That’s not even close to being the main problem in launching a Reddit competitor. Reddit moderators are unpaid. Content moderation not being automated is not a problem. People like to do it for free.

The problem is the user base. Biggest obstacle for any social media site is attaining critical mass. In order to do that, you need to be an order of magnitude better than your competitors on some metric which makes the users choose your platform over their entrenched one.

>> Content moderation not being automated is not a problem. People like to do it for free.

The old classic saying is "Anyone that wants to be a politician is not someone you want as a politician"

Well that is true for reddit mods as well, people that like to moderate subreddits are drawn to the action because they like to have power over the others, this is generally not someone you want in charge of anything

There are technical solutions to that problem though. I remember a Reddit replacement from a couple years ago called Aether. It made all moderation public, allowed clients to ignore a certain mod's actions (basically, personally removing them as a mod as far as their view is concerned), and allowed votes of no confidence for mods.

No idea if that model would actually work, as the app itself was one of those cool-on-paper-but-not-in-reality P2P types which made everything slow and likely made it rough attracting users.

I really liked Aether from a design and user interface standpoint, and even accepted it as a valid use of blockchain tech, but it was also the blockchain aspect that hampered it. Took too long to do things, and required a desktop client. Plus, of course, the network effect of no users. I hope those can be (or have been) solved some day.

Perhaps we need a sort of jury duty for content moderation where you are somehow held accountable for your actions and decisions.

I would be fine occasionally contributing some moderation to a community I am an active member of. I find I already have to do some of that on my workplace Slack when people get aggressive with the @here tag reaching the entire company when they really want one of 5 people.

I always thought Slashdot's meta-moderators was an interesting idea. Before you could ever moderate (vote up/down), you first had to get enough points via meta-moderating.

Meta-moderating consisted of being shown a single comment and how an unknown moderator scored it. If enough meta-moderators disagreed with how a moderator was scoring, then that moderator would eventually lose all their moderation points.

Those that meta-moderate would slowly gain points they could use to moderate (no one had unlimited moderation points)

Granted, slashdot was different in that it only applied to comments. Post on the front page were done by slashdot employees. Reddit would still need a different level of moderators to remove inappropriate posts all together.

Moderation is really not the problem if you just remove comments. I’ve thought for a while that crowd sourced link ranking with a good search mechanism would be more useful than google. For tech related topics I often find HN search to be more useful than google.

>Reddit moderators are unpaid.

Moderators won't exist without users.

Users won't join a shit filled platform.

The thing you're claiming as the biggest problem is dependent on the thing I'm claiming is the biggest problem.

> It's something that's not well automated, and thus requires an enormous amount of labor (ie capital) This is why those who have tried have turned into racist pornography filled shit holes. Want to start a billion dollar startup? Figure out how to automate this problem.

Can't modern AI help with that? I'd guess GPT is pretty good at determining sentiment like racism.

Try it with ChatGpt, ask it determine if a sentence you give it is racist or not.

Of course, AI can't be the only moderator, but it may help to filter out most of the obvious cases.

That's an obscenely expensive solution.

'AI' seems to be approaching a point where automated moderation (or at least AI assisted moderation with a human in the loop) is feasible.

Splitting might help:

* Open source reddit-like clone designed for a single sub.

* Hosting company specialising in hosting this clone for a small fee.

I'm working on the first. :D

> This is why those who have tried have turned into racist pornography filled shit holes.

While some reddit alternatives might qualify as "racist", they've done a very good job at not being porn filled shit holes, unlike reddit. In fact, most of the porn spam on these websites are images or videos hotlinked from... reddit.

How exactly is Reddit a "porn filled shit hole"? I don't see any porn in the vast majority of subreddits. Nobody forces you to visit the rest, and they are typically even marked adult-only.

Besides, are porn and racism really the only questionable things in social media? What about extreme violence, videos where real people get shot or blown up? Somehow messed up American society thinks that violence is quite okay thing for people to see, but nudity and porn are not.

> How exactly is Reddit a "porn filled shit hole"? I don't see any porn in the vast majority of subreddits. Nobody forces you to visit the rest, and they are typically even marked adult-only.

My bad, I should have said "reddit is a porn spammer filled shit hole" on top of really being a "porn filled shit hole" /s

The fact that you feign being oblivious to all the CSAM, depraved porn and revenge porn hosted on reddit really doesn't make any difference.

The point isn't that that content doesn't exist there.

The point is that as a Reddit user, you won't see that content unless you disable your NSFW filter and go searching for it.

Because the platform is moderated.

> The point is that as a Reddit user, you won't see that content unless you disable your NSFW filter and go searching for it.

The point is that people use reddit to hotlink depraved, illegal porn images and videos and spam these media on other forums. That depraved content is not moderated, it should not be possible to hotlink porn material from reddit. The platform is not moderated like you claim, a checkbox in the preferences is not moderation, if evading "moderation" consist simply in hotlinking porn hosted on reddit.

That's completely tangential to the stated problem though.

That sounds like your own personal beef, which as we can see from Reddit would not hamper user growth.

> That's completely tangential to the stated problem though.

No it isn't "tangential", it is what makes reddit a "porn filled shit hole", these are your own words.

> That sounds like your own personal beef, which as we can see from Reddit would not hamper user growth.

Sounds like you have your own personal beef against reddit alternatives when reddit itself is a "porn filled shithole" which you refuse to acknowledge for some reason.

> feign being oblivious to all the CSAM, depraved porn and revenge porn hosted on reddit

And that is the point where you out yourself as a troll.

That is not a technological problem, as what is one persons "hate speech" is another persons political speech, and no one can agree on where the line is

For me the line is no line... All Speech should be allowed, the problem with moderation is attempting to do it in the first place at the platform level

Give people their own personal controls to see or not see, to block, mute, or otherwise disassociate.

Anything beyond personal tools to control your own overton window will be a problem. Reddit is a prime example of this where by countless issues with power tripping mods who over censor communities exist.

Probably the line is decency, is what you are saying decent or nasty. Keep it simple, keep the nastiness to yourself (I’m not saying you, I’m saying everyone).

The world would be a kinder place.

Have you been to reddit? Or other places where political conversations are had?

Also let me ask, is declaring "The US should have strong border security and limited immigration" being "decent" because I have seen people respond to that exact statement with claims that is hate speech...

What about "There are only 2 genders"

Reddit is designed to propagate content to the front page that invokes indecency. Take, for instance, a video of someone jumping turnstiles at a subway station or a homeless person bathing themselves in free hand sanitizer solution. What kind of comments is a video like this going to solicit?

If you want to keep things decent and avoid nastiness, prohibit content that evokes a strong emotional response. This strong emotional response, unfortunately, is what keeps people coming back.

It can’t be subjective like that. Nobody will ever agree on the definition or trust whoever is making the judgment call.

I think there is a much bigger problem and capital cost. Servers. 7 billion API requests per month from a third party app that's just one of them. There would be several. Then the much larger traffic from official sources. If you can't afford the servers moderation isn't even worth thinking about.

Who is going to pay for all of those?

No startup goes from zero to 7 billion API requests immediately. That's a problem that can be scaled in to over time.

We're talking about a Reddit replacement. It's realistic to think a half decent one would get a 1 million or so users straight away within weeks. That's 7 billion requests.

BlueSky had to be invite only to stop getting too much traffic straight away. 1.5 million joined Mastodon when Twitter got bought. The numbers for a feasible replacement are still skyhigh at the start.

You're assuming integrations exist on day one.

Hell you're assuming that an API even exists for people to build integrations against on day one, rather than the company just providing it's own UIs/APPs and scaling into APIs when it's ready.

It doesn't matter if it's web requests or API requests. Traffic is still traffic and it still needs to be handled by web servers that cost money. And the idea that a start up can go from 0 to 1 million users in a few weeks without a lot of cash for servers isn't realistic. No real competitor for Reddit would slowly grow, they either take advantage of people being unhappy and do so in a quick manner and have an intital surge or they don't even have a chance.

BlueSky has millions of users, people still complain that it's basically dead. Even though lots of people are using it. It's just dead compared to Twitter. That makes adoption harder.

Why not apply some commonsense to onboarding? Accounts on HN start their lives shadowbanned, once you've gained some karma you can downvote comments and some point after that the ability to vouch for dead posts. I don't have any data on to what degree it works, but it would make the lives of those seeking instant "gratification" somewhat harder. Alternatively you could apply rate limits. Maybe new users could create a small number of posts on their first day and it gradually increases over time before the training wheels are kicked off entirely. Obviously neither option eliminates the need for moderators, but in theory you should be able to get by with less of them.

Accounts on HN do not start shadow banned, but it's very easy for established users to flag their posts which hides them from most. I assume the flag threshold goes up over time.