Many people here believe that social media can't be both large and have good discussion because the human brain isn't made to interact with large numbers of people. What do you think of this?
p.s the difference between this post and this post is that I want to ask questions and get people's opinions and answers in this one more.
Here's a few examples, last one being an argument between a few people where most people, including Deimos agreed with this idea.
Personally, I find this idea almost terrifying because it implies social media in it's current form cannot be fixed by changing or expanding human or automoderation, nor fact checking, because moderation can't reasonably occur at scale at all.
However, I have 2 questions:
1: If large social media platforms can't really be moderated what should we do to them? The implied solution is balkanizing social media until the 'platforms' are extended social circles which can be moderated and have good discussion (or more practically, integrate them to a federated service like mastodon which is made to be split like this or something like discord.) An alternative I've heard is to redo the early 2000s and have fanforums for everything to avoid context collapse and have something gluing the site's users together (something I am far more supportive of) or a reason for invite systems and stricter control of who enters your site but doesn't explain the idea that once your site hits a certain usercount, it will inevitably worsen and that is something that stems from human nature (Dunbar's number aka the max amount of friends you could theoretically have) and so is inevitable, almost natural.
2: Why is moderation impossible to do well at large scales? While I think moderation, which I think is analogous to law enforcement or legal systems (though the many reddit mods here can definitely give their opinions on that) definitely likely isn't the kind of thing that can be done at a profit, I'm not entirely sure why would it be wholly impossible. A reason I've heard is that moderators need to understand the communities they're moderating, but I'm not sure why wouldn't that be a requirement, or why would adding more mods make that worse (mods disagreeing with eachother while moderating seems quite likely but unrelated to this.)
I actually totally disagree with most of the people you linked to about this particular issue. I think large platforms absolutely can be properly/effectively moderated, and have quality discussions on them, given the right site policy structure and enforcement, moderator numbers (achieved through crowdsourcing), good moderation tools, support for those moderators by site admins, and by fostering the right culture. Reddit has its issues, but overall still functions reasonably well despite its size, and despite lacking most of those; So imagine how much better it would be if it did have them. That's basically the root idea for Tildes in a nutshell.
All the above beliefs are precisely the reason I joined Tildes in the first place, and why I accepted the offer to help with theorycrafting possibilities for the trust system, and various other site mechanisms, all those years ago. And TBH I find it incredibly disheartening to see so many people here basically just throwing their hands up in despair, and claiming effective moderation at scale, and quality discussions at scale, are impossible. IMO they're not, and ironically enough, reddit is the proof. Its model just needs to be slightly improved for all that to come to full fruition. E.g. By forming as a non-profit to start, not taking on VC money, and disallowing advertising (and memes), it makes many of the pitfalls reddit has encountered over the years much easier to avoid.
You make a lot of good points here.
Why do social media companies think the right business model is to spend a tiny, tiny amount of their overhead on moderating and curating their actual content, and almost all of it on other things?
We have to question that business model. Throwing one's hands in the air and saying it's impossible is just avoiding responsibility or a lack of willingness to look at solutions.
It's tantamount to saying: "Well our diesel cars can't be made environmentally friendly. There's no way to make sustainable cars!"
Good moderation costs money. Or it requires an army of volunteers. For-profit firms don’t have the incentive to spend that kind of money.
I think there's a lot of value in a Reddit-like approach towards moderation. Small communities can be curated and generally moderated by community moderators, and up-level moderators can ban abusive/misbehaving communities and users. I certainly agree perverse incentives have hurt the community aspect of Reddit.
Removing the karma system and allowing trusted community moderators to ban users site-wide might go a long way as well.
Instead of phrasing "moderation is impossible to do at scale" I would put it like this:
The problem is in the incentive structure.
Just think: What does the action of voting/retweeting/liking/commenting do on Reddit/Twitter/Facebook/Tildes? Who is this action directed at? Whose experience is changed as a result? In all cases these actions are directed at the reader. The lurker.
Who is incentivized to vote/retweet/like/comment then? Those who want to influence the experience of the reader.
It is not a problem at a smaller scale (Tildes) when your user base consists of people with good intentions. They vote/comment to help other users.
As the system grows, this mechanic attracts bad actors who will game your system to death. And the interesting thing here is that moderation is not going to solve this completely. The goal of a bad actor is often to silence a productive discussion. And to do that they don't need to outright violate any content policy. They can simply derail the discussion, sow "fear uncertainty and doubt". In other words, create enough noise that would bury the signal. Conspiracy theories are an effective tool in that regard.
How do we solve that?
My attempt to solve this is through a different incentive structure. Instead of the vote being directed at other users, it is directed at yourself. That is, when you "like" something you say that you want more from other users who liked it. See more details in this post: https://tildes.net/~tech/u7f/linklonk_a_link_aggregator_with_a_trust_system
I don't think it's the amount of people that is the problem, at least from a behavioral perspective. It's the nature of the individuals that comprise the group. Most people don't give a crap about discussion online, or else Facebook wouldn't have a tendency to turn into a festering septic tank when discussion happens. Every interaction happens on a small scale within a larger system. You're talking to one person at a time while millions simply can see your conversation.
This site is the way it is for some general reasons, but mostly to do with people actively trying to find something different. You'll see similar in the fediverse, in general. Here, discussion is the goal, not the thing that happens by accident when encountering memes, pictures, and factoids.
The issue is you're trying to moderate something that behaves like a public square with millions of people. The people are doing what people will do, we're just better at managing it in meatspace with things like decency laws, police officers, and the ability to punch the asshole next to you.
I don't have a strong opinion about whether it's possible or not because I don't generally participate in those sorts of forums. What I do see, though, is that the defaults of everything public, automation driven by engagement numbers, farcically bad moderation, and showing a score of how much others like what you said, is causing many of the problems we see with social media. Unfortunately, those are also the things that drive the revenue of the companies building these platforms. If you take those things away, the spaces will be less toxic and also less profitable. I won't say you can't have social media without them, just that it will be an uphill battle to have sustainable social web sites because you don't have the outrage driving revenues like they do today.
That's exactly it. Today's social media are optimized for growth and engagement, regardless of content or anything else. This leads to sites designed to be addicting through any available means, user wellbeing be damned; they're like the digital equivalent of smoking or refined sugar.
It depends what you mean by "large." I don't think we're limited to Tildes' current scale. We could probably reach Hacker News scale using the current structure.
Beyond Hacker News scale, I think moderation gets more difficult. It seems like a key to good moderation is that the moderators are part of the community. The moderators should be people you know. This is because they need to be part of the community to understand context, and you should be able to ask questions.
Big Tech outsourcing content moderation to armies of anonymous workers in another country doesn't seem like a good approach to moderation. It's what they have to do because they went so far beyond the human scale that allows people to know the moderators.
To be honest I don't really have a personal definition since I've never used any platforms that fall between Reddit's 52 million daily users and this site's 260. If I had to guess, "relatively large" would be 1 million daily, 5 million monthly.
Hacker News seems to have been getting about 5 million readers a month last year, according to a New Yorker article about the two moderators.
For me, if a website has a comment section and those comments can be read forever and are publically accessible barring moderation or user deletion is counts as social media to me, with a few splits like chan boards, private and semi-private chats (ie Whatsapp and discord respectively) and excluding sites where the comment section isn't often used or there are high requirements to comment.
The Wikipedia definition definitely would also include tildes.
260 is just the amount of people that answered that during last unofficial census. @Kuromantis is contrasting two totally different types of stats in that reddit vs tildes #s comparison. The last time @Deimos opened up about active unique users was quite a while ago, but it was far higher than that (almost double, IIRC).
edit: It was almost
a yearsix months ago, and was 538 unique user contributions (created a topic or comment) in 30 days previous. And so given the 90/9/1 rule, you were probably looking at ~5,000 daily users on Tildes at that time if you included lurkers (who would have also likely been included in the reddit 52M daily users stat).To be more accurate, that's the amount of people who answered that they use tildes at least daily according to the census/survey we had (as mentioned by cfabbro, which includes lurkers). I'm not sure if daily data has actually been collected and uploaded. Bauke collected data on a weekly basis and came up with 250 people who comment and under 200 people who post (note that these overlap a lot and this is from March last year.) Deimos does monthly (previous 30 days of) data and gave the figure of 538 people who have posted or commented here.
If the one percent rule holds, you can expect 90 lurkers for every user who makes a submission or comment over the same time period. Tildes is very young and still all but invisible out there so our numbers are probably on the low end. We have to have at least a 10 to 1 lurker to submitter/commenter ratio.
According to more of Deimos' stats, the ratio might be as low as 2:1, using this stat on user themes as a proxy for monthly users (because you can only set one account default), although all the disclaimers are in his comment and the data is a year old.
I'm not sure whether I buy it or not, but I think that it's worth bearing in mind that the biggest sites are the ones focused on their growth metrics above all else, with fairly negligible focus on what users do once they're on the site. I feel like there was this sense for the longest time among the mid-2000's social media disruptors, dedicated to learning nothing from earlier and smaller social networks, that this sort of thing would just work itself out, and we're only just now seeing half-hearted, slapdash attempts at cleaning things up a little after several fascist uprisings by movements fostered, connected, and promoted on their platforms. Hell, algorithmic content promotion can be even worse than doing nothing, because with no awareness of what it's doing, it drives people down political rabbit holes and connects them with other people experiencing the same; I don't think we'd see quite the same post-rational far right death spiral ruining our political system were it not for the way mainstream social media is run.
So to summarize my somewhat disconnected thoughts, I think it's a little premature to say that big social media is impossible to do well, when your sample size is today's dumpster fire sites that consider growth infinitely more important than what they grow in to. Optimizing for growth is often directly counter to the interests of existing users, and unless that contradiction is resolved, big social media will continue to be bad.
But I don't think it's an iron law that it will always be so, it's just that the first big social media sites, i.e. the ones we have right now, are the ones optimized to grow the fastest, and those sites will pretty much always be bad for that reason. In ten, twenty years, it's possible that these sites will be displaced by more user-driven ones that grow slowly, but sustainably in the background, while these VC lenders figure out that there's a limit to growth, you can't make up for a loss leader on volume, and they're never getting their money back.
This is a neat sort of system, but one of the things I like about the internet is having been able to connect with people who I've never met in real life, and likely never will. This sort of cuts that out, doesn't it? I suppose it would work somewhere like Facebook which is sort of geared towards connecting with people you already know, but then I already pretty much use it like this (connecting with people after I meet them irl).
I think the people who believe it can't be done lack imagination. It certainly can be attempted, and as far as I can tell, almost no one ever even takes it seriously, so giving up now is giving up before we've even started. Every major site's business model doomed them out of the gate before their first users signed up. It's hard to believe the doom and gloom when nobody takes the problem seriously in their implementations.
In the end it's simpler than everyone thinks. It's not about fact checking. It's not about the topics being discussed. It's about one thing only: the health of the conversation itself. All one needs to do here to make progress is make the conversation itself (what we think of as threads) more resistant to bad behaviors (screaming during a civil debate, that sort of thing). Think of the website as a town hall and the moderation as a prosthetic for civility at scale.
If you want to make that happen instantly, it's simple: Force everyone to use their real world names and pay a fee to use the site. It's when you go for free and anonymous that you get into trouble. The accountability alone will instantly make people think twice about what they say.
I feel like tying your social presence on the internet to your name, even if you don't have any socially questionable opinions, is a mistake. Leaving that stuff connected to your name is just leaving data on a silver platter for all the varied hostile entities that gather online.
Making websites pay to use on the other hand, that could definitely be used to make discourse better quality. Though that would still have the problem of moderation at scale.
Somethingawful does it that way - just the $10 signup fee. Then if you hassle them they ban you and you get to pay to play again.
Yep - the oldest still in existence. That's how you know it works. If you can dodge the winds of September for 35 years you are doing something right.
Calling it a social media site might be a misnomer, though. It's more like a digital town whose members just happen to be all over the globe. They were never after the 'big scale' you see with Twitter etc.
I wonder how this works out with people who are not of working age or haven't gotten a first job yet.
If I remember correctly, they require credit/debit info on file, so they aren't courting that sort of membership. Adults only, no children.
And a pretty large share of teenagers, unless they're allowed to use a parent or close relative's account or they're allowed to pay for an account for them (do you know if they let people do that?). Definitely not the biggest loss, but that's a question worth asking for me.
Teenagers don't generally have bank accounts, do they? I'd be surprised if The Well had many teenagers (well, at least under 18) and if they did it would be obvious who they were. The entire place is intended as an adult forum - but not for crap like porn, it's adult level discussion (just like here) that they are after. Their 'conversations' are a lot like our groups or reddit's subreddits. No small achievement since it is only predated by usenet and mailing lists.
Frankly, their biggest stumbling block is the $150/year membership. Most people would never pay that for a forum, and you can imagine how much of a roadblock that is for people in the developing world who are lucky to see that much money in a year.
You can check here and read a couple of the few world-readable conversations on the site. Nearly all of it is behind a paywall and private, members-only.
I just joined the Well again temporarily to participate in the annual “State of the World” discussion and look around a bit. It’s pretty quiet these days. Not only are there not many kids, it’s also turning into an old folk’s home. It’s a bit surprising how little has changed in the five years since I was there last.
Although neither the Well nor Tildes are attracting many new users, I think Tildes has a better chance, due to how it’s funded and having more modern software and policies. Keeping most conversations off the Internet seems like security theater when anyone with a credit card could copy everything pretty easily. Using a pseudonym if you need to and being careful what you say seems safer?
I agree completely and that's why I love Tildes. Our pseudo-anonymous posting (once that's a real thing) should give us the best of both worlds. Offer them some invite codes, I'd love an AMA with their moderators. ;) I'm sure they've had some interesting insights into moderation, which also becomes more complex if you have everyone's real world name.
I know the general attitude on The Well is that there's no way to maintain civility without the real world accountability. I'm not convinced of that yet. Just because it works doesn't make it the best or even the only way forward. The trust system should be just as effective.
Outside those whose who buy microtransactions (which are probably managed by something like Google pay rather than a bank account anyway), no.
Fair. I guess it's just that you said it was a good solution and I interpreted that as you saying this should probably be widespread. Either way, requiring pay is fine, but I think that figuring out what happens to teenagers under such a system is something that should be considered.
(P.S, I'm 15, that's why I care about this.)
I think Tildes already has the better solution, honestly. Charging high fees and getting rid of anonymity does work for the health of a forum, but I can name zero examples of large forums using these models.
A lot of people I knew while growing up had bank accounts to split pocket money between "saving up for something bigger" and "cash I can buy snacks with" and I'm aware thats probably due to the more affluent area I was in but it did mean there was the ability to verify over 18 places with it if I chose to. It also helped with learning to save money and budget.
This is one of my own personal theories. The human brain just can't handle the amount of responses and data points required to facilitate such a large amount of conversations and knowledge on so many lanes.
I feel like it's strictly a volume issue. There's just too many people and too much information. Who has time to read 25,000 comments, let alone make sense out of it?
After a thread dies down, it seems like something needs to come along, clean it up, fact check it, summarize it, and present the consensus. I feel like there's layers of work that need to be done to condense and streamline the content into something we actually can fit into our heads on a reasonable time schedule. We never do this - we just throw the thread away until the next time it appears in the forum's life cycle, often repeating itself many times. Threads are mayflies, they live for a day in most forums, and that's it.
There's more to this, focused on the long view, I'm sure of it. File it under problems to be worried about after the trust system is done. Tildes is already on good footing here with the wiki and threads able to bump. That's enough to be going on with.
My nuanced response is that it is not the amount of data per se, it is the inability to understand the different context that a person and their argument is coming from. For a multitude of reason, context is not readily available in any discussion online. It requires a lot of work to give all the context that brought about a view point.
So having a discussion online you never know how a person is coming to your argument. Its like an optical illusion where from one vantage point you see a square but from another you see a circle 1. Both can be "correct" if you come to it from a certain angle. And unless you are willing to spend a lot of time understanding different viewpoints and context, you will never understand that it is not a circle or a square but some other shape entirely