Hacker News new | past | comments | ask | show | jobs | submit login

Every single social media platform that has ever existed makes the same fundamental mistake. They believe that they just have to remove or block the bad actors and bad content and that will make the platform good.

The reality is everyone, myself included, can be and will be a bad actor.

How do you build and run a "social media" product when the very act of letting anyone respond to anyone with anything is itself the fundamental problem?




You're confusing bad actors with bad behavior. Bad behavior is something good people do from time to time because they get really worked up about a specific topic or two. Bad actors are people who act bad all the time. There may be some of those but they're not the majority by far (and yes, sometimes normal people turn into bad actors because they get upset about a given thing that they can't talk about anything else anymore).

OP's argument is that you can moderate content based on behavior, in order to bring the heat down, and the signal to noise ratio up. I think it's an interesting point: it's neither the tools that need moderating, nor the people, but conversations (one by one).


I think that's right. One benefit this has: if you can make the moderation about behavior (I prefer the word effects [1]) rather than about the person, then you have a chance to persuade them to behave differently. Some people, maybe even most, adjust their behavior in response to feedback. Over time, this can compound into community-level effects (culture etc.) - that's the hope, anyhow. I think I've seen such changes on HN but the community/culture changes so slowly that one can easily deceive oneself. There's no question it happens at the individual user level, at least some of the time.

Conversely, if you make the moderation about the person (being a bad actor etc.) then the only way they can agree with you is by regarding themselves badly. That's a weak position for persuasion! It almost compels them to resist you.

I try to use depersonalized language for this reason. Instead of saying "you" did this (yeah that's right, YOU), I'll tell someone that their account is doing something, or that their comment is a certain way. This creates distance between their account or their comment and them, which leaves them freer to be receptive and to change.

Someone will point out or link to cases where I did the exact opposite of this, and they'll be right. It's hard to do consistently. Our emotional programming points the other way, which is what makes this stuff hard and so dependent on self-awareness, which is the scarcest thing and not easily added to [2].

[1] https://news.ycombinator.com/item?id=33454968

[2] https://news.ycombinator.com/item?id=33448079


You definitely hit the nail on the head.

If someone points out a specific action I did that can/should be improved upon (and especially if they can tell me why it was "bad" in the first place), I'm far more likely to accept that, attempt to learn from it, and move on. As in real life, I might still be heated in the moment, but I'll usually remember that when similar cues strike again.

But if moderation hints at something being wrong with my identity or just me fundamentally, then that points to something that _can't be changed_. If that's the case, I _know they are wrong_ and simply won't respect that they know how to moderate anything at all, because their judgment is objectively incorrect.

Practically at work, this has actually been a good policy you described when I think about bugs and code reviews.

> "@ar_lan broke `main` with this CLN. Reverting."

is a pretty sure-fire way to make me defend my change and believe you are wrong. My inclination, for better or worse, will be to dispute the accusation directly and clear my name (probably some irrational fear that creating a bug will go on a list of reasons to fire me).

But when I'm approached with:

> "Hey, @ar_lan. It looks like pipeline X failed this test after this CLN. We've automatically reverted the commit. Could you please take a second look and re-submit with a verification of the test passing?"

I'm almost never defensive about it, and I almost always go right ahead to reproducing the failure and working on the fix.

The first message conveys to me that I (personally) am the reason `main` is broken. The second conveys that it was my CLN that was problematic, but fixable.

Both messages are taken directly from my companies Slack (ommitting some minor details, of course), for reference.


> I try to use depersonalized language for this reason. Instead of saying "you" did this (yeah that's right, YOU), I'll tell someone that their account is doing something, or that their comment is a certain way. This creates distance between their account or their comment and them, which leaves them freer to be receptive and to change.

I feel quite excited to read that you, dang, moderating HN, use a similar technique that I use for myself and try to teach others. Someone told my good friend the other day that he wasn't being a very good friend to me, and I told him that he may do things that piss me off, annoy me, confuse me, or whatever, but he will always be a good friend to me. I once told an Uber driver who told me he just got out of jail and was a bad man, I said, "No, you're a good man who probably did a bad thing."

Thank you for your write-up.


I think your moderation has made me better at HN, and consequently I'm better in real life. Actively thinking about how to better communicate and create environments where everyone is getting something positive out of the interaction is something I maybe started at HN, and then took into the real world. I think community has a lot to do with it, like "be the change you want to see".

But to your point, yeah my current company has feedback guidelines that are pretty similar: criticize the work, not the worker, and it super works. You realize that action isn't aligned with who you want to be or think you are, and you stop behaving that way. I mean, it's worked on me and I've seen it work on others, for sure.


> I try to use depersonalized language for this reason. Instead of saying "you" did this (yeah that's right, YOU), I'll tell someone that their account is doing something, or that their comment is a certain way. This creates distance between their account or their comment and them, which leaves them freer to be receptive and to change.

I use this tactic with my kids when they do something wrong. Occasionally I slip up and really lay into them, but almost all of the time these days I tell them that I love them, I think they are capable of doing the right thing, but I didn't love some action they did or didn't do and I explain why. They may not be happy with this always, or with the natural (& parent-imposed) consequences of their actions, but it reinforces that they have a choice to do good in the future even if they slip up from time to time. If all of us were immutably identified by the worst thing we ever did, no one would have any incentive to change.

Thanks for the thoughtful & insightful comment, dang.


I think you do a good job on HN and I appreciate, as someone who moderated a similarly large forum for a long time, how candid you are in your communications on and off site. You're also a very quick email responder!

> I try to use depersonalized language for this reason. Instead of saying "you" did this (yeah that's right, YOU), I'll tell someone that their account is doing something, or that their comment is a certain way. This creates distance between their account or their comment and them, which leaves them freer to be receptive and to change.

My sense is that this is a worthy thing to do (first of all because it's intellectually correct to blame actions rather than people, and second of all because if you're right about the effect it's all upside). But I suspect this will produce very little introspection, maybe a tiny but on the margins.

It's pretty normal in an argument between two people IRL that one will say something like "That was a stupid comment" or "Stop acting like an asshole" -- both uses of distancing language -- and the other person will respond "Don't call me stupid" or "Don't call me an asshole". I think most people who are on the receiving end of even polite correction are going to elide the distancing step.

On the social psych side, I have no idea whether there's any validated way to encourage someone to be more introspective, take a breath, try to switch down into type-II processing, etc.


The other tricky thing is a bad actor will work to stay just this side of the rules while still causing damage and destruction to the forum itself.


Yes. But in our experience to date, this is less common than people say it is, and there are strategies for dealing with it. One such strategy is https://hn.algolia.com/?sort=byDate&dateRange=all&type=comme... (sorry I don't have time to explain this, as I'm about to go offline - but the key word is 'primarily'.) No strategy works in all cases though.


You're doing a good job dang.

... kinda wondering if this is the sort of OT post we're supposed to avoid, it would be class if you chastised me for it. But anyway, glad you're here to keep us in check and steer the community so well.


For stuff like that I go by what pg wrote years ago: https://news.ycombinator.com/newswelcome.html

Empty comments can be ok if they're positive. There's nothing wrong with submitting a comment saying just "Thanks." What we especially discourage are comments that are empty and negative—comments that are mere name-calling.

It's true that empty positive comments don't add much information but they have a different healthy role (assuming they aren't promotional)


This "impersonal" approach to also works in the other direction. Someone who said something objectively bad once doesn't have to be a "known bad person" forever.


> persuade the user to behave differently

That scares me. Today's norms are tomorrow's taboos. The dangers of conforming and shaping everyone into the least controversial opinions and topics are self evident. It's an issue on this very forum. "Go elsewhere" doesn't solve the problem because that policy still contributes to a self-perpetuating feedback loop that amplifies norms, which often happen to be corrupt and related to the interests of big (corrupt) commercial and political powers.


I don't mean persuade them out of their opinions on $topic! I mean persuade them to express their opinions in a thoughtful, curious way that doesn't break the site guidelines - https://news.ycombinator.com/newsguidelines.html.


Sufficiently controversial opinions are flagged, downvoted til dead/hidden, or associated users shadow banned. HN's policies and voting system, both de facto and de jure, discourage controversial opinions and reward popular, conformist opinions.

That's not to pick on HN, since this is a common problem. Neither do I have a silver bullet solution, but the issue remains, and it's a huge issue. Evolution of norms, for better or worse, is suppressed to the extent that big communication platforms suppress controversy. The whole concept of post and comment votes does this by definition.


Almost completely disagree.

There are a few sacred cows here (I won't mention them by name, though the do exist), but I have earned my rep by posting mostly contrarian opinions, and I almost always have quite a few net upvotes - sometimes dozens. It's not too difficult: First, I cite facts that back up my claims from sources whose narratives would typically go against my argument. I cite the New York Times, Washington Post, the Atlantic, NPR, CNN, etc.; I only rarely cite Fox News, and never cite anything to the right of Fox. Second, I really internalize the rules about good faith, not attacking the weakest form of an argument, not cross-examining, etc. Sometimes I have a draft that has my emotions, and I'll edit it to make it more rational before posting. Third, I ask open-ended questions to allow myself to be wrong in the mind of other commenters. Instead of just asserting that some of my ultra-contrarian opinions are the only way anyone can see an issue, I may propose a question. By doing that, I have at times seen some excluded middle I hadn't considered, and my opinion becomes more nuanced. Fourth, I often will begin replying and then delete my reply because I know it won't add anything. This is the hardest one to do, but sometimes it's just the way you have to go. Some differences are merely tastes and preferences, and I'm not going to change the dominant tastes and preferences of the Valley on HN. I can only point out some of the consequences.

The content moderation rules and system here have encouraged me to write better and more clearly about my contrarian opinions, and have made me more persuasive. HN can be a crap-show at times, but in my experience, it's often some of the best commentary on the Internet.


Completely disagree about HN. Controversial topics that are thought out, well formed, and argued with good intent are generally good sources of discussion.

Most of the time though, people arguing controversial topics phrase them so poorly or include heavy handed emotions so that their arguments have no shot of being fairly interpreted by anyone else.


That's true to an extent (and so is what ativzzz says, so you're both right). But the reasons for what you're talking about are much misunderstood. Yishan does a good job of going into some of them in the OP, by the way.

People always reach immediately for the conclusion that their controversial comments are getting moderated because people dislike their opinion—either because of groupthink in the community or because the admins are hostile to their views. Most of the time, though, they've larded their comments pre-emptively with some sort of hostility, snark, name-calling, or other aggression—no doubt because they expect to be opposed and want to make it clear they already know that, don't care what the sheeple think, and so on.

The way the group and/or the admins respond to those comments is often a product of those secondary mixins. Forgive the gross analogy, but it's as if someone serves a shit milkshake and when it's rejected, say, "you just hate dairy products" or "this community is so biased against milkshakes".

If you start instead from the principle that the value of a comment is the expected value of the subthread it forms the root of [1], then a commenter is responsible for the effects of their comments [2] – at least the predictable ones. From that it follows that there's a greater burden on the commenter who's expressing a contrarian view [3]. The more contrarian the view—the further it falls outside the community's tolerance—the more responsibility that commenter has for not triggering degenerative effects like flamewars.

This may be counterintuitive, because we're used to thinking in terms of atomic individual responsibility, but it's a model that actually works. Threads are molecules, not atoms—they're a cocreation, like one of those drawing games where each person fills in part of a shared picture [4], or like a dance—people respond to the other's movements. A good dancer takes the others into account.

It may be unfair that the one with a contrarian view is more responsible for what happens—especially because they're already under greater pressure than the one whose views agree with the surround. But fair or not, it's the way communication works. If you're trying to deliver challenging information to someone, you have to take that person into account—you have to regulate what you say by what the listener is capable to hear and to tolerate. Otherwise you're predictably going to dysregulate them and ruin the conversation.

Contrarian commenters usually do the opposite of this—they express their contrarian opinion in a deliberately aggressive and uncompromising way, probably because (I'm repeating myself sorry) they expect to be rejected anyhow, and it's safer to be inside the armor of "you people can't handle the truth!" than it is to really communicate, i.e. to connect and relate.

This model is the last thing that most contrarian-opinion commenters want to adopt, because it's hard and risky, and because usually they have pre-existing hurt feelings from being battered repeatedly with majoritarian opinions already (especially the case when identity is at issue, such as being from a minority population along some axis). But it's the one that actually has a hope of working, and is by far the best solution I know of to the problem of unconventional opinions in groups.

Are there some views which are so far beyond the community's tolerance that any mention in any form will immediately blow up the thread, making the above model impossible? Yes, but they're rare and extreme and not usually the thing people have in mind. I think it's better to stick to the 95% or 99% case when having this discussion.

[1] https://hn.algolia.com/?dateRange=all&page=0&prefix=true&sor...

[2] https://hn.algolia.com/?dateRange=all&page=0&prefix=true&sor...

[3] https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

[4[ https://news.ycombinator.com/item?id=6813226


Just wanted to say that it's great to have you posting your thoughts/experience on this topic. I've run a forum for almost 19 years as a near-lone moderator and so have a lot of thoughts, experience and interest in the topic. It's been frustrating when Yishan's posted (IMO, solid) thoughts on social networks and moderation and the bulk of HN's discussion can be too simple to be useful ("Reddit is trash", etc).

I particularly liked his tweet about how site/network owners just wish everyone would be friendly and have great discussions.


Indeed—if only.

I'd like to hear your thoughts/experience as well.


> The more contrarian the view—the further it falls outside the community's tolerance—the more responsibility that commenter has for not triggering degenerative effects like flamewars.

This sounds similar to the “yelling fire” censorship test

it’s not that we censor discussing combustion methods, there would be no effect if everyone else was also yelling fire

But people were watching a movie and now the community’s experience has been ruined (with potential for harm), in exchange for nothing of value


And bad behavior gets rewarded with engagement. We learned this from "reality television" where the more conflict there was among a group of people the more popular that show was. (Leading to producers abandoning the purity of being unscripted in the pursuit of better ratings.) A popular pastime on Reddit is posting someone behaving badly (whether on another site, a subreddit, or in a live video) for the purpose of mocking them.

When the organizational goal is to increase engagement, which will be the case wherever there are advertisers, inevitably bad behavior will grow more frequent than good behavior. Attempts to moderate toward good behavior will be abandoned in favor of better metrics. Or the site will stagnate under the weight of the new rules.

In this I'm in disagreement with Yishan because in those posts I read that engagement feedback is a characteristic of old media (newspapers, television) and social media tries to avoid that. The OP seems to be saying that online moderation is an attempt to minimize controversial engagement because platforms don't like that. I don't believe it. I think social media loves controversial engagement just as much as the old-school "if it bleeds, it leads" journalists from television and newspapers. What they don't want is the (quote/unquote) wrong kind of controversies. Which is to say, what defines bad behavior is not universally agreed upon. The threshold for what constitutes bad behavior will be different depending on who's doing the moderating. As a result the content seen will be influenced by the moderation, even if said moderation is being done in a content-neutral way.

And I just now realize that I've taken a long trip around to come to the conclusion that the medium is the message. I guess we can now say the moderation is the message.


++

A giant amount of social quandaries melt away when you realize:

"Good guys" and "Bad guys" is not a matter of identity, it's a matter of activity.

You aren't a "Good guy" because of who you are, but because of what you do.

There are vanishingly few people who as a matter of identity are reliably and permanently one way or another.


> Bad actors are people who act bad all the time

I'd argue that bad actors are people that behave badly "on purpose". Their goals are different than the normal actor. Bad actors want to upset or scare people. Normal actors want to connect with, learn from, or persuade others.


I think you agree with the parent.

They pointed out that everybody can be a bad actor and you will not find a way to get better users.


I can "behave well" and still be a bad actor in that I'm constantly spreading dangerous disinformation. That disinformation looks like signal by any metadata analysis.


Yes, that's probably the limit of the pure behavioral analysis, esp. if one is sincere. If they're insincere it will probably look like spam; but if somebody truly believes crazy theories and is casually pushing them (vs promoting them aggressively and exclusively), that's probably harder to spot.


It's not a mistake. It's a PR strategy. Social media companies are training people to blame content and each other for the effects that are produced by design, algorithms and moderation. This reassigns blame away from things that those companies control (but don't want to change) to things that aren't considered "their fault".


Its really all media not just social media that profits from propaganda. Turn on CNN. You might agree with what they are saying versus what the talking heads on fox news are saying, but they use the same state of constant panic style of reporting because that works really well to fix eyeballs on advertisements, both overt ones and the more subtle ones that happen during the programming.


That is very much a problem in the US (AFAIK) where news and entertainment are merged. Other countries have laws to ensure that news are presented emotionless and factual.


This is something Riot Games has spoken on, the observation that ordinary participants can have a bad day here or there, and that forgiving corrections can preserve their participation while reducing future incidents.


Did Riot eventually sort out the toxic community? If so that would be amazing, and definitely relevant. I stopped playing when it was still there, and it was a big part of the reason I stopped.


Anecdotally: outright flame / chat hate is a bit better than it used to be, but not much.


I’ve been playing very very active from 2010 to 2014 and since then on-off, sometimes skipping a season.

Recently picked it up again and I noticed that I didn’t had to use /mute all anymore. I’ve got all-chat disabled by default so I’ve got no experience there, but overall I’d say it has come a long way.

But I’d also say it depends which mode and MMR you are in. I mostly play draft pick normals or ARAMs in which I both have a lot of games played - I heard from a mate that chat is unbearable in low level games.


The only success I've seen in sorting out random vitriol is cutting chat off entirely and minimizing methods of passive aggressive communication. But Nintendo's online services haven't exactly scaled to the typical MOBA size to see how it actually works out


It sounds like a insurmountable problem. What makes this even more interesting to me is that HN seems to have this working pretty well. I wonder how much of it has to do with clear guidelines of what should be valued and what shouldn't and having a community that buys in to that. For example one learns quickly that Reddit-style humor comments are frowned upon because the community enforces it with downvotes and frequently explanations of etiquette.


If we follow the logic of Yishan's thread, HN frowns upon and largely doesn't allow discussion that would fall into group 3 which removes most of the grounds for accusations of political and other biases in the moderation. As Yishan says, no one really cares about banning groups 1 and 2, so no one objects to when that is done here.

Plus scale is a huge factor. Automated moderation can have its problems. Human moderation is expensive and hard to keep consistent if there are large teams of individuals that can't coordinate on everything. HN's size and its lack of desire for profit allow for a very small human moderation team that leads to consistency because it is always the same people making the decisions.


I have a suspicion that the medium is the message at HN:

No pictures and no avatars.

I wonder how much bad behavior is weeded out by the interface itself ?

A lot, I suspect …


Nope. There's been abuse in text-only environments online since forever. And lots of people have left (or rarely post on) HN because of complaints about the enviroment here.


> No pictures and no avatars

This is essentially moderation rule #0. it is unwritten, enforced before violation can occur, and generates zero complaints because it filters complainers out of the user pool from the start.


The no-avatars rule also takes away some of the personalization aspect. If you set your account up with your nickname, your fancy unique profile picture and your favorite quote in the signature, and someone says you're wrong, you're much more invested because you've tied some of your identity to the account.

If you've just arrived on the site, have been given a random name and someone says you're wrong, what do you care? You're not attached to that account at all, it's not "you", it's just a random account on a random website.

I thought that was an interesting point on 4chan (and probably other sites before them), that your identity was set per thread (iirc they only later introduced the ability to have permanent accounts). That removes the possibility of you becoming attached to the random name.


Why would one be concerned with being wrong at all? Being wrong, thus being able to learn, is the whole reason for having discussions with others.

Once you’re confident that you can’t be wrong, you’re not going to care about the topic anymore. There is good reason why we don’t sit around talking about how 1+1=2 all day.


I suspect HN succeeds due to heavy moderation, explicit community guidelines and a narrow topic set.


Some areas of reddit do similar things with similar results. AskHistorians and AskScience are the first two to come to mind.

This may be a lot easier in places where there's an explicit point to discussion beyond the discussion itself - StackOverflow is another non-Reddit example. It's easier to tell people their behavior is unconstructive when it's clearly not contributing to the goal. HN's thing may just be to declare a particular type of conversation to be the goal.


I think that has far more to do with this site being relatively low-traffic. Double the traffic, while keeping the exact same rules and topic, and it would become unreadable. It's easy to "moderate" when people clearly break the rules; but "moderation" can't do anything if the only problem is that most comments are uninsightful. Large numbers always ruin things, in real life or online. You can see that on this very website on Musk-related stories, with a terrible heat-to-light ratio in the comments.

It's controversial, but if the average IQ was 120 rather than 100, I doubt you'd have 1/10th as many issues on massively popular social media; most of the moderation issues would go away. The problem comes from the bottom-up, and can't be fixed from the top down.


The only thing HN has going for it imo is its size. Once it becomes a larger and therefore more attractive market for media, the propaganda will be a lot more heavy handed, like what happened to reddit as it grew from something no one used and into a mainstream platform. You definitely see propaganda posted on here already from time to time.


HN works very well, because it's about as far from free speech as you can get on the internet, short of dang personally approving every post.


I think most posts are short lived so they drop off quickly and people move on to new content. I think a lot of folks miss a lot of activity that way. I know I miss a bunch. And if you miss the zeitgeist it doesn’t matter what you say cause nobody will reply.

The twitter retweet constantly amplifies and the tweets are centered around an account vs a post.

Reddit should behave similarly but I think subreddit topics stick longer.


There's also the fact that there's no alerts about people replying to you or commenting on your posts. You have to explicitly go into your profile, click comments, and then you can see if anyone has said anything to you.

This drastically increases time between messages on a topic, lets people cool off, and lets a topic naturally die down.


Very good point about the "fog of war". If HN had a reply-notification feature, it would probably look differently. Every now and then someone builds a notification feature as an external service. I wonder if you can measure change in the behavior of people before and after they've started using it?

Of course, that also soft-forces everyone to move on. Once a thread is a day or two old, you can still reply, but the person you've replied to will probably not read it.


It's proof of the old adage : the best possible government is a benign dictator.


What kind of free speech is not allowed? What can’t you say right now that you feel should?


Category 1 from Yishan's thread, spam, obviously isn't allowed. But also thinking about house general framework of it all coming down to signal vs noise, most "noise" gets heavily punished on here. Reddit-style jokes frequently end in the light greys or even dead. I had my account shadow-banned over a decade ago because I made a penis joke and thought people didn't get the joke.


Free speech doesn't mean you can say whatever, wherever, without any repercussions. It solely means the government can't restrict your expression. On a private platform you abide their rules.


Well, no. Free Speech is an idea, far more expansive than the law as written in the US constitution, or many other countries and their respective law of the land/documents.

Free Speech does mean what you describe it not as. But there is no legal body to punish you for violating the principle. It is similar to 'primum non nocere', translated roughly to 'do no harm', extremely common in medicine and something you may see alongside the 'Hippocratic Oath'. You can be in violation of that principle or the oath at any time, some people even get fired for violating either as a pretext to malpractice. Some even argue that it is quite impossible to abide by this principle, and yet, it is something many people take on as responsibility everyday all across the globe.

I won't argue about websites and their rules, they have their own set of principles and violate plenty of others, sometimes they even violate their own principles. But Free Speech is not just the law and interactions one may have with their government.


You seem to be conflating the general principles of free speech with the specific protections defined by the US first amendment.


There are no "general principles of free speech," there are conflicting principles over which people constantly disagree.


What now? You’re suggesting that the removal of the word “general” turns it into a concept that can exist and be disagreed upon? There can’t be conflicting general principles of free speech over which people consistently disagree? What a bizarre correction.


Where are the people arguing about Donald Trump? Where are the people promoting dodgy cryptocurrencies? Where are the people arguing about fighting duck-sized horses? Where's the Ask HN asking for TV show recommendations?


> The reality is everyone, myself included, can be and will be a bad actor.

But you likely aren't, and most people likely aren't either. That's the entire premise behind removing bad actors and spaces that allow bad actors to grow.


> But you likely aren't, and most people likely aren't either.

Is there any evidence of this? 1% bad content can mean that 1% of your users are bad actors, or it can mean that 100% of your users are bad actors 1% of the time (or anything in between.)


I assume all of us have evidence of this in our daily lives.

Even the best people we know have bad days. But you have probably also encountered people in your life who have consistent patterns of being selfish, destructive, toxic, or harmful.


> you have probably also encountered people in your life who have consistent patterns of being selfish, destructive, toxic, or harmful.

This is not evidence that most bad acts are done by bad people. This is evidence that I've met people who've annoyed or harmed me at one or more points, and projected my personal annoyance into my fantasies of their internal states or of their essence. Their "badness" could literally have only consisted of the things that bothered me, and during the remaining 80% of the time (that I wasn't concerned with) they were tutoring orphans in math.

Somebody who is "bad" 100% of the time on twitter could be bad 0% of the time off twitter, and vice-versa. Other people's personalities aren't reactions to our values and feelings; they're as complex as you are.

As the OP says: our definitions of "badness" in this context are of commercial badness. Are they annoying our profitable users?

edit: and to add a bit - if you have a diverse userbase, you should expect them to annoy each other at a pretty high rate with absolutely no malice.


Your logic makes sense but is not how these moderation services actually work. When I used my own phone number to create a Twitter, I was immediately banned. So instead I purchased an account from a service with no issues. It’s become impossible for me at least to use large platforms without assistance from an expert who runs bot farms to build accounts that navigate the secret rules that govern bans.


> and spaces that allow bad actors to grow

I believe that's GP's point! Any of us has the potential to be the bad actor in some discussion that gets us irrationally worked up. Maybe that chance is low for you or I, but it's never totally zero.

And even if the chance is zero for you or I specifically, there's no way for the site operators to a priori know that fact or to be able to predict which users will suddenly become bad actors and which discussions will trigger it.


Spam is a behavior, not a fundamental trait of the actor.

Would be interesting to make a service where spammers have to do recaptcha-like spam flagging to get their account unlocked.


Which definition of spam are you operating under? I think it is a fundamental trait of the actor.


So you would expect a spammer to only ever post spam, even on their own personal account? Or a spam emailer to never send a personal email?


I know sales bros who live their live by their ABCs - always be closing, but that's besides the point. if the person behind the spam bot one day wakes up and decides to do turn over a new leaf and something else with their life, they're not going to use the buyC1alis@vixagra.com email address they use for sending spam as the basis for their new persona. thus sending spam is inherit to the buyC1alis@vixagra.com identity that we see - of course there's a human being behind it, but as we'll never know them in ant other context, that is who they are to us.


We have laws around mobs and peaceful protest for a reason. Even the best people can become irrational as a group. The groupmind is what we need controls for: not good and bad people.


I think the point is that anyone and/or everyone can be a bad actor in the right circumstances, and moderations job is to prevent those circumstances.


The original post is paradoxical in the very way it talks about social media being paradoxical.

He observes that social media moderation is about signal to noise. Then he goes on about introducing off-topic noise. Then, he comes to conclusions that seem to ignore his original conclusion about it being a S/N problem.

Chiefly, he doesn't show how a "council of elders" is necessary to solve S/N problems.

Strangely enough, Slashdot seems to have a system which worked pretty well back in the day.


I think the key is that no moderation can withstand outside pressure. A community can be entirely consistent and happy but the moment outside pressure is applied it folds or falls.


Slashdot moderation is largely done by the users themselves, acting anonymously as "meta-moderators." I think they were inspired by Plato's ideas around partially amnesiac legislators who forget who they are while legislating.


What about fox news? AM radio? These are bastions of radicalization but they dont let anyone come on and say anything. At the end of the day this sort of rhetoric played by these groups is taught in university communications classes as a way to exert influence. Its all just propaganda at the end of the day, and that can come in the form of a pamphlet, or a meeting in a town hall, or from some talking head on tv, or a tweet. Social media is just another avenue for propaganda to manifest just like how the printing press is.


Some people are much more likely to engage in bad behavior than others. The thing is, people who engage in bad behavior are also much more likely to be "whales," excessive turboposters who have no life and spend all day on these sites.

Someone who has a balanced life, who spends time at work, with family, in nature, only occasionally goes online, uses most of their online time for edification, spends 30 minutes writing a reply if they decide one is warranted - that type of person is going to have a minuscule output compared to the whales. The whales are always online, thoughtlessly writing responses and upvoting without reading articles or comments. They have a constant firehouse of output that dwarfs other users.

Worth reading "Most of What You Read on the Internet is Written by Insane People"[1].

If you actually saw these people in real life, chances are you'd avoid interacting with them. People seeing a short interview with the top mod of antiwork almost destroyed that sub (and lead to the mod stepping down). People say the internet is a bad place because people act badly when they're not face to face. That might be true to some extent, but we're given online spaces where it's hard to avoid "bad actors" (or people that engage in excessive bad behavior) the same way we would in person.

And these sites need the whales, because they rely on a constant stream of low quality content to keep people engaged. There are simple fixes that could be done, like post limits and vote limits, but sites aren't going to implement them. It's easier to try to convince people that humanity is naturally terrible than to admit they've created an environment that enables - and even relies on - some of the most unbalanced individuals.

[1] https://www.reddit.com/r/slatestarcodex/comments/9rvroo/most...


How do you build and run a "social media" product when the very act of letting anyone respond to anyone with anything is itself the fundamental problem?

This isn't the problem as much as giving bad actors tools to enhance their reach. Bad actors can pay to get a wider reach or get/abuse a mark of authority, like a special tag on their handle, getting highlighted in a special place within the app, gaming the algorithm that promotes some content, etc. Most of these tools are built into the platform. Some though, like sock puppets, can be detected but aren't necessarily built in functionality.


At the very least you could be susceptible overreacting because of an emotionally charged issue. Eg. Reddit's boston marathon bomber disaster, when they started trying to round up brown people (actual perp "looked white")

Maybe that wouldn't be your crusade and maybe you would think you were standing up for an oppressed minority. You get overly emotional, and you could be prone to making some bad decisions.

People act substantially differently on reddit vs. hackernews; honestly I have to admit to being guilty of it. Some of the cool heads here are probably simultaneously engaged in flamewars on reddit/twitter.


The business plan of massive user scale, user generated content, user “engagement” with ad driven revenue leads to the perceived issues about polarization and content moderation. That and the company structure are the fundamental problems that attract what we see on social media. The data about users is the product sold to advertisers. The platform only cares about moderation in that it supports the goal of more ad revenue, that is why Yishan said spam moderation was job #1, its more harmful to ad revenue than users with poor behavior.

If a social media company’s mission is to have no barrier, anyone and everyone to share ideas, information and “all are welcome” then maybe a company structure like a worker cooperative [0] would be a better match to that mission statement. No CEO that gets massive pay/stock, instead employees are owners. All employees. They decide what features/projects the company does, how to allocate resources, how to moderate content, etc.

[0] https://en.wikipedia.org/wiki/Worker_cooperative


> The reality is everyone, myself included, can be and will be a bad actor.

Customised filters for anyone, but I am talking about filters completely under the control of the user. Maybe running locally. We can wrap ourselves in a bubble but better that than having a bubble designed by others.

I think AI will make spam irrelevant over the next decade by switching from searching and reading to prompting the bot. You don't ever need to interface with the filth, you can have your polite bot present the results however you please. It can be your conversation partner and you get to control its biases as well.

Internet <-> AI agent <-> Human

(the web browser of the future, the actual web browser runs in a sandbox under the AI)


I'll raise you a forum-trained AI spambot to defeat the AI spamfilter. It'll be an entirely automated arms race.


The critical part is who controls the filter, it doesn't have to be the best to be useful, but it must obey the user.


>How do you build and run a "social media" product when the very act of letting anyone respond to anyone with anything is itself the fundamental problem?

That's...that's not a problem.


Not true at all - everyone has the capacity for bad behaviour in the right circumstances but most people are not, in my opinion, there intentionally to be trolls.

There are the minority who love to be trolls and get any big reaction out of people (positive or negative). Those people are the problem. But they are also often very good at evading moderation or laying in wait and toeing the line between bannable offences and just every so slightly controversial comments.


Having a a verified public Internet /Reputation ID system for those who want to be bad or good publicly is one way!

All others are just trolls not backed up by their verified public Internet / Reputation ID.


The reality is everyone, myself included, can be and will be a bad actor.

Based on this premise we can conclude that the best way to improve Reddit and Twitter is to block everyone.


To be honest, I would not disagree with that. Very 'the only winning move is not to play'.


To be honest, and maybe this will be panned, but the real answer is for people to grow thicker skin and stop putting one's feelings on a pedestal above all.


When I make comments about other people, they need to grow thicker skin.

When other people attack me personally, it's a deep and dangerous violation of social norms.


I don't block people because they hurt my feelings, i block people because im just not interested in seeing bird watching content on my timeline. No one deserves my eyeballs.


Look - I don't even particularly disagree with you, but I want to point out a problem with this approach.

I'm 33. I grew up playing multiplayer video games (including having to run a db9 COM cable across the house from one machine to another to play warcraft 2 multiplayer, back when you had to explicitly pick the protocol for the networking in the game menu)

My family worked with computers, so I had DSL since I have memories. I played a ton of online games. The communities are BRUTAL. They are insulting, abusive, misogynistic, racist, etc... the spectrum of unmonitored teenage angst, in all it's ugly forms (and to be fair, some truly awesome folks and places).

As a result - I have a really thick skin about basically everything said online. But a key difference between the late 90s and today, is that if I wanted it to stop, all I had to do was close the game I was playing. Done.

Most social activities were in person, not online. I could walk to my friend's houses. I could essentially tune out all the bullshit by turning off my computer, and there was plenty of other stuff to go do where the computer wasn't involved at all.

I'm not convinced that's enough anymore. The computer is in your pocket. It's always on. Your social life is probably half online, half in person. Your school work is online. Your family is online. your reputation is online (as evidenced by those fucking blue checkmarks). The abuse is now on a highway into your life, even if you want to turn it off.

It's like the school bully is now waiting for you everywhere. He's not waiting at school - he's stepping into the private conversations you're having online. He's talking to your friends. He's hurling abuse at you when you look at your family photos. He's in your life in a way that just wasn't possible before.

I don't think it's fair to say "Just grow a thicker skin" in response to that. I think growing a thicker skin is desperately needed, but I don't think it's sufficient. The problem is deeper.

We have a concept for people who do the things these users are doing on twitter in person - They're called fighting words, and most times, legally (even in the US) there is zero assumption of protected speech here. You say bad shit about someone with the goal of riling them up and no other value? You have no right of free speech, because you aren't "speaking" - you're trying to start a fight.

I'm not protecting your ability to bully someone. Full stop. If you want to do that, do it with the clear understanding that you're on your own, and regardless of how thick my skin is - I think you need a good slap upside the head. I'd cheer it on.

In person - this resolves itself because the fuckwads who do this literally get physically beaten. Not always - but often enough we have a modicum of civil discussion we accept, and a point where no one is going to defend you because you were a right little cunt, and the beating was well deserved.

I don't know how you simulate the same constraint online. I'm not entirely sure you can, but I think the answer isn't to just stop trying.


> The computer is in your pocket. It's always on. Your social life is probably half online, half in person. Your school work is online. Your family is online. your reputation is online (as evidenced by those fucking blue checkmarks). The abuse is now on a highway into your life, even if you want to turn it off.

It is still a choice to participate online. I'm not on Twitter or Facebook or anything like that. It doesn't affect my life in the slightest. Someone could be on there right this minute calling me names, and it can't bother me because I don't see it, and I don't let it into my life. This is not a superpower, it's a choice to not engage with social media and all the ills it brings.

Have I occasionally gotten hate mail from an HN post? Sure. I even got a physical threat over E-mail (LOL good luck, guy). If HN ever became as toxic as social media can be, I could just stop posting and reading. Problem solved. Online is not real if you just ignore it.


The attitude of "If you don't like it, leave!" is allowing the bullies to win.

Minorities, both racial and gender, should be able to use social media without having vitriol spewed at them because they're guilty of being a minority.


They can use a different social media. That toxic social media then gets a bad reputation.


What happens when every major social network has a bad reputation?

"Start your own" isn't really a viable solution since the greatest challenge of making a social network is getting people to use it.


This is yet another antitrust issue, regulation should be put in place so that a private company cannot have own a platform with a market share large enough to become "the" public square.


That's asking human nature to change, or at least asking almost everyone to work on their trauma until they don't get so activated. Neither will happen soon, so this can't be the real answer.


Interesting, that wasn’t my interpretation of the twitter thread, it was more that spam and not hurtful content was the real tricky thing about moderating social media.


Spam was more of an example than the point, I think -- the argument Yishan is making is that moderation isn't for content, it's for behavior. The problem is that if bad behavior is tied to partisan and/or controversial content, which it often is, people react as if the moderation is about the content.


Paul Pelosi's skin wasn't thick enough to deflect hammers.

(OK, that wasn't a twitter problem, but the attack on him was 100% a product of unmoderated media in general)


I respectfully disagree. Beyond the reason that there is no way you can be 100% certain 'unmoderated media' was the primary motivator. Nobody can presume to know his motivations or inner dialogue. A look at that mans history shows clear mental health issues and self-destructive behavior so we can infer some things but never truly know.

Violence exists outside of mean tweets and political rhetoric. People, even crazy ones, almost always have their own agency even if it runs contrary to what most consider to be normal thoughts and behavior. They choose to act, regardless of others and mostly without concern or conscious. There are crazy people out there and censoring others wont ever stop bad people from doing bad things. If so, then how do we account for the evils done by those prior to our inter-connected world?


We really don't know much. The attacker used drugs, his ex partner said the he went away for year and came back a changed person.

He lived in a community with BLM signs and a rainbow flag. He did hemp jewellry.

He registered a website three months ago and only recently filled it with standard extreme right garbage.

This is all so odd that for all we know someone else radicalized him offline, the old fashioned way.


You hit the nail on the head, but maybe the other way around.

"Block" and "Mute" are the Twitter user's best friends. They keep the timeline free of spam, be it advertisers, or the growth hackers creating useless threads of Beginner 101 info and racking up thousands of likes.


After using several communications tools over the past couple of decades (BBSes, IRC, Usenet, AIM, plus the ones kids these days like), I'm convinced blocking and/or muting is required for any digital mass communication tool anyone other than sociopaths would use.


Doesn't Twitter give the option of a whitelist (Just who you follow + their retweets) already?


Not really, even the it still does recommended tweets and I don’t want to see retweets or likes and you have to turn that off per person.


Yes really: you can get this experience with Twitter: https://lee-phillips.org/howtotwitter/


Edit: I apologize that my original post was dismissive of your effort to help people use twitter.


That's unnecessarily dismissive of someone trying their best to share some tips. It's not like they charged you to read it.


Fair enough it wasn’t meant to be a commentary on them, but I will edit with an apology


Don’t worry about it. I didn’t exactly understand your original comment, but I don’t have a problem with people having opinions.


Yes, really. I never see 'recommended' or 'so-and-so liked...' in my feed.


I had to create a new account after losing mine, and without following many people it seems like easily 30-50% of my feed is recommended content.


There are "home" and "newest" feeds. I agree it's shitty that the default shows this stuff, but you just have to switch it over to "newest."


Charge them $10 to create an account (anonymous, real, parody whatever), then if they break a rule give them a warning, 2 rule breaks, a 24 hour posting suspension, 3 strikes and permanently ban the account.

Let them reregister for $10.

Congrats, i just solved spam, bots, assholes and permanent line steppers.


This is how the SomethingAwful forums operated when they started charging for accounts. Unfortunately it probably wouldn't be useful as a test case because it was/is, at it's core, a shitposting site.


I think metafilter still does this ?


And twitter isn't?


Unless you generate more than $10 from the account. For example in presidential election years in the US billions is spent in advertising the elections. A few PACs would gladly throw cash at astroturf movements on social media even at the risk of being banned.


Sounds good to me. That would mean that your energy in moderation would directly result in income. If superpacs are willing to pay $3.33 a message, that's a money-spinner.


This was the strategy at the SomethingAwful forums. They seemed pretty well moderated, but definitely never hit the scale of Reddit or Twitter.


Having posted there in its heyday, it made for an interesting self-moderation dynamic for sure. Before I posted something totally offbase that I knew I'd be punished for, I had to think "is saying this stupid shit really worth $10 to me?". Many times that was enough to get me to pause (but sometimes you also can't help yourself and it's well worth the price).


You solved bots, but destroyed the product.


I mean, who here remembers app.net? Love the Garry Tan endorsement! https://web.archive.org/web/20120903182620/https://join.app....

EDIT: Lol Dalton's PART of YC now. Hey dude, why not pitch it then


I don't even know if it solved bots. Rich countries, rich organizations, rich people could do a lot. $100M would buy you 10M bots.


I think the idea is that it shifts the incentives. Sure, a rich nation state could buy tons of bot accounts at $10 a pop. But is that still the most rational path to their goal? Probably not, because there are lots of other things you can do for $100M.


The problem is the meaning of those rules. Any rule looks reasonable when it is written down. And after some time it becomes a weapon.

For instance, the three deletions (forget the exact term) rule in wikipedia. It is now a tool used by “the first to write”…


This kind of thing worked for a few forums that tried it before FB/Twitter came around.


Give the user exclusive control over what content they can see. The platform should enforce legal actions against users only, as far as bans are concerned.

Everything else, like being allowed to spam or post too quickly, is a bug, and bugs should be addressed in the open.


I'm not a bad actor, I only have 3 tweets and they're all reasonable IMO. So your premise is wrong.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: