Meta is building a decentralized, text-based social network

The bastardization of the term "decentralized" in this post has energy levels over 9000 for sure. Should be flagged as click-bait.

I think you're wrong and not the typical user. Personally, I would never use any social network without strong moderation and banning of toxic users, and I'm that many if not most people share this sentiment. Places without strong moderation turn into hellholes and are generally less interesting.

Whether a network is decentralized or not is a completely different, purely technical question. I don't understand why two issues get mixed up so often. The design of a community should never be based on technical considerations.

That's roughly how Mastodon works. Each instance is responsible for moderation of their own users. If an instance does not meet arbitrary standards of moderation, other instances may choose not to federate with them, but nothing can stop willing instances from federating.

You see mods on reddit go on ego trip ban spree's all the time. It's classic 90's mod meltdown stuff. I don't think subscribing to a benevolent master is a good model. What happens if their gaze turns to you?

Moderation is really hard. And it's especially hard because a lot of people have very para-social relationships between themselves and rando's they talk to online.

Whilst I know that sometimes those rando's turn into real friends in the real world. For the most part people are strangers, and the better form of moderation is getting off social media for a while.

If meta wanted to do something amazing. It'd enforce one reality tunnel on the networks. Rather than have a base set of rules. Have a base set of news and facts that everyone shares in. Bubbles are generally bad. And everything goes to shit, even on HN, when we start getting into politics.

So everyone needs to explicitly block cheese pizza for themselves just so a few perverts can share that stuff in the name of "free speech"?

Just because over 10 million Americans believe that Tom Hanks violates, kills, and eats children - a widely held belief - does not mean that a private company can't tell those people to go suck a bug.

A private company does moderation and it's the end of the world. Ron DeSantis bans books and he is just "anti-woke". The actual government.

The huge problem is scaling, the bigger the social network gets the more moderators you would need to subscribe because the ones you already subscribed to are not cutting enough noise anymore, so it's almost inevitable that there would be a scarcity of moderators, at some point the only useful moderators would be paid ones or more likely bots, the bots with the most success at banning "toxic people" using AI or whatever would quickly win that race... and then you just would have created Facebook again.

Widely held beliefs

Can you elaborate on these "Widely held beliefs"? You appear to have done some research into the topic so I think it would be beneficial to your point if you gave examples of the beliefs.

Because of Bandwagon fallacy, because we know just because something is widely-held doesn't necessarily make it valid, so these popular beliefs, I'd like to see what they are.

That same property of inviolable non-excludable makes a social network inaccessible to those who are frequently harassed. The right to shun some participants is a necessary function of social spaces, to mitigate known and documented human tendencies over the past couple thousand years to drown out, abuse, and in general commit conversational warfare upon others.

“Just block them” doesn’t scale when each harasser is harassing thousands of people. “Just block scripts” is no more effective than blocking adblockers. “Opt-in moderators” is no more legally viable than “No moderators”.

If you want to moderate a space, then set up and run a space of your own. If you want to federate with others, you’ll comply with the moderation requirements of your federations. If you want your space to stay online, you’ll comply with the moderation requirements of your hosting and transit providers. If you want your federation to stay online, you’ll comply with the moderation requirements of your organization’s laws. This is true today for all social sites, whether solo or distributed.

Facebook thinks they can opt out of expensive and difficult moderation duties by transferring the legal requirements for moderation to their users. They may succeed briefly, but the EU will not look kindly upon their attempt to circumvent the law.

Suppose I create a bot and create millions of accounts. post in each article, now you're spending a huge, myriad of time just filtering/ignoring users.

You're probably interested in Aether then: https://getaether.net/

> But users should always have the final say on whose posts they view.

The difficulty there is the end users are not hosting or distributing content. Some server(s) need to host content. No one can be forced to host content. A host is always going to be the final arbiter of what content is ultimately available.

Even if someone wants to run a Mastodon (or whatever) instance filled with Nazi content no one else is obligated to federate with them.

How is it different from whatsapp?

Social networks are hotels, right.

- First they are new and cool. The young, hip and rich stay often.

- New becomes old. Shoddy carpet, old fixtures, bad music in the lobby but older people stay until they die.

- New hotels sprout up with cool carpet, music and fixtures. The cycle repeats.

The only way we can ditch them is to ban them and then they’ll go underground :(

What is dead may never die.

> Is the proposition of “being more connected” really worth the downsides?

And 'proposition' is the right word to use, since its really just the promise of social networks that people like. In practice it has not worked as a means of connecting people to each other. Certainly not better than the alternatives it replaced. We should call the experiment a failure.

I'm all for getting rid of social networks, especially general-purpose ones. I don't see it happening though. Most people see them as a fact of life, inevitable as death or taxes. If you're younger than 30, you probably don't have an intuitive feeling for what the world is like without them. I guess this is understandable, if sad.

> What if, and hear me out: we ditch the idea of a better replacement social network and just ditch the concept in general?

Good luck with that. Social networks is an idea that is as old as the Internet. MySpace and Facebook were merely a step in a progression that included IRC, Usenet, email. Arguably hypertext itself. If you're not connecting people, what is the point of the Internet?

You do realize that social networks are optional? No one is holding a gun to your head and forcing you to use them.

They bring way too much value for this. They’ve introduced a slew of “new” problems, and these problems are bad. The problems are still outweighed by the benefits. I enjoy the connection social media gives me to people outside of geographical area and outside the sphere of who I’m comfortable phone calling.

I hear arguments like this all the time, and it bothers me that such arguments miss that there is something that people like about social media - connecting with people. It clearly is worth it and that's been around since the beginning of the internet in some shape or form be it IRC, forums, usenet, whatever. There's innovations that constitute as modern "social media," like the feed or social search, but the core notion of using the internet to try to connect people is going to be a demand as long as there are people and there is an internet. The best we can possibly do is identify maladaptive innovations and try to provide some alternative, until one is dominant enough that "social media" is as dated as "forums."

The ads aren't the problem of social networks. They are the problem of the Web.

Usenets were the first social networks, but their monetization scheme didn't allow for mega international companies to exist: they were paid for to ISPs and ISPs only pay other ISPs for traffic, they don't pay for individual services.

Email used to be the same way... sigh.

Web became the platform that enabled for advertisement to be mixed into the product. I cannot tell if Hotmail or similar saw the promise of ads on the Web, or did they "accidentally" succeed -- whichever the case, it's the Web that took over other Internet applications because it made it possible to monetize it in a different and a more scalable way.

Ideally, we'd have an Internet application for social networks, instead of it being a Web application. So that we pay our ISP for usage. We just really don't have that kind of application (yet?) And, I'm not sure ISPs are thrilled about this (will the effort be worth the revenues?). Definitely, companies like Meta aren't interested in that, as that would undermine their core business.

But then, who knows... ad spam seems to be able to find its way no matter the technology. You can think about SMS as being similar to an Internet application, and you do get spam ads over SMS too.

Another problem of social network as Internet application would be the search. Search works better when it's governed by a single entity which cares about indexing the data in various ways. If it's distributed, that'd be very challenging to implement, especially because it wouldn't be clear who's responsible for storing the indices and who pays whom in the end.

Another challenge would be the connection between user's identity and their ISP... well, there are a lot of things that would need work, but I don't think it's not possible overall. In legislative terms, we'd need to recognize social networks as a service similar to postal service, and regulate it accordingly...

>What if, and hear me out: we ditch the idea of a better replacement social network and just ditch the concept in general?

Why do "we" need to ditch anything? People are building it because they want to, and will use it if they want to. Or not.

I think you could literally take Twitter's code base from 10 years ago and set up a server and people would flock to it. Building something new is really not what is required, the bigger question is what people, what community.

Welcome to hacker news, the place that is definitely not a social network.

This puts the cart before the horse. Otherwise, you seem to be assuming that all people on social networks were somehow tricked into being there.

They're there for a reason. It just so happens that this current implementation of entertainment is particularly popular. Is that because they lack other options or because this is actually prime entertainment? Or is it because it's essentially free and mildly fulfilling and socially useful for people?

Either way I doubt a "solution" is to essentially "make it illegal."

young people will always want a social network that connects them to their peers before they realize the damaging effects of it. Even if warned, guaranteed they won't care at first

I think the ideal social network would just extend the natural social networks of real life.

Moxie at Signal was right when he said the real “social network” is just your phone contacts.

I think facebook realizes that their uptake among younger audiences is way down and their userbase is aging. They've buttressed themselves so far via instagram, but I imagine they are maybe seeing a bit of the same trend. A generation seems to, more or less, engage with some social network and it becomes pretty sticky for them. So maybe this is just their 'next wave' play.

You mean, ditch the HN comments section?

[Successful] Companies build what consumers want.

You aren’t locked into social networks. Don’t use them.

Consumers vote with their dollars.

The constant noise of “Paywall” is why so much is ad based.

So, if you’ve got a better idea, build it and see if they come.

[Updated]

Clarified that I meant successful companies.

It’s decentralized in name only, it’s marketing talk. It’ll still be behind Meta’s walled garden

The same way gmail is but with ActivityPub instead of email. You sign in with your Google account, but you can send emails to an email server run by anyone.

For an example of decentralized servers + centralized authentication just look at video games. It's common to have a common identity, but be able to join a server hosted by anyone.

It's just borrowing language that is hot right now, just like Discord telling you to "start your own server". Its not your own server and you do not own or control it.

In the best case it will speak ActivityPub as to embrace, extend and extinguish. Because all this effort is, imho obviously, an EEE tactic.

It’s called embrace, extend, extinguish for a reason. Is ActivityPub next? Though given Meta‘s recent lack of output I seriously doubt anything tangible will emerge out of this.

You have to really be rewriting history to blame google for the downfall of usenet.

The Google graveyard is expansive but they most certainly didn't kill Usenet. Spam put Usenet on life support and ISPs dropping NNTP hosting pulled the plug.

By the late 90s spam killed actual conversations on Usenet. Eternal September had already raised the noise floor but spam just got out of control.

First ISPs dropped retention of even text-only hierarchies and then by the early 2000s dropped NNTP hosting entirely. So even people willing to wade through the spam had to go hunting for independent Usenet hosting.

Google didn't have anything to do with either of those things.

Usenet was destroyed by spam.

Even if you set up your client to have your groups clean, for any new arrival without guidance it was a horrible experience that required a lot of effort in bringing signal out of the noise.

What killed it was not having a distributed spam filtering as it became popular.

BBS networks (eg FidoNet, PC-Relay) were pretty great too.

FWIW, I believe, but cannot prove, that content moderation on Internet accessible "open" messaging systems (aka store & forward, decentralized) is not feasible.

Back in the day, I was hub for a small BBS network as well as moderated a forum on CompuServe. It was manageable, but a lot of work, esp for volunteers like me. These days, there's no way I could do a reasonable job.

I guess we'll see if I'm correct with the Mastodon experiment.

You're conflating the limitations of Mastodon with the limitations of ActivityPub.

ActivityPub does not require "email addresses" for identity - that's a Mastodon idiosyncrasy. Identities in ActivityPub are URLs, so as long as you own your domain you can preserve your identity. That's not too different than your identity being a public key, with the benefit of human readability.

Mastodon doesn't offer you export/import of your activities because - again due to its idiosyncrasies - it does not care about past activities. What a user "sees" when they visit another account is only what that account has posted and has already been disseminated to the current account's Mastodon server. A sane approach to past activities would be to load them in response to the user interaction, instead of serving stale data from the db. For this case exporting activities from one service to another would allow for continuity.

This could probably be achieved even at protocol level, ActivityPub in its simplest form has enough vocabulary to allow loading a collection and adding its elements to a another one.

Some of these things exist in one form or another in other clients than Mastodon.

I did a quick ctrl+f 'nostr' on the article and when no results came up I closed the browser tab.

They obviously haven't done their research.

> How can a decentralized network provide any performance guarantee?

Why can't it just be like message boards like Xenforo and PHPbb? There are numerous sports-related boards created 10+ years ago on that old tech that are still handling thousands of users concurrently.

They're also well moderated too; that's something that becomes impossible if you are trying to manage hundreds of thousands of users.

> A central server is better if they have a good moderation plan built in

If that's true, then the internet itself was (and continues to be) a mistake.

Weirdo billionaires can't buy a decentralized network.

> Why is this obsession with decentralized?

One reason could be that decentralization reduces the liability on the company. They could save huge costs in letting users self-moderate and in not being responsible for illegal content.

> a good moderation plan

^ This is the reason decentralized is better for everyone, including Meta. At a certain scale, content moderation goes from "hard" to "actually impossible."

With enough users, you will eventually have an Kurdish knitting group that meets in Turkey, who uses ethnic slurs ironically as a commentary on the government's treatment of immigrants, and you'll have to unravel the whole thing when someone complains. And the whole conversation is happening in Kurdish.

A central authority that has total control over the conversation is considered by many to be a bad thing. Some think that a more democratic approach is preferable.

"Why is this obsession with decentralized?"

From a business perspective: absolution from liability. If it's decentralized then you can conveniently say you can't control it and you can't shut it down. All you can do is make money from it.

Decentralized implies smaller scale than centralized. Smaller scale means less burden on moderation. Smaller scale means less resources required. Smaller scale means a better community.

> Why is this obsession with decentralized? I don’t see the benefits of it at all

To me this is the angle: we want to be available in China, this is the way to be available in China.

> A central server is better if they have a good moderation plan built in.

"An authoritarian regime is better if they have a good dictator and the proper watchdogs"

It depends how decentralized.

If they go so far as to be partition tolerant then you get a lot of resiliency wins. i.e. if a solar flare knocks out anything attached to a long enough cable to induce circuit-damaging current, you can still gossip messages between colocated devices, some of which might be en route to someplace distant. Secure Scuttlebutt Protocol works in this mode for example.

In such a scenario, being able to piggyback messages on travellers you don't know (even if it takes a few days to propagate) could be lifesaving.

There's also something to be said re: censorship. If you can facilitate comms between otherwise firewalled groups of people by simply sneaking a SD card across the border from time to time, that's right in line with Meta's mission statement.

I doubt that Meta is going to make it this decentralized. Whatever they build is going to be decentralized™. But it would be pretty cool to see this design space explored.

All of it has always served a commercial use. It's meant to be for you, whether you like it or not.

Or you could just go outside.

Remember when Facebook (and Google, etc.) supported federated XMPP? I can't bring myself to expect any different outcome this time.

If my memory serves me well, their Messenger was once XMPP-based (perhaps still is somewhere beneath) and you could use it without a fuss in any Jabber client. But then smartphones bloomed and the possibilities having a mobile app arrived.

So, I'd expect same scenario here: open at the beginning and walled later - once they capture enough people within. That is, if this thing ever manages to gain attention.

>What's in it for them?

Commoditize your complement.

That's really what Usenet is - I'm kind of surprised we don't see it used more.

As someone who uses it but isn’t a fan of it, it’s exactly the same as it’s always been.

I've been using Twitter for the last 10 years in write-only mode and have even given up that a couple of month ago.

Yep, Urbit is incredible. It does require some technical know-how, and has a sign-up fee (unless you can get a planet for free from someone). I think that has really maintained quality of discussion. It has also come a long way in the last 3 years. I was given a free planet in early 2020.

ICYMI, they've stepped this requirement back so now you just need a "Meta account".

VR goggles? Dude. How about the part where they tried to consign your grandchildren to a dystopian technocratic panopticon?

https://hackernoon.com/on-the-infestation-of-small-souled-bu...

My wild guess is that this is their way to operate in China. It'd be much harder for the GFW to block a decentralized app.

they might want some sort of franchise. push some costs and risks down the food chain for a share of the ad pie

It's remarkable to me that Meta is still pushing Reels so hard despite the fact that most of the popular ones are literally advertising their competitor. The TikTok logo splattered all over Facebook and Instagram is a branding disaster.

Decentralized doesn't capture what you want it to mean anyways. When crypto people say decentralized, they don't mean AWS edge locations or The Federal Reserve Central Bank which are both decentralized. How about we start saying exactly what we mean?

"Non-excludable" captures the important aspect. It doesn't matter if your network is decentralized if it can't maintain this feature. Bitcoin being non-excludable implies that it is secure in terms of consensus control - even if a powerful agent is out to exclude you from the network, they would fail. That is the core property and it leaves no room to hide behind as a buzzword; it means exactly what it means.

The real problem is decentralized is too ambigious. There are a billion different ways something can be decentralized. The term is meaningless.

Pier to pier is what we always called it, or shortly, P2P. It's how the internet was designed. Now, all piers are equal, but some piers are more equal than others.

On a side note, the Napster model was almost perfect, its downfall was centralized indexing. I wonder if a blockchain could be used for distributed indexing using the same P2P communication model of Napster.

Its not as bad as "democratizing" something

Permissionless

No, I think it's gonna be a shared textual virtual reality, metaverse if you will, in short, a MUD.