This is from January 30th. Since then Bluesky has opened up to everyone (no invitation required), allows you to host your own PDS and even added new community based moderation tools known as Ozone internally, but acts a a "labeling service". It allows those who subscribe to them to get more specific tailored content they don't (or in some cases do!) want to see and it is labeled with a tag on accounts or posts, then the end user can choose whether they want those posts to have a warning on them, be completely hidden or nothing at all. It allows much more customized options & Bluesky doesn't need to actively close or ban accounts that might just include content other people don't want to see.
From the stats of people who ran their own PDS for their own accounts, it used an less than 1mb of data transfer a day and insignificant amounts of CPU/RAM (less than a 20% spike). But, it will depend on how big your account is, how many followers you have, number of posts you make, etc etc.
In native Bluesky with a default account, they have decentralized the servers so there are maybe around 20-30 servers all on the East Coast of the US. There's currently not much incentive to run your own. The PDS software is fairly new and unless you just wanted to have a non-US host for privacy reasons, you could do so and get away with running it on a RPi or standard VPS no problem. No one has yet started any major instances to rival the primary ones yet, however.
> it sends a copy per follower, meaning if 10 users on one server follow the same user on a remote server that remote server sends 10 copies of the message
That's not right. Most AP servers operate a shared inbox. So you only need to send one message - no matter how many followers you have on that server.
With ActivityPub you are still putting everything in one basket. All the federation is completely optional and identities are not transferable. But unlike big corps, that generally care very little about what you do, a whole lot of ActivityPub instances are heavily politically motivated and love to block or defederate.
That's not entirely correct. Mastodon already implements a feature to migrate to a new instance. Also there is a draft for identities that are not bound to an instance.
You can't even post titties on Facebook/Insta. Big corps are intensely politically motivated, they succumb to American prudery through advertiser pressure.
Even the more centrist tend to block instances. Especially because the others are politically motivated. Who in the center of the political spectrum would want to federate with gab for example? The same goes for the other extreme, like some instances which are full of tankies. When you federate with them you can't have a single news topic without getting "NATO evil, Putin great" pile-ons. It derails every discussion.
One of the nicer centrist instances is Beehaw and they defederated a lot of stuff just to keep the atmosphere alright.
I used to think it was a bad thing when instances defederated each other but since I saw the above happening I agree with it.
This post was written in late January. Bluesky opened up early access to federation about a month later, which does change some of the perspective. https://docs.bsky.app/blog/self-host-federation
"This is just for a brief period during polishing/debugging since this is the first time third-party PDS hosts have been added to the network.
The goal is just to help PDS operators join the network and make sure that if there's a problem we have a channel of communication open."
ActivityPub is ahead, but it feels like a dead end. It has the same issue with Reddit's little fiefdoms: moderators still lord over everyone and control speech, the algorithm, your feed, and your freedom.
Nostr and Bluesky's AT protocol are the most promising. I love the truly distributed nature of Nostr, but the ecosystem is hard to get into. Bluesky has strong technical underpinnings and accessibility, but they're the only ones developing and implementing AT protocol.
Social media should be more P2P and learn from the 2000's era before the platform giants stole away the dream. Bittorrent, RSS, Atom, semantic web (FOAF, microformats) were the way to complete digital freedom.
I would rephrase it: mastodon is full of cliques. The cliques form around moderation, politics, whether they federate with Threads.net, what content warnings are used, and even whether alt text is “required” for images.
Then some instances defederate otherwise innocent instances that don’t have sufficient agreement on all those points, regardless of actual content.
It is kind of tiring and honestly hard to recommend.
>I would rephrase it: mastodon is full of cliques.
It's social networking. Of course there are going to be cliques. That's what happens when you put a non-trivial number of humans in one space (be it virtual or physical).
I think it's more of an arcitectural issue. If you were a node operator, would you peer with a dedicated racial hatred AP node? What if a small minority of your users complained that you blocked it.
With something like nostr, you are free as a relay operator to block whatever makes you feel uncomfortable, while users can just get that content from some other relay without having to create a new account and move their entire social graph over.
Why should minds be put in jails? Why should we segregate into like-minded and undesirables?
We're better than this.
I grew up with [x views, y political party], but because the early internet was open and accepting and not at all like this, I was able to see the broader world and wholly new perspectives. Now the bias is to shit on, shame, and block everyone. That's not going to build a better society.
The internet pioneers would be aghast at all of the "happy little" censorship happening today.
I've been banned from several subreddits for leaving comments in subreddits the moderators dislike. Removed completely by automation. They don't even realize my comments weren't in support of their favorite issue to hate on.
Censorship is wrong. Put blinders on your own feed all day long if you so choose, but stop trying to put cages on other people's minds and voices.
But they do not have a right to unfettered access to me.
I'm free to choose to set standards, and the standard of requiring them to have a second account if they both want access to me and to content I find deplorable is a really low barrier.
And others are free to choose to commune with me in the spaces I want to spend time in, and communities are free to set their standards.
I've not argued for censorship. I've argued for my right to shield myself very mildly from people I want nothing to do with, and for others to choose to do so with me.
You're free to say what you want on your instance or those of people who believe as you do or are fine with your speech.
Nobody is stopping your speech, just your ability to impose it on those of us who does not want it near us.
Because... you are terrified that you might manage to figure out how to access such content, and need to make it extremely difficult for you to do so, even on purpose? (I clearly don't understand.)
No, because I want to be able to choose to interact with people who are fine with there being barriers to similar types of content.
If some feel they have a compelling reason to straddle both, they still can, so it denies them nothing, but sometimes even small extra barriers helps set a tone for communities.
> I can't understand the toxicity of mind that makes people think it's OK to impose others view their own, unwanted points of view.
You're using words like toxic to describe me and you don't even know me. This goes to show the incredibly rotten and highly polarized the status quo has become. Systems that behave the way you're describing are the current norm, and they're broken.
You can choose for yourself to ignore me. You can advertise to the world that you ignore me and people can opt into that if they choose. But you shouldn't ever have the right to unilaterally make that call for everyone. Nobody should have that power.
Imagine if [opposite viewpoint] could technologically shut you down for everyone else? The systems we design today could enable that tomorrow. Just because you're in a comfortable spot today doesn't mean that eroding rights and freedoms won't ensnare you tomorrow or that the zeitgeist won't change.
> If you're not welcome in a community
This is also frankly a disturbing trend. Creating little fiefdoms of unacceptance. Who makes you the judge of that? You don't even know what my views and values are, and you're seeking to shun me already.
We're all the same. We just sample the world differently. We all need to be more accepting and loving and stop playing petty team sports.
In the meantime, you're free to keep on tooting. Nobody is taking Mastodon away from you. It's just not a system I want to spend time or energy advocating for.
> But you shouldn't ever have the right to unilaterally make that call for everyone.
In this hypothetical I'm only making the call for the people of my instance, not for everyone. Like I said, you'd be free to join other instances where people are willing to entertain your point of view.
The audacity of assuming that the server that I pay and care for should be a vehicle for content that I don't agree with is baffling to me. It's my house, my rules, simple as.
I just want spaces for me and others where this isn't the default imposed upon everyone. I don't want to subscribe to [x filtering] unless I choose.
House rules and filtering should be opt-in and not the default for the public commons.
The legislators are going to make rules that enforce these technological straddles, and the defaults will slowly change in directions you neither anticipate nor enjoy.
We shouldn't accept ActivityPub as the end state solution to these problems. If anything, large instances (and thus their owners) will win out, and this just entrenches the status quo of tiny cabals in power censoring things that displease them.
My Mastodon instance is not a public commons. If I choose to let you join it, it is on my terms. If you don't like that, there are others, or you can run your own. The "public commons" is the system of federation, not the instances.
> House rules and filtering should be opt-in and not the default for the public commons.
I can't understand how you can think that fediverse instances are "public commons".
Each of them is run under the rules of the people that keep them running and not everyone can be moot to be able to suffer everything on their servers.
The public commons part of the fediverse is formed of the many software that you can just pick up and run yourself.
They mean the freedom to impose their views on others, because if the ignorant can just be made to hear the 'truth' they will believe it too. Having the ability to remove such toxicity from the much larger community of people who genuinely want share ideas is a strength of the platform, not a weakness.
The larger instances, the larger moderation issues.
People are only people so fierce moderation will happen.
With smaller instances "blast radius" is smaller and it shouldn't be problematic like in centralized social media.
> It has the same issue with Reddit's little fiefdoms: moderators still lord over everyone and control speech, the algorithm, your feed, and your freedom.
That's a feature, not a bug. Whatever your preferences, there's an ActivityPub instance with moderation you'll like. Want to be in a liberal thought bubble? It's there. Want nothing but all Trump, all the time? They have that. Only the good things about astrology? Yep. Aggressive blocking of anti-science woo? Sure.
Moderation decisions are made at the instance level, not the network level. There are instances that take a laissez-faire approach. Others are quite heavy handed. Most are somewhere in the middle. Users can pick which sounds best to them and find servers that align with their wishes.
What you are describing -- a system where I have to maintain accounts on a number of separate networks that don't interoperate due to content control disagreements -- is an ecosystem of competing centralized systems, not a decentralized one, and that people who run their own servers are reportedly presumed evil and defederated either by default or at the drop of a hat is a visceral demonstration that this setup undermines most of the benefit you think you are getting by calling it "decentralized". The reality is that many of the people using Mastadon would be not just "happy enough" but "actively happier" (per all the clearly ridiculous controversy surrounding people doing the supposedly-evil work of building search engines or, heaven forbid, being able to quote-tweet without permission, lol... if you are going out of your way to prevent hyperlinks, something seriously wrong has happened) using a centralized competitor to Twitter that was merely run by someone they liked.
The number of people who have to maintain more than one account is minuscule, and most of the time they ought to reflect over why they are part of communities with such mutual animosity that they can't find one home that accepts both/all.
I run my own, and have seen no trace of this being "presumed evil" you speak of.
Issue is that the admin is the one who decides what you can see. I remember a few months ago some meme/map or the network divided in "normie" and "weird" sides of the network. The impression I got is that the normie side instance admins apparently cooperate on blocking instances.
What if I want to be on both? Well you need two accounts.
Two accounts are trivial to manage, especially with a third party client. I managed more than that through a bookmarks folder. Many people have multiple accounts on various platforms anyway.
Also you could always host your own instance and be your own admin. I started doing that a couple of weeks ago. It's easy to find a host plan, if you don't want to do all of the work yourself. But you could still do that if you wanted to.
That schism is inevitable. To pick an extreme case, imagine a server for LGBTQ people and another for self-declared neo-Nazis. The people on the 1st might have very strong feelings that they want no contact whatsoever with people on the 2nd. There’s no world where it’s reasonable to say they have to keep their servers connected.
Now suppose you want to follow people on the 2nd server for journalism purposes. It’s unreasonable to have an account on the 1st server and then complain that they’re not talking to the 2nd. Even if you have legitimate purposes that aren’t creepy, they’re not going to go along with it. So instead, you get another account to follow the 2nd server and everyone’s happy.
The Mastodon server I own turned 7 years old today. We’ve blocked probably a couple dozen servers over the years. In every case I can tell you exactly why we blocked that instance. Most of them involve, literally, their users sending mine swastika imagery, death threats, or other utterly indefensible content, and the moderators of those instances being OK with it. My users stick around because they trust my explanations for the moderation actions I take. They don’t want to manually filter actual Nazi content from their timelines. Those that do can pick any number of servers that wouldn’t block that content.
I think that says more about the content of those instances being blocked. If you're transmitting content bad enough to get blocked by a large portion of the fediverse, the problem has more to do with what you're saying than their decision not to hear it.
I think that’s more like Signal, right? Maybe you could tunnel AP or BlueSky over it? I don’t really know what I’m talking about, but that’s how it seems to me.
The next hurdle to create distributed social media seems to be distributed likes and follows. Those are the social currencies of today's generation.
In a centralized system, you trust the central authority to show the correct number of likes a post got and the correct info who liked a post. Same for number of followers and who follows whom.
Do any of these 3 protocols have an approach to do the same in a decentralized world?
It is not obvious to me at all that this problem needs to be solved or that it is even a good thing for social networks to provide such stats; but, if it actually somehow is important, it is also a problem that is trivially reduced to the same problem as spam: likes and follows are simply signed messages, not pre-aggregated counts, and if you balk at sending someone a list of a hundred thousand signatures they are trivially compressed using zero-knowledge proofs. Now... how to prevent people from making up a million real-looking bot accounts? That is harder (though I absolutely believe is tractable), but you have to solve that anyway lest rather than inflating likes and follows people just flood the system with replies (and it is also a problem the centralized system will also have to contend with, though it might have more strategies available to it).
This is just such a strange thing to point out when the more relevant thing is that the comment itself does NOT display the number of likes even though that USED TO BE how the site worked because the people who maintain it decided doing so was actually a bad idea and actively removed the display of display of like counts one day, and while some people momentarily whined pretty much no one does anymore as it absolutely did turn out to be better to not have that mental clutter in the user interface as, in practice, it was irrelevant information for how people use the website.
Karma and voting are ruining Hacker News as well, just not as quickly because the algorithms here are not as aggressively tuned to maximize engagement.
Regarding whether it is "a good thing for social networks to provide such stats":
It's a central part of social media, even here on HN.
Look at the front page. Each post shows 6 pieces of information:
Title, Domain, Points, Author, Time, Comments
2 of them = 33% are numeric indicators we trust because we trust HN.
Likes and follows are probably two of the main reasons that centralized social media sites have displaced websites, rss, forums, mailing lists, newsgroups etc.
And yet, Hacker News long ago removed the ability to see vote point counts on comments, and there is no mechanism for people to follow other users at all.
And, hell, while "author" is the most trivial feature of all to get right as it simply can't be forged in a signature-based system, it is incredible to me how many people on Hacker News actively try to ignore who wrote a comment (not me! the first thing I look at is your name as I think it is actively asocial to dehumanize the people you are communicating with).
I honestly think you will be surprised at how usable the Hacker News feed would still be if you remove everything except the URL itself, and had your client generate a URL preview, and encourage you to try hiding some of these fields for a while using a local browser extension / script / stylesheet.
The real magic of Hacker News -- both of the feed and of the comment thread -- is the algorithmic sorting that makes sense of the cacophony of submissions, not the silly indicators (which are likely serving more as distractions than as beacons) it is bothering to surface to you: those are merely serving the purpose of making its decisions seem credible (and if you go hunting you will realize there are many posts you likely didn't see for many reasons despite having gotten many likes or even comments).
The question then is really: "how can I algorithmically rank content on a decentralized network", and I think that phrasing of the issue unlocks tons of doors that myopically insisting on a scant handful of specific statistics that are merely one of many many inputs to the current system that actually rules your usage of the website uses precludes.
Sometimes, information the user might actively be using to argue with the recommendations is even best hidden from the user: TikTok notably will happily recommend a post from years ago and hides the timestamp from you as people doing manual mental filtering incorrectly deweight the value of old content and/or find it awkward to interact with.
Now, is sorting and filtering in a decentralized system for sure solvable? I don't know... I think it is, and have tons of ideas for how to do it! Yet, I would not be shocked to try really hard and fail, or even to discover some "trivial" (in retrospect) proof that it is impossible. But, either way, one thing I am very confident of is that likes and comments isn't it, as real world systems -- including Facebook and TikTok -- manage to surface tons of interesting content to me that have low numbers of likes: the world simply isn't best sorted by thumbs up.
>The question then is really: "how can I algorithmically rank content on a decentralized network"
An interesting thing would be if you could write your own ranking algorithm and then apply it to your followers' content or if you could at least tweak the algorithm e.g. "I want 70% of content to be from let's say cybersecurity and 30% from gaming) or (I want 70% of content to be from my friends and family and 30% of content from "Internet people" that I follow).
Edit: Custom ranking algorithm would be hard to design and implement but not unfeasible.
My friend runs an activity pub instance, but the software he uses is called snac2. One of it's stated features is that is does not report follower counts. It also only reports likes or boosts on the home instance (not on federated feeds). The atmosphere on this instance is markedly different than elsewhere in the fediverse.
These all feel quite over-engineered. I feel like a combination of a webpage with your bio and a feed of posts with an associated RSS feed could cover the "follow a user/microblogging" side of social media, and email could cover direct messaging and threaded conversations. What am I missing that these protocols bring to the table?
Semantics. Notions of tags, replies, metadata, what kind of entity a link is refering to. Stable identifiers (how do you canonically refer to content rehosted elsewhere?). Authenticity for rehosted content. There was/is an idea and attempt at "semantic web" aka "web 3.0" but it never caught on.
Also, a common requirement is authenticated reads and fine-grained access control of pull-based content. E.g. I want to share my vacation photos with a different subset of friends/followers vs my tech blogging, only some of which should be public.
I’m with you. The more I follow all these various projects the more I’m convinced we already have the best form of social media with personal websites + rss + email.
If I understand this right: A supposed benefit of Nostr and Bluesky is the idea of having a strong cross-server identity.
Nope. This is most likely a bad idea. And we've already essentially proven this because the overwhelmingly most important method of internet communication DOESN'T centralize like this and has no need to. Email works fine, probably better, WITHOUT this strong centralization.
So while I agree with all the current short term clique-related critcisms of ActivityPub/Mastodon, it's still the smartest model.
From the stats of people who ran their own PDS for their own accounts, it used an less than 1mb of data transfer a day and insignificant amounts of CPU/RAM (less than a 20% spike). But, it will depend on how big your account is, how many followers you have, number of posts you make, etc etc.
In native Bluesky with a default account, they have decentralized the servers so there are maybe around 20-30 servers all on the East Coast of the US. There's currently not much incentive to run your own. The PDS software is fairly new and unless you just wanted to have a non-US host for privacy reasons, you could do so and get away with running it on a RPi or standard VPS no problem. No one has yet started any major instances to rival the primary ones yet, however.