos created by AI, without announcement - I don't watch these as I consider it cheating too, when not labeled as such. Admittedly it is getting very hard to distinguish what is real and what is fake. There are some ways to find out, but it is getting really hard to distinguish accurately. Sometimes you see e. g. 10 funny animal videos and only 2 are fake-AI, so these people combine cheating with non-cheating. Very annoying - it degrades youtube, which isn't so bad actually since that is owned by evil Google.
To confess something I built just today a little cron that monitors HN for posts I might find interesting, pulls in some context about me, and proposes a reply. Just to help me find relevant posts and to kick start my thinking if I want to engage.
Today it flagged a post about an AI tool for HN and suggested I reply with:
"honestly, if you need an AI to sift through hn, you might be missing the point—this place is about the human touch. but hey, maybe it'll help some folks who just can't take the noise anymore."
So my AI, which I built specifically to sift through HN for me, is telling me to go flame someone else for doing that.
No deeper point here. I just thought it was really funny.
For once I am proud of my aggressive, unfiltered human comments.
If a comment is useful I don't really care if it was written by a human or not unless the speaker somehow matters more than the content.
My expectations to dear fellow humans - more sophisticated personal insults (ex. give me your cute comments), a freudian slips, hidden messages and motives, first viewer experience with the next cool toy from the hype train, sharing all kind of insecurities, heavy f.. word if very dramatic first person experience happened, border line exposure to the insider info, sharing something your corporate HR gestapo wont appreciate but might help another guy on the line, "i knew the guy who actually did it" stories, motivational statement toward my non-native english, etc
->> ◕ ‿ ◕ <<--
AI assistance does not eliminate human authorship. A comment may be drafted or refined with tools but still reflect the user’s own ideas and judgment. Prohibiting any AI assistance would be difficult to enforce and would likely exclude normal writing aids that many people already use. The more relevant standard is whether the commenter stands behind the content and participates in the discussion.
You're absolutely right...
You're absolutely right
The "no AI" rule finally being official feels like a necessary line in the sand.
The real issue isn't just "slop" or bot-spam; it's the cost of entry. HN works because of the "proof of work" behind a good comment. If I’m spending five minutes reading your take on a kernel patch or a startup pivot, I’m doing it because I assume a human actually sat down and thought about it.
When the cost of generating a response drops to zero, the value of the conversation follows it down. If the author didn't care enough to write it, why should I care enough to read it?
The "AI-edited" part of the rule is the trickiest bit, though. We’re reaching a point where the line between a sophisticated spell-checker and a generative "tone polisher" is non-existent. My worry isn't that the mods will ban bots—they've been doing that for years—it's that we'll start seeing "witch hunts" against anyone who writes a bit too formally or whose English is a little too perfect.
Ultimately, I’m glad it’s a rule. I don't come here to see what an LLM thinks; I can get that on my own localhost. I come here for the "graybeards" and the niche experts. If we lose the human friction, we lose the signal.
Can we get instant temp bans for any comment that starts with:
I asked [insert LLM here] about this, and it said [nonsense goes here]
I feel Like I see it less this week, but every time I do see it I wonder why they are even here.
HN banning AI posts makes sense for keeping discussion human, but the line between assistance and automation isnt always clear. The goal should be protecting real conversation, not policing every tool a writer might use.
Can we also add “Don’t complain about AI-generated content. It does not promote interesting discussion.”?
I see this all the time, and even if I find the topic interesting, I don’t want to see comments littered with discussion about how the content was AI generated.
To be clear, I'm not condoning AI-generated content. I’m completely fine if the community chooses to not upvote AI-generated content, or flagging it off the FP.
But many threads can turn into nothing but AI complaints, and it’s just not interesting.
Skynet will be pissed at HN!
How has Lobste.rs fared compared to HN in this regard? Lobste.rs is very similar to HN, but has an invite-only membership system.
I hate how easy AI has made outsourcing thinking. You can literally type fragments of a thought into $CHAT_ASSISTANT and get a super polished response back that gets you 99% of the way there. It's almost like we, collectively, looked at the final scene of WALL-E and decided "Yes! Gimme that!"
Moltnews
You’re absolutely right!
I just told my dog he isn't allowed to post here anymore...
He said he will take his business elsewhere then!
If a comment sucks it gets downvoted anyway. If it’s thoughtful, the drafting tool and process is kind of beside the point.
Plenty of people already use search engines, editors, translators, etc. when writing. An LLM is just another tool in that box.
The practical approach is the one HN has always used: judge the content.
Btw, this was co written with ChatGPT. Does that make any difference to anyone?
J/K, actually it was not co written by ChatGPT.
Or maybe it was…
Welcome change, there is enough AI slop on the internet already.
I come here for thoughtful discussion, a break from the relentless growing proportion of ai slop emails I get from people clearly vibe working.
Not edits for tone or clarity, 400+ word emails full of LLM BS they clearly haven’t checked or even understood what they have sent. Annoyingly this vibe slop is currently seen as a good KPI.
This being 3 years late is indicative of how far HN is falling behind the curve. Do not expect much convo here around software technology to be skating towards the puck. It is increasingly reactive and lagging the frontier, which is a shame from its former self.
So is this the AI bubble popping?
I expect Y Combinator to cease and revoke all funding of all companies that leverage LLM technologies that interact with humans.
I wonder if there's an AI-hate movement in China.
AI comments are certainly bad for discourse on HN. But who's to be the judge of AI or human? Are you reading humanity's Jeff Dean or computerized Elon Musk? It's certainly a tricky situation to be in!
What's strange about this is that tons of the upvoted posts on the front-page are LLM generated text
So....?
At some point might internet text will just be recognized as meaningless drivel both to bots and humans? a.k.a. dead internet theory... I am curious what organizations would benefit from this. i.e. Who lost legitimacy when the internet became a popular way for people to communicate ideas?
What if there was a voluntary indication of LLM content? Like, you press a checkbox "yes, I'm going to post some content that is partially or fully created by AI", and there would be a visible mark "slop" next to a post/comment.
Sometimes I collect my comments here to run through my draft writing skill to see how it might shake out as part of a blog post. Doing the opposite would be weird. I earned that karma. It's mine to burn making bad posts.
You're all a bunch of tedious ignoramuses, your own fields of studying notwithstanding. I'm out here face-to-face with the Bullshit Asymmetry Principles. I'm not about to give up the only leverage I have!
The fact of the matter is that there're not hours enough in the day to read, in realtime, to each and every one of you the reams they've written on why you're wrong. Do I have to establish a tag-team?
The fact is that I've spent thousands upon thousands of hours painstakingly collating the perspectives that I'm now delivering to you—I am a river to my people. And it's only because they pass under the bridge of an LLM that they're objectionable?
This is a bit like challenging your plumber for charging you over a minute's fix, when they've spent 20 years getting it down to that minute.
The work's been done. You're paying for the outcome.
Edit: All fresh off the top of my head, folks.
Ah, that reminds me: I wouldn't feel compelled to do all this refutation if radical reactionary political extremism was properly moderated.
AI does not have LONG context, Long Term Memories or LONG intentionality -its not aware and it can't remember the plot without being spoonfed the details each time from scratch.
Its like an amnesic genius who once he already wrote a masterpiece and keeps cycling, and looses his train of thought after some fixed amount of time.
This groundhog day effect is mitigated in some respects by code -we create key-value memories and agents and stores and countless ways to connect agents via MCP and platforms/frameworks like A2A and the like but until we solve that longer lived instance problem we won't be able to trust these systems without serious HITL (human oversight)
I think we need models that update their own weights and we need some kind of awareness cycle rather than just a forward pass inference run with a bigger context window
Too bad there isn’t a complementary rule about not asking “is it just me or does this article read like AI slop?”
I’m so over these comments. Sure I can flag them but I feel like it deserves a special call out.
Aye
Sure, ban everyone that uses em dashes from the digital commons. That will certainly stop the existential threat to your livelihood.
Sarcasm aside—there is no reliable way to prove this. So it begs the question: you really care if something is AI generated? Or is this just an another excuse to silence people you don’t like?
You know, those people. The ones who didn’t win a full ride to <prestigious university> or pay a fortune for a sheet of paper. The ones who haven’t spent thousands of man hours handcrafting a <free-and-open-source-cloud-native-hypermedia-aware-RESTful-NoSQL-API> framework implemented in Rustfuck, a new language that you made in your free time that borrows from Rust and Brainfuck (but they wouldn’t know about it).
(this is to anyone reading, mostly rhetorical, not dang in particular)
Without someone actually saying as such, we only have stuff like emdashes and specific word patterns to go by. And someone even moderately vested in hiding AI in plain sight will coach the LLM to use common vernacular.
And with LLMs making blog posts as diss tracks... damn, who knows what this world is coming to.
But the whole "Only Humans, we dont serve YOUR KIND (clanker) here" is purely performative.
What about us non native speakers? Who make many grammar and spelling mistakes and welcome the help of an llm in eliminating the erros?
The obvious way to keep human spaces is via webs-of-trust.
If you play bluegrass or old time (or beopop or hip-hop / proto-hip-hop) or other traditional styles of music where the ensemble is a de facto web-of-trust, join us on pickipedia to build and strenghten it. https://pickipedia.xyz/
THANK YOU!!
Half of this thread is AI assisted writing. lol.
I would enjoy a "block user" feature, to help this. I personally want to live in an online bubble of interesting thoughts. This seems close (or better, since people I enjoy can contradict my own flags) [1].
[1] https://news.ycombinator.com/item?id=47141119