Ask HN: How do I build a new social network?
I am interested in building a new social network but I want it to be limited to North America only. I dont like what platforms like Reddit has become nowadays (too much moderation and there is too many power mods fighting for control) and rampant bot abuse to grow communities. As someone that has been on the internet for a while I also dont like how political platforms like reddit has become so I want to make sure all political content is restricted or super duper limited. The internet that I grew up with wasnt this political or bombareded with political news and such, I feel like this is really bad for peoples mental health and would like to help them. What are your guys thoughts, what would make a new social network be successful? I have been modding on Reddit for a while and one of the good things the platform has done is that it hasnt turned into Twitter or platforms like Bluesky which everyone is still uncertain about. But reddit also sucks, they had the whole API fiasco and people blacked out there subs. There are still some small subs to this day that are abandoned from that period of time.
You can't. Your best bet is to joint the fediverse so you can interoperate with existing networks, like Mastodon.
> too much moderation
> rampant bot abuse
You can't solve the bot problem without thorough, labor-intensive moderation. These two of your goals are at odds with each other.
I will also say that almost 100% of the value of reddit is the moderation. After Musk eviscerated Twitter's moderation, it became more of a cesspit and people have been leaving for better-moderated networks.
> I want to make sure all political content is restricted or super duper limited
How are you going to do this without moderation?
What's your definition is political? Is discussing climate change political? What about health care?
If filtering certain topics or moderating a social network were a simple problem, bigger companies (like Meta) would've solved it without armies of human moderators.
Which also begs the question: how would you keep CSAM off your network?
> You can't solve the bot problem without thorough, labor-intensive moderation.
I want to see a social network that solves the bot problem by requiring a fee to post.
Medium does the inverse of this, but sort of the same idea. Content creators are rewarded money from a pool of subscription revenue based on community involvement. I think rewarding good content is more powerful than some barrier to entry.
Albeit the effect may very depending on the amount of the fee, I find it hard to imagine fee would stop the bot problem. World of Warcraft requires fee for accessing the most recent expansion, as well as the subscription, but it is swarmed by the bots all the same. IMO as long as there are traffic aka money, there will be bots.
It's a question of price.
> I want to see a social network that solves the bot problem by requiring a fee to post.
I think this is doable without losing massive numbers of users. You could do it by requiring a small fee to post publicly, but you could post for free to your followers.
BlueSky kind of flirts with this model because you have to pay to get your own domain if you don't want to use theirs.
> I want to see a social network that solves the bot problem by requiring a fee to post.
Why would people pay to post? Why do people post in the first place?
To attach a fee as proof of their belief in the post content being valuable.
It exists: well.com
Requiring the fee to post will suppress poor people, not spammers. Bad actors will happily pay to spread/promote spam/misinformation.
Metafilter.com
for CSAM part i'm sure AI solution could work
It's incredibly expensive to do that with flagship models, and many countries have laws that require human intervention for CSAM reports anyway
Perhaps a naive take but money, resources and even legislative friction are non issues to try to solve such a disturbing humanity-level problem on a global scale. Which platform or entity wouldn't want to invest or otherwise back it up? Meta alone should shower such initiative with money and proccessing capacity. So then OP could just integrate simple API into their new social network.
... So wait, you're suggesting that they simply use a non-existent Facebook product which Facebook will, for reasons unclear, provide to their competitors for free?
Not necessarily for free, but since this problem has criminal aspect and lasting negative societal impact, i.e. hurt people hurt people, the solution should not be profit-first.
Meta does offer LLama, don't know their reasons, for free, and i only mention them as a largest platform. This not-yet-existent product should be whatever FAANG and governments effort which will save everybody a ton of money.
Edit: some other reasons i mention Meta
https://news.ycombinator.com/item?id=43293004
https://news.ycombinator.com/item?id=32521609
https://news.ycombinator.com/item?id=35726102
Have a goal in mind. A social network is a tool to an end. What is the end?
I have found several use cases all trying to go on at the same time.
1. People with important group communications needing a large audience; i.e. scientists sharing research, organizations needing to contact people, special interest groups like Ham Radio (cuz they can't broadcast on their gear), photographers, musicians and so on.
2. Hot news feeds. These were added not to share news, but to lock people in with echo chamber news.
3. People simply wanting to communicate privately with friends and family and so on.
4. Local information of interest and need. Some is quite urgent, as in evacuation alerts; Some is important, like road closures and events, and some is caustic bullshit as in community gossip and flame wars. And LOCAL advertising can actually be useful and not just the "sports talk radio" carpet-bombing of tax attorneys, divorce lawyers, dick stiffeners, and house flippers.
5. Politics. I find that this is best in affinity groups that are heavily moderated. Trolls and bots gonna troll.
6. Probably more.
The more you address, the more diluted the product.
And the money model?
I have the same urges.
You need to carefully consider what you truly want and the realities surrounding those desires. How large do you envision this social network becoming? Do you aspire to be the next Reddit, Hacker News, or Lobsters?
Large social networks need to grow to survive, and growth requires funding.
Are you pursuing a commercial venture or a non-profit one?
If it's commercial, you will likely need to seek external investors. They will expect a return on their investment and will exert influence -- often in the form of ad revenue or otherwise increasing costs to your end users. It almost always ends this way, unless you stay a small bootstrapper.
If it's non-profit, you will still need to secure funding through various fundraisers and donations.
It would help to brainstorm (to yourself, at least) what you want to accomplish in lucid detail. This post is not it.
> dont like how political platforms like reddit has become > I want it to be limited to North America only
I obviously don't know your reasoning but IMO you're already going political
I think what we need is something that wouldn’t replace online content but would add a new layer, an alternative for those who want more real-world interaction… and maybe even for those who don’t realize they’d enjoy it yet, a new way of ditributing content is what i expect.
> The internet that I grew up with wasnt this political or bombareded with political news and such, I feel like this is really bad for peoples mental health and would like to help them.
maybe we had different internet (I grew up in the USA), but I remember a lot of shock porn/sites. I remember 4chan posting truly horrific content. I remember "prank" videos. I remember revenge porn. I'm kinda glad the internet has cleaned up a lot.
Me and my friends still make goatse jokes today. But the shock stuff is nowhere near as bad as the politics today IMHO.
I grew up on IRC back when 9/11 happened. We'd talk about how radicalized the mujahideen were. And it was puberty, we went from Nickelodeon to C&C Generals. That was the politics, plenty of caricatures, but the caricatures were a safety barrier.
These days it's completely common for one party to be radicalized against another party. Words like genocide and dictator are thrown about regularly. People cheer when a CEO is murdered. This is akin to people cheering for the suicide bombers back then. I mean, the Internet I'm in is already cheering for Hamas, even on the videos of Hamas blowing up ambulances.
The old internet was a spectator violence internet, but the new internet is about violent participation. If it keeps up, we'll likely see more things like the Capitol Attack, all these things that make for great television.
How do you define success? This little website seems to be a great example of having clear guidelines and enforcing them.
The first thing is to check if a social network already exists that meets your needs such as a closed group using Telegram.
I picked the first search result...
https://www.jploft.com/blog/best-social-media-apps
I dont like Telegram, so far I have been testing Lemmy and we have like an instance or something working here: https://canzuck.zone/ to experiment with. Telegram kind of scares me I dont like what it has become. I use Discord nowadays, and if Discord made their chats or some of their communities visible online on Google, they would give other social networks a run for their money imho. For instance the 3d printing community on Discord is very nice when I need help with my 3d printing it would be great if they were on Google, but with all the LLM and scraping stuff going on I can see why they might not want to do that even if they had plans before.
I'm not a fan either but it serves as an example of my point that existing social networks might offer a solution for the poster.
All you have to do is simply solve the problem of moderation and bots
You can set up a Mastodon instance, choose your method to filter who can register and choose whether it will federate with other instances.
From your comment here, you clearly have a strong idea about what you want to see, so stop pretending to be offended by strong moderation. Just stick to making a forum with a niche that you like. That way you can own it and ban whoever is making trouble for you.
Imagine going to a dinner party where politics was not allowed to be discussed. After a few wines, people would break the rule. People in social settings enjoy talking politics, even if it means animated debate.
Politics is essentially about finding the best way to improve society by fixing issues or proposing new ideas. Isn't it? If so, perhaps it's specific conduct you want restricted, not politics. After all, bad conduct can happen when talking about any topic.
Am I the only one that gets an icky feeling when I see things like “I am so tired of all this political stuff, I want to filter it out”?
All the political news you don’t like is a symptom of the toxic society we’ve built. Maybe we should be making new social networks that bring people together fix the root problem, not ignore it because it’s harshing the vibe.
I'd help you build it if you hire me.
By definition, building a social network requires building something for other people.
There’s nothing wrong building something just for yourself.
But “you can’t build an airplane out of bricks.”
Good luck.
Bluesky has an algorithmic feed even though people think it doesn't.
I am annoyed by emotionally negative political content on Mastodon and had to make a number of rules (no "fascist", "republican", "Trump", ...) to make it tolerable.
I found Bluesky's algorithmic feed eliminated about 75% of emotionally negative content and thought it was a good feed without any rules (though I was being careful who I follow.) Then the inauguration happened and there was a huge exodus of people from X and the feed either got overwhelmed with them or they were deliberately giving activists and journalists a lot of visibility so they build up followers quickly.
I've been wanting to make a "negative people" filter to make it easier to choose people to follow, the idea is to make a ModernBERT + LSTM classifier which I think will outperform my miniLM + pooling + SVM classifier. I need something that can look at images and detect "text in images" such as screenshots and image memes to block those too.
So from my viewpoint people can post what they want, but I won't read it. Probably Threads is using some technology like this to suppress politics and negativity in general.
To tell if something is true you need to make a god. Politics and negativity aren't quite the same thing, but they are fellow travelers.
a lot of memes have texts in them, so you potentially remove ones where context matters a lot of people abuse these to redirect traffic from what I have seen.