Ask HN: Please restrict new accounts from posting

SG-

I'm honestly surprised HN isn't used to share more malware/githubs with new accounts too.

cebert

HN does a good job moderating and blocking spam from new accounts.

saltyoldman

Amen. I think the purpose of the bots is to create high-upvoted accounts for the purpose of later flag and downvoting things they've programmed the bots to suppress.

toomuchtodo

“shownew” : “no|yes” option would be nice.

BalinKing

I furthermore wish that "posting an LLM-generated comment (i.e. and passing it off as your own)" was worthy of an instant ban, because I see this sort of behavior from non-green accounts as well. EDIT: I meant (but totally forgot) to qualify that my "proposal" would only apply when the LLM-ness is self-obvious—idk, make up a "reasonable person" standard or something. Presumably, the moderators would err on the side of letting things slide. Even so, many comments I've seen are simply impossible for any reasonable person to claim as "human-written"—the default ChatGPT style is simply too distinct.

amadeuspagel

Let's turn HN into a place where we all grow old together until it slowly dies when we do.

verdverm

This is largely the same pattern that happened during the crypto hype cycle, spam posts and complaints. It will likely subside as reality sinks in. There are still quality submissions by new accounts and HN is good at pulling those needles from the haystack.

_alternator_

Devils advocate take: I think the quality of the ShowHN projects are in fact getting higher, at least the ones that land in the front page. The issue is that projects that used to take weeks, months, or even years of work now can be done in a weekend or so. It’s been democratizing, but it also means that when we look at these posts we (rightly) see that these new projects aren’t that much effort _with AI assistance_. So maybe we should just be honest about this: our standards have raised. We want to see Show HN posts that require effort and dedication, that require more than a few hours of prompt flogging.

neom

I almost emailed dang this morning to offer to help out tho I'm not particularly technical. Few solutions I thought of: 1 - honeypot, hide some links llms can follow if stuff gets posted in it, unlikely to be a human. 2 - Make an captcha that only llms can answer, I recently made 2 social networks, one that humans couldn't join by making the submission question too difficult to figure out quickly. 3- Use an LLM to detect LLMs, the other social network I did for fun (that a small number of people use), an llm that looks for moderation issues does a good job of flagging them. 4- Invites but vary the number you have to give out by account age + karma. The first 3 seem like they'd stop some % for some time, but eventually get old.

castral

I don't understand how this is supposed to solve anything, and I've seen it suggested as a solution multiple times. If you restrict comments to older accounts, all it's going to do is make the bot creators speculatively open and proactively age accounts for future use.

ThrowawayR2

I'd suggest instead a lower threshold for [dead]-ing posts and submissions by new accounts when flagged by HN users.

refulgentis

I'm honestly surprised how well it's going. From the perspective of usually just swinging into a post from the front page, when I do see green, it's usually overtly political trolling, and dead from the start. So I had assumed new account = everyone sees your post in gray, at least for a week or two. I don't envy the "Show HN:" case. It can be intractable, story time: Last week, there was a "Show HN:" post for a GitHub link, made it all the way to #2. It was a Flutter app, written up as if it did all the stuff you'd want from an open source LLM client. I said to myself "geez, I knew I took too long to deliver the thing I've been working on for 2 years. the MVP version is insanely popular." -- only after digging into the repo for 10 minutes, with domain expertise, did I realize it was a complete Potemkin village, built by Claude. And even then, I was afraid to post something pointing this out because it required domain expertise, and it could have read as negative rather than principled. All that to say, some subsets of The AI Poster Problem now require having intimate domain expertise and 10 minutes to evaluate it. :/ Additionally, the Claude 4.6s and GPT-5.4s are better than me at posting on HN now. :/ And I've been here 16 years. The past couple days, any comment I write involving some sort of judgement or argument is by Opus 4.6 or GPT-5.4, via: 1) dump HN post into prompt 2) say "I feel $X about this, write me an HN post that communicates this but not negatively". I'm a little ashamed to admit if you look through my post history, you'll definitely see a repeated pattern over 16 years of someone who is very negative and has a hard time communicating it constructively. They're smart enough now to extrapolate observations in the way I want to, while avoiding my own tarpits.

abtinf

I really wish there was a setting whereby I could simply hide all comments from accounts less than a year old. The correlation with LLM slop is simply off the charts. It almost feels like new accounts should be treated like new posts -- it is sort of a service that a select few are willing to undertake to upvote interesting stories early on. I wish even more I could block specific users (there are some highly prolific, high karma users here who are extremely irritating), but that's harder and is probably best handled client side.

delichon

There is an epistemic silver lining. This is in fact a Red Queen's race that cannot be won. So in the end the only solution is to evaluate the text on its own merits without reference to the writer's status, because that status can no longer be reliably detected. For a public feed like this one, the only alternative is to ignore it. The fire hose of data will inevitably become ever more fecal. We can only walk away from it or be more careful about the pearls we pluck out. It ends well only if we get better at pearl detection.

bmacho

Accounts have to start posting somewhen. Moderators don't have the capacity (and fairly, it is impossible) to check if they are bots or humans. There are no good solutions, there are hundreds of thousands of intelligences out there, trained millions of hours on how to scam humans, capable of spitting out text tirelessly and shamelessly, and there will be only more of them, tens, hundreds, thousands times more.

dang

We're going to at least restrict Show HNs for a while. I do think this is relevant though: "HN can't be immune from macro trends" - https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

alabhyajindal

100%. Not sure what the solution is but I have lost interest in Show HNs these days. Part of it is because when someone posted before, it usually meant they spent a fair amount of time thinking, and found it worthwhile to spend energy on the project. This was a nice first filter for bad ideas and now no longer exists. Even for posts that are interesting to me, I get the feeling that it's not worth looking at because it was probably made using LLMs. Nothing against them, but I personally thought of Show HNs as doing something for the love of it, the end result being a bonus.

neural_thing

Eventually HN is going to need to charge people $1 to post, just for spam filtering. Maybe donate the money to open source or something.

Lerc

There have been numerous stories on HN where someone directly involved with the story has created an account specifically to engage in discussion about whatever the story was about. Losing that seems too high of a price to pay. Yes there are AI generated comments, in the past there has been script generated comments. You can report, downvote, or just ignore and move on. I am aware of posts like this existing, but I feel they are being effectively managed. Try not to be too offended about the notion of these posts existing. Many of them are not malicious, they just caused by users stepping outside what is considered appropriate, but in a landscape where the footing is quite dynamic, everyone is making their own judgement calls in a field where the consensus is not clear, guidance seems more appropriate than punishment here.

DetroitThrow

It used to be so pleasant to read Show HN and find such interesting projects, but nowadays it's rare that any project posting their GitHub has ever read their source code or even comes close to functioning in the way the OP claims. Such a sad development.