Ask Dang: What tools do you use for moderating HN?
Hi Dang,
HN is undoubtedly one of the best networks that I am part of. The site is very well moderated. Wondering what tools / help do you get in moderating HN?
HN is undoubtedly one of the best networks that I am part of. The site is very well moderated. Wondering what tools / help do you get in moderating HN?
> The site’s real-life moderators are Daniel Gackle and Scott Bell, two wildly polite old friends. […] Gackle and Bell, by contrast, practice a personal, focussed, and slow approach to moderation, which they see as a conversational act. They treat their community like an encounter group or Esalen workshop; often, they correspond with individual Hacker News readers over e-mail, coaching and encouraging them in long, heartfelt exchanges.
My impression is that it's just an old-fashioned manual process.
https://jargon-file.org/archive/jargon-1.5.0.dos.txt
Login names are often used as nicknames, pronounce[d] if possible and spelled if necessary. My wife and I met at MIT, and she still calls me "Gliss", because my login name was GLS. "Guy" sounds very weird to her.
* as a publisher or a subscriber, and to some extent even as a reader
I've been part of a brief but very positive exchange like this with dang, and it's not an exaggeration. This is part of why I still visit the site; there's a real personal touch that makes this forum what it is.
I felt stupid about a comment I'd left and thought "ugh, I should ask someone to delete this". dang responded saying it was a great comment and that I should consider keeping it as it was the kind of content he felt made HN a constructive and interesting place. I was totally caught off-guard.
Influence campaigns are an older problem, so we have many years' worth of defenses built up, although it's an arms race where the influencers also evolve.
I'm aware that I didn't answer the "how" part of your question—the trouble is that there's no secure way to do it, so everything relies on a stack of ad hoc techniques that would stop working if we published them. It's basically the old "I'd tell you but then I'd have to kill you" joke. But the answer to the "if" part is yes.
But on the whole it does seem like the old-fashioned method of clear consistent communication.
I think this is outdated, or flat out not true.
I have had particularly sour engagements with dang specifically. I asked him to reach out to me via email and he just kept pushing the matter in public.
Pretty poor handling all around, and completely lacking inclusivity or self-awareness for a "seasoned professional." I'm glad it's daisies and roses for most of you, but some of us have not seen such kindness. Ultimately you have to ask yourself if moderation has fostered an environment of growth and evolution, or an echo chamber. In my experience I have seen more of the latter on HN, and it's concerning.
It is entirely true, as of a couple of weeks ago.
I hear you that you didn't get a satisfying email response, and I'm sorry about that. It's possible that I screwed up. Actually, it's likely that I screwed up, because you're posting this some time later and it's my job not to leave people feeling this way.
Without details, though, it's impossible to say what happened. I probably wouldn't email a user who just said "email me" in a public comment. But if you wrote in to hn@ycombinator.com and I didn't read or hear you or reply in a way that met your concern, then I'm sorry.
You'd think this small thread you replied to would be enough detail if you were willing to reply in the first place.
Look at my original post. Simply sharing my experience (which is on-topic). It currently sits at -3, while other comments praising moderation here are mostly positive. This is a pattern I am noticing across more than one thread, in more than one day, with more than one user. It's not an anomaly.
Replies to me are from folks lacking self-awareness, who think their experience somehow trumps or invalidates mine. Nobody is directly addressing my concerns. Just brushing them off and piling on.
You should recognize there is a growing majority at HN who perpetuate this "SV" flavor of culture, and you should recognize that your inability to scale efficiently is detrimental to the long term growth of HN. Ironically, against the users most willing to help you do better, you level threats publicly and cite HN rules as if they have any relevance.
Frankly, it seems you are blissfully unaware of how much of an echo chamber exists here, or that you are directly perpetuating it. It's concerning. Flagging comments and downvotes are actively abused for censorship, and you clearly have no idea how to fix it. You think the system is working, what if it has always been horribly broken...?
It's even more concerning that seemingly every other member at HN gets white glove treatment, yet I am still wondering when I will be treated like a human by the staff.
I’m not really active online for the majority of my usage of sites like this.
I’m just here for the content.
Rarely I will submit something interesting I come across but over the years I’ve noticed the cycles of the same sites and the same content.
The “Dead Internet Theory” really does resonate with me. I’ve been browsing HN since ‘06 when Tesla was just the weird roadster maker.
Just look at how popular my original comment is. There's clearly a culture lean here and I'm not going to feed into it.
I'm going to call it out and laugh when I'm proven right.
hn@ycombinator.com
They're kind enough to respond when time allows if it ain't urgent
only good words from my side for Dan G.
Specific mention of a custom browser extension: https://news.ycombinator.com/item?id=30374040#30378776
You don't often see de-escalation on the internet. Especially between moderator and user when they could just as easily disappear you from the site. It works well.
I think the more upvotes something gets the less ability for downvotes to bury it, if the ratio is close. People always post things that are majority-liked because even something slightly disliked more than liked will bury it.
I've (almost) stopped posting those kinds of replies, because it's just too easy to shut down conversations that go against the common narrative/view.
On the other hand, when people have taken the time to argue logically with me on some unpopular opinion, I've had meaningful back and forth. Sometimes I've even changed my mind.
It's a loss, that for many people being "right" is more important than having meaningful and potentially enlightening conversation that benefits everyone.
Let's argue the ideas, not the ideology they're supposed to be attached to.
This is still the internet, there is a lot of spam, trolling and malicious people.
An unmoderated space doesn't last very long.
But I knwo this is never going to happen because it's too transparent and welcoming diverse opinions and you could suddenly see all those anonymous internet heroes.
Most of the time, and only if you have show dead enabled, you will see that dead comments are that way for a reason. They are low quality, nonsense bots, etc. Very, very rarely do I see a dead comment that represents some well thought counter argument.
It's also easy to confuse the flag link as a downvote. It absolutely is not that. If a person abuses the flag button too much, the flag button gets taken away from them. I should know, it happened to me years ago. I was quite judicious with the flag button because my default behavior from digg/reddit/youtube was to downvote things I disagreed with, didn't like, or thought was of low value.
So it was taken away for a couple of years. I've since gotten the flag button back and today I only use it for what its supposed to be used for.
Case in point, here are the currently dead comments on this post (asterisks are my own on that last one):
"I know I’m shadowbanned but I hope you see my crap, Dang."
"Jesus H Christ"
"It's called "feet" bcuz it's done mainly with the foots :D"
"Just go to the point and ask him if you can suck his d***"
Unless, of course, the topic at hand happens to be a hot issue in the US. At which point anything outside the established narratives (the propaganda) bites the dust almost immediately.
For example I have posted and debated a little about my counter-narrative opinions on "the covid response" here before and had them stay and even gather the occasional upvote, to their credit. But it is also extremely easy to get moderated when walking that line, and likely highly dependent on whether or not the users participating or reading the comments become upset by what they see. This is not just a case of covid in general being an off topic or controversial issue, because you could post until the cows come home about the Covid response if you were cheerleading and parroting the schizophrenic and ever-changing position of western governments and media.
I give them credit at least, I think what they have is better than a lot of places for such debate. You just have to know your place and be extremely careful and polite and factual when you deviate, and you have to be prepared at some point the conversation will be shut down if it is upsetting and angering other users.
I also like that deleted comments can be made visible, another really good transparent part about the moderation. I do agree there is a lot of junk, but it's nice to see what's verboten. I assume posts can be nuked entirely with no trace , I give the benefit of the doubt and hope that's used sparingly.
My biggest issue is the unavoidable US centrism. Again, much better than what's out there, but it's absolutely real. Real talk gets stuck down immediately for being inflammatory
I've noticed that I have to shy away from certain ways of expression when I say things critical of a US-positive narrative. I have to avoid "real talk" as it were. It's kinda frustrating because that's how I prefer to speak but I can also see a difference in the discussion that ensues.
I feel you though. It can be pretty tough when you say something genuinely concerning, which people don't like thinking about, only to have it dismissed with the downvote button. Still, oftentimes such things gain more points while I'm sleeping and seemingly people in other parts of the world view my comments (which I think is kinda telling; it's what it is).
(For context, I am a born US citizen.)
https://hn.algolia.com/?query=whataboutism%20by:dang&dateRan...
Whataboutism […] denotes in a pejorative sense a procedure in which a critical question or argument is not answered or discussed, but retorted with a critical counter-question which expresses a counter-accusation. From a logical and argumentative point of view it is considered a variant of the tu-quoque [1] pattern, which is a subtype of the ad-hominem [2] argument.
[1] Attacking the arguer for hypocrisy.
[2] Attacking the arguer rather than the argument.
If you find yourself accused of whataboutism it tends to mean you’re engaged in political rhetoric rather than intelligent focussed debate - which is fine, but not (as far as I understand it) what HNs is for.
What I'm saying is that accusations of "whataboutism" are the sort of canned argument we want to avoid on HN, as well as an internet trope (which we also want to avoid), a shallow dismissal (which the site guidelines ask to avoid) and a form of name-calling (ditto).
It's also a logical fallacy, since pointing out a contradiction via a comparable situation is a reasonable thing to do. No doubt people do sometimes bring up comparable situations in unhelpful, irrelevant ways, but if so, it needs to be explained how—not met with a lazy label like "whataboutism", which is a blanket that smothers a good argument just as well as a bad one.
Btw, internet users are orders-of-magnitude too quick to jump to the conclusion that they're dealing with "disinformation campaigns" and that sort of thing. As far as I can tell, most of the time they're just encountering other, equally disgruntled internet users who happen to have the opposing view of a topic and feel just as strongly about it. But even if it's a sinister campaigner masquerading as a disgruntled internet user, the solution is still what I described in the previous paragraph. That's the only thing that can work in the long run.
There's another, highly underrated and effective way to respond, though, and that is to simply ignore it and walk away. Responding to these things feeds them. If people would simply starve bad information/arguments of attention, they would fizzle out.
The firehose of bullshit makes responding to every bad argument untenable. It's a DDoS of online discussion. Making a counterargument is one strategy against toxic and harmful online content, but it's not the only one and not always an effective or appropriate one.
>That's the only thing that can work in the long run.
The only thing that works in the long run is removing and banning those who engage (knowingly or otherwise) in disinformation distribution.
> There's another, highly underrated and effective way to respond, though, and that is to simply ignore it and walk away. Responding to these things feeds them.
This is demonstrably false. As a particularly bad example, Rwanda learned this one the hard way. Russia wants Ukraine to lose popular support - if that happens many more innocent people will die than would otherwise. Ignoring the guy on the radio who tries to make others question you and wants to cause you harm will not prevent the machete from slicing you open.
>If people would simply starve bad information/arguments of attention, they would fizzle out.
They get even less attention when the material is removed and the user reasonable prevented from spreading more of it. This isn't high school debate club, it's a nation-state that wants to kill its neighbor's civilians.
People highlight instances of whataboutism to inform others and to point it out to moderators. Punishing them for doing that is clearly the objectively wrong thing to do.
Yes, I am aware what the guidelines say about this.
Usually when people post things that they know are against the majority view, they compensate by adding extra invective or vitriol into their post. Then the post gets flagged or moderated, but because of the invective/vitriol, not because it's a minority/contrarian view. It's not easy to find a way out of this dynamic but I think it's the most common case.
Past explanations about this:
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
That leaves me the option of not saying anything, getting modded down/flagged for pointing out blatant hypocrisy, or spending an hour crafting a long-winded comment that is either murky enough to confuse possible down votes, or clearly boils down to the simple controversial statement that many don't like, at which point it bites the dust anyway. So, where's the incentive to do better?
In my experience, the "vitriol" usually comes the above dynamic plays for the nth time, which is the opposite of what you are saying.
Maybe you are unable to recognise this because it appeals to your own biases. Or maybe you do but it simply suits you. Either way, I am not entirely convinced that I should take your opinions at face value and assume good faith no matter what.
I believe a word was missed out at a crucial point so I'm also struggling with the point. (The caps are for clarity with my request, not some kind of slap down! I'm genuinely interested).
Not always but mostly, I'd say.
But with topics where there is a media and government wide push against ‘desinformation’ like during COVID, a statement like ‘two weeks to stop the spread doesn’t work’ or ‘the vaccines will not end the pandemic’ the hivemind medical science parrots will very swiftly act to kill and bury this dissent under the proscribed talking points. And yes, that is frustrating and causes crass reactions.
Especially the pandemic really has opened my eyes to a very significant weakness in our society, media, governance and scientific institutions and in the general common sense of the population and its willingness to be manipulated. And it reflects in this site. And it’s not limited to COVID, there are many topics that are just untouchable like most of Woke.
That is not a criticism of the moderation team, it’s on the people who vote on comments and I don’t have a solution either. Personally I just don’t care about the points and try to remain reasonably distant to avoid becoming toxic.
>For me, you disqualify your comment by using 'This.' and by using strong words that are borderline insults. To 'non-progressives' it may look like dissenting comments are downvoted but my guess is that people are just protecting the non-aggressive spirit of the site. HN needs far more moderation resources if the discourse descends into aggressive territory. To discuss controversial topics, non-mainstream arguments should be written as non-aggressively as OP's argument about overtone windows.
I am sure that without aggression, fair discussions about controversial topics can be held on HN. What I miss is a path to pass the knowledge about how to argue. My guess is that dang doesn't message korroziya and no user can reach him. People like korroziya learn to 'hold back'.
It's more like passive-aggresive here. It's also harder to tell signal from noise if someone has to write a 500-word essay instead of saying "you're an idiot".
The echo chamber is hard to miss when you see it, but it's pretty remarkable how rosy tinted some users view the site. Censorship exists in a real way here, and it suppresses the flow of information and ideas.
Meta moderation could deal with these users more effectively. Take points from their precious reputation and ultimately take away their moderation rights in case of repeat offenders.
The second sub-thread (for me) is often much more interesting.
I also observed that it looks like there are two "waves" of commentators, the first one has a pack mentality, if they don't like what they see they downvote instead of discussing.
The second "wave" often upvote what was previously downvoted (and it helps that an algorithm removes the downflag button past a few days).
I very rarely see dead comments that did not deserve it (including occasionally my own), and I see all of them in the threads I read. Do you have any example?
> On any topic, there is an acceptable HN Overton window of acceptable opinions and it skews very Silicon Valley.
Reasonably constructed comments can be downvoted, but are very rarely hidden. And the tone of the community varies dramatically depending on the story and its subject. HN is not a homogenous hive mind.
However, I put as much effort in those comments as I do in any other ones, as none of it is trolling or just attempts to be edgy, the comments try to fully abide by HN guidelines.
With that in mind, I have never gotten flagged (unless i missed it somehow), despite definitely being downvoted for "unpopular" opinions before (especially in hot-topic US politics adjacent threads). I know that's just a sample size of 1, of course, and I don't doubt there are some who got flagged unjustly. But from my personal experience, it doesn't seem like a common scenario.
There was this recent doozey https://news.ycombinator.com/item?id=34707305
I doubt this is true but oh well, a man can dream can’t he?
With terminal utilities, the potential for automation is continuous - until the process reaches Shannon limit and can no longer be improved.
That will almost never happen for a GUI program, they usually stop far short, and often cannot be improved/fixed at all.
Of course, this sometimes has the echo chamber effect where it's difficult to get an alternative point of view across, but I'd rather have that, than see threads devolve into wild tangents and shout matches. There's always Reddit for that.
> this is nothing else than Reddit sub for IT guys
Hard disagree. There are plenty of IT subreddits where the discussion is nothing like here. There's actually less of a hivemind here than on any subreddit, as there is some tolerance of dissenting opinions, as long as they're presented tactfully and they make the discussion more interesting.
On most reddit subs, you get banned/flagged/downvoted for simply disagreeing with the hivemind of a given subreddit, your tone doesn't matter.
On HN, I found the exact inverse, you can bring up dissenting points with no issues, at worst you are going to get downvoted (but even that doesn't happen as much as one would expect), as long as you stay polite and reasonable. But if you are being inflammatory or act like an asshole, even if your opinion aligns perfectly with the hivemind of a given comment section, you are getting downvoted/flagged (often by the same people who would agree with your point otherwise).
That alone makes the quality of discussion and the entire interaction with HN community massively different from reddit, and for the better.
That's true, and in this sense the moderators proper are really just a proxy for the community. We don't know how to build an entirely self-moderating community (I wish we did...it would be so much less work) but we do know how to take the temperature of the community and fill in the bits that the system can't accomplished unsupervised. Or at least that's what we're gunning for.
Sure, HN has "smarter" discussions and likes to masquerade as some sort of upscale establishment, high and above the petty squabbles of Eslewhere, but when it comes to any topic on which people have varied opinions, this place too devolves into a predictable echo chamber just like the rest of them.
This did not happen overnight; I noticed this trend deliberately forming a few years ago, when I started seeing perfectly fine comments getting buried almost instantly.
This community does not brook any disagreement, because this service is not designed for it, and the mods cannot appreciate these problems until they try participating as a regular user.
I don’t think it’s up to Dang to correct that.
HN can both be a very well moderated network AND one that doesn’t allow for a certain type of discussions because of the up/down vote system. Both things can be true at the same time.
He is also banning users (like me) from submitting articles with "You are posting too fast" BS instead saying honestly "You are banned from submitting articles because XX", while I can post comments without any issues.
I do know that I only had to deal personally with dang once and it was a very enjoyable experience.
As for user abusing flagging, again, that’s up to the users. It’s not up to dang to override what most users do I believe.
My guy, that is literally what moderation is.
* Users can be banned
* IPs can be banned
* Websites can be banned
* Posts can be deleted
* Submissions can be deleted
* Submissions can be merged
* Submissions can be renamed
Those are the main tools I know of
> their posts are rate limited
I don't like this one... if your comments are shadow banned or automatically hidden.. what's the point...
There's also a lot of editing going on, but most of the time, they let you know what was edited, which is nice, thanks.
When a "veteran" user flags a post, it apparently also gets a LOT of weight from what I heard... It might be why in some cases a post with a lot of momentum quickly falls off from the front page. There's probably other reasons too.
I know that if a user has multiple accounts, they also somehow connect them together to the same person (not sure if it is automatic). They connected multiple of my account but not all... Dang confirmed it (I try to increase privacy by switching account once in a while)
When you get 500 karma, you can down-vote...
The point is to prevent your continued participation in the community from doing more harm or, in the case of filters like rate limiting, to frustrate you until you leave or apologize and ask for the filter to be removed.
But if it helps, they didn't even used to warn people beforehand. PG just pushed the button on you and you might not even know you were shadowbanned for years. At least you can vouch up unfairly banned comments now, and the mods give you warning.
Oh I missed that one. It's separate from having your comments down ranked.
“You’re posting too fast, please slow down” is just a signal that I have enough free time to find something better to do.
I emailed once, got a positive response and am still rate limited so, whatever…
Shadow banning is a key moderation tool. If a user is trolling and they get banned, they just make a new account. If they get shadow banned, then no one engages with their content and they loose interest and go away.
Wouldn't surprise me.
If you consider that HN logs every hit, click, vote, thread collapse, etc., then it is big data.
One of the best moderation bots out there uses a baysian approach to predict malicious changes to wiki edits.
Regardless - they’re still quite green in the grand scheme of things and lack transparency or ability to dynamically shift at their core.
Won’t take long to correct but for now, humans are still much better as crappy as they are.
Most of the moderation action on the site comes from readers --- what's voted up or flagged off the submissions queue, content moderation, and flags. If I'm remembering correctly, "Flagged" means that it's user actions, "Dead" can be either reader or moderator intercession (see, e.g., <https://news.ycombinator.com/item?id=12984398> and <https://news.ycombinator.com/item?id=8628891> and <https://news.ycombinator.com/item?id=30273707>). Users may also be banned, which will be tagged to "dead" comments or submissions.
For specific actions, such as title edits or URL substitutions, email the mods: hn@ycombinator.com I've been told that including the submission or comment ID ("34920400" is the submission ID for this post) is useful as there are automated tools that latch into this and assist with the moderation task itself.
I try to use the following format for most of my mod correspondence:
Subject: <Issue>: <item title as presently presented> <item ID>
Body:
<post URL>
<link URL>
Brief issue description, e.g., clickbait title, bad automunging ("how", enumerations, and a few other terms often get truncated.), link disambiguation suggesting a better source, etc.
If possible: a suggested fix. Mostly as this makes the mod's action easier --- they can use my suggestion *or* select something different, but don't have to start from scratch in identifying a solution.
If another member's identified the issue, I'll usually point to that and the relevant comment.
Close with "Thanks".
Example description from a recent comment: <https://news.ycombinator.com/item?id=34768459> (I'd either copied that from, or to, the email I'd sent the mods, full email would have included the title and post URL as described above.)
Issues I raise are usually title or URL related, and occasionally nominations for the 2nd-chance queue (see <https://news.ycombinator.com/pool>). I've found that most spam and self-promoting submissions / users tend to get picked up without email, though I'll note cases that seem to have been missed. I'll rarely highlight a grossly-offending comment though those tend to get noticed by mods in most cases. My email frequency ranges from a few per week to a few per month.
There are numerous automated features, including some spam and dupe detection, weighting of various domains, and some "cooling off" algorithms for deeply-nested threads / highly-active commenters, and some stricter standards for "green" (new) accounts. These seem to work without direct moderator interaction in most cases, though they may misfire.
Searching the HN archives for dang's comments (and specific keywords or topics) is a good way to get a fuller sense of site guidelines and standards than is explained in the Guidelines and FAQ (linked at the bottom of most pages). Dang frequently links earlier comments himself.
I usually read with "showdead" on, and may flag or vouch dead comments as seems appropriate. Occasionally (rarely) emailing a vouch for something that seems particularly unfairly killed.
If there's an area with which I've disagreements with dang over moderation, it's that I'd grant far more latitude to negative emotion and tone in comments from a viewpoint that's disadvantaged, intellectually rigorous, and hews closer to truth. There's a thread from over a year ago now ... which still bothers me, in that the original submission itself was a shallow ideological rant. Dang's admonishment in that case was itself flagged (which I believe is highly unusual). I do understand that HN's core consideration is "intellectual curiosity and thoughtful conversation" (see: <https://news.ycombinator.com/item?id=13108404>). Though there are times I feel that's better served by highlighting grave faults in source material or earlier comments.
It's worth noting that that's probably the strongest disagreement I've had in my time on HN, certainly for many years, so it's exceptional in that regard. And that the instance is arguably close to the border, with my disagreement being as to which side it fell. I'll occasionally sense the same friction occurring in other threads. I've observed for some time that there are issues HN seems unable to discuss meaningfully, a frustration I've seen dang hint at occasionally as well.
But in general HN has been and remains among the best general discussion sites I've known online since the mid-aughts. And I've been engaged in online discussions since the 1980s.