Comment by dang
8 days ago
Yes, we need to do something about this and tomhow and I are talking about it - it's not clear yet what.
Raising the quality bar would likely cut down on quantity as a side effect, and that would be a nice solution. One idea that a user proposed is a review queue where experienced HN users would help new Show HN submitters craft their posts to be more interesting and fit HN's conventions more.
I recommend making "What are you working on [this weekend/weekly]" official like whoishiring and encouraging pre-Show HN comments there. (https://news.ycombinator.com/item?id=47041973#47043174 "what's the right venue for sharing [LLM-built side projects]")
Also requiring disclosure of the use of AI in repos and especially (or perhaps specifically discouraging its use) when responding with comments to HN feedback.
I'll take this opportunity to strongly encourage sharing prompts (the newest tier of software source code) as the logical progression of OSS adding additional value to Show HN.
Combining the two approaches might work. A "pre-moderation queue" for submissions that are are solid enough to pass the "Show" bar, and then the monthly "what are you working on" threads as a more free-form creative outlet.
And yes, disclosing the use of AI should be par for the course.
My issues with the super posts is its really hard to grep for relevant information.
I've tried asking ChatGPT to recommend projects based on my interests, but ChatGPT heavily hallucinated or projected my interests onto irrelevant projects (for example, a project might be about developing a new programming language and chatgpt was like, "you could use this in your soap making hobby!").
The Who's Hiring posts have community sponsored indexers and its easy for me to query job titles relevant to me, but keyword search is not as useful here.
Bruce Schneier had the Friday Squid Blogging posts that he allowed random stuff on. Most of the noise was contained in the Friday threads.
Lobsters had a weekly thread by caius called, "What are you doing this week?" People would post personal projects or experiences in it. On top of interesting tech, that let us pray more specifically for and encourage some in need.
The weekly threads might work here.
My new favorite thing to do is grab the 'What are you working on threads' and have a LLM group them categorically with one line descriptions of each app.
Interesting take on the sharing of prompts. I don’t think this is a bad idea. How would this work though given different prompts occur in different context windows?
Internally, we have a standard that any AI written code simply includes a cut-and-paste of the chat prompt (if that were used), and/or the .md files (if those were used).
I would like to see it as extended comments in each git commit. There have been a few examples of some doing so manually but it needs to be supported by the tooling... with all the half-arsed "standards" like MCP etc. I'm surprised there isn't something already.
[dead]
Sharing prompts, not sure it works if your project required hundreds of prompts? It’s all in history though (.jsonl) so I’m sure the AI can condense it somehow.
What I see is new users who are trying to share something without having yet understood HN. I get the impression that they think of "Show HN" as no different than "Show and Tell", and that putting the label on their post is communicating the message of "Here is something I want people to see", instead of "Here is something you can try out".
So while I understand that new features on HN are few and far between, a quick validation of "Show HN" posts that says, "I see you are trying to post a Show HN..." with some concise explanation of the guidelines might help. I want to believe that most new users mean well, they just need better explanations.
> What I see is new users who are trying to share something without having yet understood HN.
From their perspective, HN is another place to post and get views on their project, part of a check list for their "launch" or whatever, not everything comes from within the ecosystem.
Some posts their projects then never reply to any of the comments, while for me (and many others I bet) half the reason of posting a Show HN is because I'm looking for participating in discussions about my thing and understanding different perspectives thinking about it too.
> I want to believe that most new users mean well, they just need better explanations.
Yeah, so far the only thing I know of is the "Please read the Show HN rules and tips before posting" blurb on the /show list, and the separate pages. Maybe some interstitial or similar if the title prefix-matches with "Show HN" could display the rules, guidelines and "netiquette" more prominently and get more people to be aware of it.
The explanation should also prominently include a "no advertising or marketing" admonition since the submitters that don't mean well often seem to think putting Show HN will let them get away with it.
Not sure if it would work for HN / how it could be adapted to HN, but something I noticed on opensource projects, is that once they hit a hurdle, submitters of low quality AI-written PRs don't try to solve it and go elsewhere.
For example, in one project, PRs have to be submitted to the "next" branch and not the default branch. This is written in the CONTRIBUTING.md file, which is linked in the PR template, with the mention that PRs that don't respect that will be close. Most if not all submitters of low-quality PRs don't do anything once their initial PR is closed.
Pretty bummed about that as I just submitted a show HN I'm pretty happy about (it solves an annoying problem I had for years, which I know many people have) and I was looking forward to talk about it (https://news.ycombinator.com/item?id=47050872)
Back when I ran a WoW guild, the first sentence in our recruitment post emphasized the importance of reading the whole post (because the way to access the application form was to click the only smiley in the post, and this detail was mentioned in the last paragraph).
Most people did not read the post, which was immediately evident from how they posted their application by copy-pasting and editing an application posted by someone else before them.
> Back when I ran a WoW guild, the first sentence in our recruitment post emphasized the importance of reading the whole post (because the way to access the application form was to click the only smiley in the post, and this detail was mentioned in the last paragraph).
I mentioned the "no brown M&Ms rule" in a recent comment, and someone pointed out to me that this is more likely to by adhered to by an LLM - humans might miss a single line in a three pages of text, but the LLM won't.
I am starting to think that a better approach might be to move to mailing lists; this means that valuable drive-by PRs (like I did in the past) are going to be unintended victims by preventing all drive-by PRs, because the friction is too high. Submitted needs to have an email address, make sure it isn't marked as spam, sign up to the mailing list, email their PR, wait for a response, etc.
The upside is that projects can start marking that email address as spam if it submits AI slop. The downside is that actually valuable drive-by PRs will be a thing of the past.
1 reply →
> For example, in one project, PRs have to be submitted to the "next" branch and not the default branch. This is written in the CONTRIBUTING.md file, which is linked in the PR template, with the mention that PRs that don't respect that will be close. Most if not all submitters of low-quality PRs don't do anything once their initial PR is closed.
Few things in life are as reliable and trustworthy as the laziness of others.
In my opinion, for open-source projects, scoring the project's AI sloppiness based on the timeline of commits would be a good indicator. If it's completed within a few days, it should require more thorough human review. On the other hand, if the project has been active for a while and received contributions spread throughout that timeline, I think that would indicate accumulated effort (human and/or AI) and higher quality.
> In my opinion, for open-source projects, scoring the project's AI sloppiness based on the timeline of commits would be a good indicator.
You can’t necessarily judge by timeline. I’ve always developed my projects privately and then squashed to one initial public commit. I’ve got a private repo now with thousands of commits developed over years and I still intend to squash.
Show HN has never been restricted to open source projects and it would be weird to make the criteria more restrictive for open source than closed source work.
I thought so too, untill I looked for 1-point Show HN posts with a repo with a long commit history. Some of these are really cool (see my article), but others were not compelling at all, at least to me.
Not a good bar - I develop projects privately, and make public only when I am good and ready.
Eh IMO any metric like this can be gamed. My project that reached hn front page was coded in a short time (and yes some ai was used), but otoh I think it was something that showed hey you can do this really interesting thing (in my case vlm based indoor location).
Also its not uncommon for weekend projects to be done in a shprt span with just a "first commit" message dump even pre-AI.
Yes, any metric can be gamed. But I believe measuring the entropy of a repository, comparing state of the code-base over time can be done deterministically, which would make it harder to game it.
So either we are going to completely avoid automation and create a community council to decide what deserves to be shown to rest of the community or just let best AI models to decide if a project is worth show up on front page?
Or we can do all of the above :)
2 replies →
Unfortunately there's now a whole cotton-industry of "vibe-coding classes and marketing" (similar to "life-coaching" on socials) that probably target HN as well. I think HN needs to think a layer of abstraction "higher" and model around some collection of semantics/metrics that allow to filter out "gloss without quality" vaporware or voting ring tactics.
How about inverting the issue, highlight posts with an opt in label. e.g
Since it's too controversial to ban LLM posts, and would be too easy for submitters to omit an [LLM] label... Having an opt in [NOAI] label allows people to highlight their posts, and LLM posts would be easy to flag to disincentivise polluting the label.
This wouldn't necessarily need to be a technical change, just an intuitive agreement that posts containing LLM or vibe coded content are not allowed to lie by using the tag, or will be flagged... Then again it could also be used to elevate their rank above other show HN content to give us humanoids some edge if deemed necessary, or a segregated [NOAI] page.
[edit]
The label might need more thought, although "NOAI" is short and intelligible, it might be seen as a bit ironic to have to add a tag containing "AI" into your title. [HUMAN]?
I'm 90% sure this will end with endless squabbles who's right that the label is correct/incorrect, rather than actual conversations about what the project that the person is showing. It already happens without the labels, feels like it'd increase the frequency of that even more if this label gets enforced.
Is the problem that the app was written with AI assistance or that it's low-effort/bad? I don't care if you used Claude to fix a bug or something if you have a cool app, but i do care if you vibe coded something I could've vibe coded in an hour. That's boring.
Feels like effort needs to be the barrier (which unfortunately needs human review), not "AI or not". In lieu of that, 100 karma or account minimum age to post something as Show HN might be a dumb way to do it (to give you enough time to have read other people's so you understand the vibe).
A core part of the HN ethos is avoiding siloing dynamics, which is exactly what [NOAI] would be.
It's a tough problem.
Once some users have extra power to push content to the front-page, it will be abused. There will be attempts to gain that privilege in order to monetize, profit from or abuse it in some other way.
The only option along this path would probably be to keep the list of such users very tightly controlled and each vouched for individually.
Another approach might be to ask random users (above certain karma threshold) rank new submissions. Once in a while stick a showhn post into their front page with up and down arrows, and mark it as a community service. Given HN volume it should be easy to get an average opinion in a matter of minutes.
I feel like these review processes could be public threads like any other (just not linked directly from /show).
I was thinking we could call them "Show Show HN" but I suppose that joke would get old
How much would it help cut down if Show HN was prohibited for accounts that were green and/or only had 1 karma?
Meaning you would have to demonstrate that you had or were willing to contribute to the HN community before just promoting your own stuff.
As it happens, a great counterexample has been on the frontpage for hours:
Show HN: AsteroidOS 2.0 – Nobody asked, we shipped anyway - https://news.ycombinator.com/item?id=47051852 - Feb 2026 (34 comments)
What a bummer it would be to lose excellent posts by new users. They do happen!
My favorite example for a new user is this: https://news.ycombinator.com/item?id=33566924
This was a super encouraging and uplifting experience for everyone involved in the project. Thanks so much for providing this excellent platform.
on problem here is that I don't want to share my ShowHN projects with my HN account because it'll connect my real identity to my pseudonymous HN identity.
so, in the past, i've created throwaway HN accounts for sharing things that connect to my real ID.
The solution to this is to allow aliases ... or account groups if you will and then allow/disallow things based on combined karma.
I'm not entirely sure that's a bad thing. If you aren't proud enough of it to attach your real name, or your pseudonymous account, maybe it shouldn't be posted.
1 reply →
Honestly I would support only allowing Show HN for accounts that can downvote. It's really not such a high threshold.
It would filter out some people who lurk but still have interesting things to contribute, but despite the drawbacks that threshold is probably the most immediately impactful solution. Of course, it strongly incentivizes the purchase/selling of accounts, and karma farming, but that problem is perhaps less of a problem than all high-effort human content getting completely drowned out. There are already a plethora of spambots making comments getting upvoted, so it's not like that problem doesn't already exist either.
1 reply →
We would have to prepare for a deluge of accounts posting\commenting just for the sake of accumulating karma to be able to downvote.
1 reply →
(Now that the front page attention has passed...)
It should be possible to find the users with the most early upvotes on Show HN posts that wind up "succeeding" as an excellent indicator they can serve to filter them going forward. However to me it's a bit awkward for Y Combinator to plan on relying on volunteers... once these users are found one of them should be hired to do this full time.
I wonder if some kind of voluntary tagging system could help?
e.g. [20h/2d/$10] could indicate "I spent 20 human-hours over 2 days and burned $10 worth of tokens" (it's hard to put a single-dimensional number on LLM usage and not everyone keeps track, but dollars seem like a reasonable approximation)
I just have no interest in seeing code that hasn't even been read by the author. At that point, just show us what it does, don't show us the GitHub link.
On the front page, someone made a cool isometric NYC map via vibe coding - another front pager was someone who also claimed to make an ultra fast PDF parser that failed on very common PDFs and gamed the speed metric by (useless) out of order parsing.
Guess which one I installed and spent more time using? These vibe coded projects aren't interesting for their code and almost always not intended to be used by anyone if they're libraries/frameworks, but the applications made with vibe coding are often very cool.
An easy win is turning off the firehouse of vibe coded GitHub portfolio projects and just ask for a link to a hosted application. Easy.
> a user proposed is a review queue where experienced HN users would help new Show HN submitters craft their posts
A separate queue is the solution. Then the moderation system (dang, tomhow, seasoned users) can apply an adjustable threshold to promote Show HN posts to the main queue.
To my dismay, the trajectory of Show HN posts looks eerily familiar. ProductHunt followed a similar course (albeit with much more acceleration) and now is just a feed of slop. The signal to noise ratio became so meaningless that I lost all interest. I fear this happening to HN. Any attempt to slow this down is welcome.
I wonder how will this review system work. Perhaps, a Show HN is hidden by default and visible to only experienced HN users who provide enough positive reviews for it to become visible to everyone else. Although, this does sound like gatekeeping to me and may starve many deserving Show HN before they get enough attention.
Funny a year ago I used to hear from so many people who thought a Product Hunt launch was the same as a marketing plan but it's been a while since I've heard about Product Hunt...
I'm haunted by the criticism Dropbox received from HN users when they posted their project here. While I respect the views many of us have, I think this has the potential to have the StackOverflow effect where the community makes the whole process miserable and worse.
Note well that the most famous example of this is a misreading that has snowballed into a kind of cultural legend. The thread in question was about Dropbox's application to YC, not the value of Dropbox itself, and the feedback was constructive and well-intentioned.
https://news.ycombinator.com/item?id=42392302
Could also consider requiring higher account age or points
It's tempting, of course, but I wouldn't want to exclude submissions like https://news.ycombinator.com/item?id=47056384)
> One idea that a user proposed is a review queue where experienced HN users would help new Show HN submitters craft their posts to be more interesting and fit HN's conventions more.
HN has a vouch system. Make a Show HN pool, allow accounts over some karma/age level to vouch them out to the main site. I recently had a naive colleague submit a Show HN a week or so ago that Tom killed... for good reason. I told the guy to ask me for advice before submitting a FOSS project he released and instead he shit out a long LLM comment nobody wants to read.
The HN guidelines IMO need a (long overdue) update to describe where a Show HN submission needs to go and address LLM comments/submissions. I get that YC probably wants to let some of it be a playground since money is sloshing around for it, but enough is enough.
Maybe restrict Show HN posts to 70 or 100 characters - readers can then scan many, and quickly find stuff of interest.
The clarity and focus this discipline would enforce could have a pleasant side effect of enabling a kind of natural evolution of categorizations, and alternative discovery UIs.
Proposal:
- Min. 90 days account existence in order to submit
- Cap on plain/Show/Ask HN posts per week
Most of the spam I see in /new or /ask is from fresh accounts. This approach is simple and awards long-term engagement/users while discouraging fly-by-night spammers.
Yep, hide new accounts except to higher karma users is an easy win.
But very gatekeepy. New accounts are already highlighted and get mercilessly reported and downvoted if they post spam.
3 replies →
Something based on the principles of 'New'? (not clear on the details of how Show HN works, does it automatically appear?). Just shove entries under 'New' and let the group decide what is "Show HN"-worthy.
Every system can be gamed, but if it were me and I were looking for a simple filtering solution, I would do something like this ..
Set a policy of X comments required per submission in the last 30 days (not counting last 24 hours) for all submissions, not just "Show HN:" posts.
Meaning, users would need to post X comments before they could post a submission and by not counting the last 24 hours, someone couldn't join, post X comments and immediately post a submission.
It would limit new submission posts to people who are active in the community so they would be more familiar with the policies and etiquette of HN along with gaining an idea of what interests its members.
One thing I noticed recently while going through several of the Show HN submissions was that a lot of the accounts had been created the same day the submission was made.
My guess is HN has become featured on a large number of "Where do I promote/submit my _____?" lists in blogs, social media, etc. to the point that HN is treated like a public bulletin board more than a place to share things with each other in the community.
I love the Show HN section because so many interesting things get posted there but even I have cut back on checking it lately because there are simply too many things posted to check out.
I hope they do something to improve it.
> Set a policy of X comments required per submission in the last 30 days (not counting last 24 hours) for all submissions, not just "Show HN:" posts.
This will definitely improve the SNR of all submissions. I second this.
May be dont show them under "Show HN" unless the post has accumulated a certain number of points/upvotes ? Just like a regular new post comes on front page.
I think this already happens but the threshold might be pretty low - there's a separate shownew page and if it gets sufficient upvotes it appears in the regular "Show HN".
I’m in some discord servers that have both #art and #ai-art channels. This seems to work well. It’s not perfect but it’s cheap and might be good as a start.
> HN users would help new Show HN submitters craft their posts to be more interesting and fit HN's conventions more.
hah that sounds like a Show HN incubator.
You can cut 90-95% of the slop posts by checking the project's GitHub repository. If it has .cursor / .claude folders it should be automatically pushed into a slow queue with the queue shifting every X hours. For example, this really popular post from 27 days ago was just pure AI slop. It got on to front page of HN and stayed there for days. Should have been nuked the very first hour:
https://news.ycombinator.com/item?id=46710710
Maybe you should train an LLM to judge the content /s
More seriously though, I think some sort of curation is unavoidable with such topics. If you get inspired by stack overflow where you have some similar mechanics at work, then I'd say that is not too bad. But of course you risk some people being angry about why their amazing vibe coded app is not being shown. Although the more I think of it, this might be a good thing.
Edit: One more thought just came to my mind. A slight modification to the curation rule, you let everything through, just like now. However, the posts are reviewed and those with enough postive review votes get marked in some shape or form, which allows them to be filtered and/or promoted on the show page.
Please make the home page show 60 rather than 30 stories by default.