A Look at Bluesky
juliette.page
juliette.page
If ActivityPub wants to keep riding the wave it has now they should start investigating ways to support this. I think more and more people are getting fed up with centralized platforms / not owning their data.
Given the meteoric rise of TikTok, I don't think there is any sign of this being true outside a small techy bubble. At least, not as a percentage (that is, the growth in the number of people who are fed up with centralized platforms is dwarfed by the growth in the number of people who are fine with centralized platforms / not owning their data).
Network effects are real, but appear insignificant until they’re in our faces. The fact that it’s currently a minority is expected, that’s just how network effects work. It was just a minority on Reddit when Digg was the place to be as well.
Obviously I don’t have any magical predictive abilities. But the “vocal minority” argument has bitten a lot of people before. Including those that have stake in the game and are paranoid about competitors.
And they shouldn't. Part of being a kid is exploring interactions and learning from mistakes. Can you imagine growing up in a society where your account follows you from elementary school and anyone can review it? Even if you are the utmost in social propriety, but have a hobby that you enjoy, someone is going to disapprove of it or your fandom and penalize you accordingly. Post too much? Penalized. Post too little? Penalized.
The world is not a fair or equitable place, and so, we should not design systems that enable bullies.
For instance, a fire department may switch from Twitter because it’s rate limited. A community for visually impaired may move from Reddit because it’s dragging its feet with accessibility. Journalists and media houses may switch because they get censored by a billionaire who’s on a quid pro quo relationship with entities they investigate. Regular “content creators” may switch when they get copyright strikes or demonetized by an automated system. In all of these cases, the small minority of “providers” take a large amount of “consumers” with them. Cumulative resentment looks just like apathy, until suddenly it doesn’t.
Most people are followers. Other liberties like freedom of the press are inapplicable to most people, directly. But the indirect effects, usually shepherded by “vocal minorities” has been shown again and again to add up and transform society rapidly, in something like S-shaped curves. If you accept the rapid growth phase of network effects you need to also acknowledging the rapid death phase on the other side of things.
The younger Snapchat generations have normalised having their face plastered on every random "hello" message daily.
In Denmark we're in the process of banning devices like iPads and smartphones from schools and other institutions. We're also likely going to see an age limit on Social Media like TikTok that's around the same age as buying Alcohol (yes, that's a good combination) and a range of other things. In the wider EU, you have privacy and competition stepping in, both banning and taxing these platforms in ways that might make it hard to operate a profitable platform if you're allowed to have one.
It's still a little too uncertain to say anything concrete, but it's certainly not looking like the explosive growth of centralized "free-to-use" platforms will face the same profitable and non-regulated market the coming 20 years that they did the previous.
Someone is going to be well positioned to take advantage of this as people move on from the previous generation of social media platforms.
This is another advantage for Bluesky. Because as of now, a user of BlueSky would have no clue that it is "decentralized" at all. In comparison, Nostr shoves it in your face that it is totally unlike anything you're used to.
Bluesky just appears as a Twitter alternative. And I think that is very smart of them. Because realistically, only very serious people/companies should consider running alternative relay servers. Relays come and go on Nostr because people underestimate the amount of work / money involved in running them.
Another non-activitypub alternative is nostr, where your identity is a public/private key pair: https://github.com/aljazceru/awesome-nostr
Moving data only works in this case if you've thought to attach your DID to a domain you control, and the server you move to connects to the BGS server your data lives on. It's still very unclear how this will work. I suspect it will be like Usenet where most BGS indexers have expiry rules and won't hold on to every post.
This is more complicated than AP and Mastodon where you know exactly where you stand when your server goes down. So many people are going to be burned because they heard it's all portable, but didn't actually understand how it worked, and find they can't migrate their identity or their posts because it all rested under [handle].somehugecentralizedserver.tld.
People who struggle with instances on Mastodon (and eventually, Bluesky) are not going to have an easier time figuring out domains and DNS to make their identity portable. This remains a huge unaddressed issue and should concern people who think Bluesky is easier just because it's unfinished.
With Mastodon you can download your data any time, but you can’t actually upload it anywhere else. There is forwarding, but only for follower subscriptions, with the cooperation of both servers. It’s quite limited. There are brownouts from servers getting overloaded and also due to fairly frequent policy disputes. It does let you do more things without making them public, though, and that’s important to many people.
The blogging model where you have actually independent websites, links, and RSS feeds seems better from a decentralization standpoint, but it’s not popular due to the difficulty of getting people to subscribe to your blog.
That’s why we comment here, right? You could post a comment to your blog, but who would read it? Replies are important.
I do think software like gotosocial is the future of ActivityPub, it certainly isn't there yet, indeed.
Mastodon, I'm afraid, has tainted ActivityPub¹. Maybe even beyond fixing. Mastodon is architectured and developed for large servers. It's performance is terrible (It's Rails) for a one-man-show, but optimized and pretty good, for a server (or fleet) for (tens)of thousands of people. There's no incentive to change this. If a choice has to be made between "ease or operation for a single person instance" over a "ease of operation for a thousands-people instance" the latter will always win. Same for performance, security and features.
Gotosocial, OTOH, has the incentives reversed. I'm convinced it will only become easier to host for your personal instance. Only gain more features for such an instance and prioritize fixing issues so self-hosters benefit, rather than community-admins.
¹Edit: AP isn't designed around large federated servers. The protocol, albeit convoluted, is perfectly fine for a situation in which each user is her own server. It's mastodon that has turned AP into a de-facto "one-server-thousands-users" model. Without Mastodon, AP wouldn't exist, I'm not trying to make Mastodon look bad, I like the project.
[1] Technically three - I've got Akkoma, GotoSocial, and a Honk but I only really use the Akkoma one for "real life".
Not only do you have to keep it running at your expense but you need to invest significant time in moderating the content. Lest you allow illegal content e.g. CP to be hosted for which you will be legally responsible.
That way you don’t need to moderate.
There's also the fact that Bluesky is run by Jack, who founded and managed pre-Musk Twitter "1.0" and funded Musk's Twitter "2.0". I don't have high expectations nor hopes.
That's not accurate, Jack gave the Bluesky/AT Protocol organization initial grant money, but he doesn't "run" it. The CEO is Jay Graber.
Jack is a power-hungry opportunist who throws money at anything that can give him wealth and status in the future.
I don't care about Jack and neither should you. Bluesky is Bluesky run by the Bluesky team. The team and what they're building should be your only concern.
Twitter has a CEO now, but the majority shareholder is Musk. She cannot do anything without his permission. Were I CEO I would remove that childish poop emoji auto-responder but I don't think she has that kind of authority. Facebook has a board of directors, but Zuck is the majority shareholder, so they have to do whatever he wants, so they spent time and money on the metaverse when a normal corporation might have said let's not.
Looks like Dorsey had a seat on the board as of 2022. I don’t know if there’s a good way to find out if that changed? The closest thing there is to an official web page is the FAQ and blog articles that it links to.
The fact that they haven’t broadly got other servers they’re federating with - they’re not ready - is why they’re limiting signups.
> A problem with the system of having a domain as a username is that people can claim it who don’t have access to that username.
That was a bug, and is not generally speaking, the case. You must have the ability to alter the DNS records for the domain you wish to claim.
> However, that means interactions, such as user likes, blocks, follows, etc are all public. With no means to make them private. Which could lead to harassment!
It could also lead to the enlightenment of man, if we're just making random claims without explaining ourselves...
> But runs into the problem of algorithms being hijacked to show inappropriate/offensive/bigoted content.
Posts on the AT Proto can be labeled by whatever labeling service you prefer, including Bluesky itself. You can then use your preferred labeling to filter out labels you don't want, or subscribe to a feed that does that for you.
> But also it’s a group of overly complicated specifications that could’ve just been extensions to the ActivityPub protocol rather than a whole new protocol!
I don't think, "AT Proto is too complex, use ActivityPub instead" is a reasonable claim at all. The problem is a complex one, and AT Proto in particular is not even out yet.
Honestly, I can barely understand this "article". It's just a bunch of oversimplified (and sometimes outright wrong) statements about what AT Proto is, coupled with painfully misunderstood declarations that border on the illegible.
This was not a high quality piece of writing. It doesn’t appear as if you spent any time at all learning about the topic, and you did not express yourself well, even when trying to articulate your ignorant points.
If English is not your first language, I totally understand and applaud the effort to learn a second language, but if English is your first language, this is the work product of a child.
You need to put more effort into two main areas to improve your writing: research and structure. I cannot overstate how poorly constructed this post is, and how thoroughly it undermines your effort to make an argument.
You can learn to improve! It’s a skill, like anything else, and your natural curiosity is what drove you this far, but you will need to work at writing if you want to get better at it.
This is an incredibly rude and insulting way to give feedback. There is no reading of this that's kind or charitable, don't tell people their work is the "product of a child" and how you "can't overstate how poorly constructed [their] post is".
But, thankfully, you can learn to improve! Try to avoid insulting the author and their work. Instead, just try to offer actionable suggestions on how to improve.
This is not exactly true, it just means the server that holds ur data (traditionally instance or home server) doesn't need to be involved in the transfer process. However, this has centralized identity resolution, which is not a good long term solution.
> For example, someone was able to get the AWS S3 domain as their handle!
This is easily hotfixable by requiring more stringent proof of ownership of the domain (like DNS records), rather than a problem with the protocol itself. But it's a valid concern
> But runs into the problem of algorithms being hijacked to show inappropriate/offensive/bigoted content.
I thought the whole point was that if you didn't like the algorithm you could swap it out for something simpler. This point puzzles me greatly
> that could’ve just been extensions to the ActivityPub protocol rather than a whole new protocol!
Not really. If you peel back the interface they're very different, since the transferability of accounts across servers is something baked deeply into how ActivityPub works and cannot be easily changed without rewriting the protocol. That being said, I personally hope there is some network that can solve the problem of untying resources from domains, since it is one of the big problems with any federated protocol
A big improvement is possible "just" by making instances revalidate the URI against webfinger regularly (you'd still need to plan ahead for that, but it'd enable "stable" usernames) and secondly by allowing for a key pair to let users sign messages to make claims about ownership and which account is current. That would make it possible to enable fully unilateral moves.
Neither would require any changes to ActivityPub.
Like moving from GitHub to GitLab.
Each instance then has its own moderation policies, again, kind of like GitHub. But your identity is still your identity, and you can keep a copy of your data.
> and you can keep a copy of your data
Is this really a selling point / concern for anyone? I’ve never heard anyone express that the problem with tradition social media is that they can’t download and keep a backup of their data. Its about a central corporation being able to decide what is allowed to be said.
You can just take your data to another instance whenever you don’t agree with the policies of your current one. And all your connections/interactions/data should stay intact.
If it works as well as it seems to in the federation sandbox, you shouldn’t even be able to tell that you’re using a different service, the app just sends requests to a different server, and the web url may be different, and your default feeds are generated somewhere else.
Now, you may say that users won’t care about backing up their data, but that can be solved with some open (or paid) archival services.
But do you have more info on this wiping data and removing people claim? All I've seen is people pissy because someone wasn't banned for something rather minor.
>Given that this is a testing environment, we will be defederating from any instances that do not abide by these guidelines
>As part of the sandbox, we will be doing routine wipes of all network data.
It seems like it's centralised on a single "Big Graph Service", even if you run your own server.
Realistically, I don't think it could handle a full enslaught of new users and wants to onboard slowly which makes sense from a development point of view. The issue is, it's basically dead or just full of shitposts. Which has no real value.
And the number of people who want on it aren't actually that high, I've got 3 invites and no one to invite.
Hi :)
It has the potential to be, once it’s finished. But it’s not finished yet.
Is anybody already doing this? It shouldn't too be hard. Just throw all the notes into a directory and add some code which replies to the different types of requests.
I guess the best way to get ones tweets out of Twitter is the "Download an archive of your data" thingy in the settings?
https://wordpress.org/plugins/activitypub/
The Indieweb Crew of course did this first. The principle POSSE – Publish (on your) Own Site, Syndicate Elsewhere – was coined by them and is the guiding principle behind the Indieweb movement and it is exactly what you’re trying to do. You should click around a little bit in their wiki:
Most of the time the Indieweblings are doing their own “protocol“ by receiving and reacting to webmentions with rich data structures in Microformats syntax in their HTML. Some parts of Indieweb a rather opinionated about technologies. But a lot of them also interact with different federation services, for example see:
So they probably could get even more users, but I'm guessing it's already sufficiently crazy for the team right now
Besides, 'limited' account creation to form communities is a pretty well-established scaling strategy. They have scaled from 0 users to 200k users in 5 months - Facebook originally took a year to scale to 1m users so it's not unrealistic that Bsky are growing/scaling at approx the same pace at this stage in the curve. Worth noting that Facebook also just couldn't have handled scaling to 1 billion users on Y1, and that required years of engineering effort (and lots of money/resources).
Considering they can only have limited users at their current stage, invites to form close communities is a strategy to ensure that a social graph is created early. Having 20k users that all have a few friends with accounts is more important than having 20k users where nobody knows each other.
Remember Facebook also originally scaled campus by campus - which was a process that let it 'scale' in a way that encourages social graph creation (i.e. 10k signups in 1 university is probably WAY better for a social network than 10k signups in 100 universities).
Has anyone else created a decentralised social network at this scale? Arguably matrix but they also grew slower, potentially have a simpler product, and had scaling challenges along the way.
Remember scaling doesn’t just mean technical for a social network - it also means support and moderation.
With which backend? Scaling from 200k users to possibly hundreds of millions that quickly?
That isn't necessarily an issue with using a domain as a username, it was only possible because you could verify ownership by putting something in a subdirectory they defined which was available as an S3 bucket name (see https://news.ycombinator.com/item?id=35820815). If you allow verification only from the root directory or via DNS records this shouldn't be a problem AFAICT.
Au contraire, if network users control their clients, they can "block"¹ and "follow" content at-the client level without stating it publicly.
¹. Not a full block, as the blockee can still read and interact with the blocker's posts.
So do those extensions exist (even in the form of plans), especially the ones that free you up from a server or is it just wishful thinking?
I tried Mastodon for a bit and like the community there, but I found they were very deliberately trying to build a place that wasn’t like twitter: building federated search is frowned-upon; no quote-tweets; extremely low threshold for content warnings.
They’re building something different and that’s great, but it’s no surprise that people fleeing twitter want to go to the more twitter-like app.
There should be a decentralized social communication protocol that is more fit to reflect how people communicate: more ephemeral and less of a permanent record. I remember LitePub/Pleroma did some work in that direction, and it's mostly compatible with ActivityPub too I think
You have four properties of feeds: category, tree-comment, real-time and persistent.
Reddit was the winner on all those except real-time.
Decentralization is now the most important feature, and taking responsibility of your own data is the best version of that.
The only missing part is decentralized RSS servers and clients. Think backup for your posts and RSS reader services that also backup your favorite feeds.
All you need for that is a distributed db underlying it all like this: http://root.rupy.se and a reader web tool.
The only feature we lack is a category (think hashtag) for RSS.
So a combination of all services below:
Add mail and video hosting/streaming and you got the complete deal.
Reply works fine and comments too even if only local.
I got my invite from a website where people shared their invite links. After creating an account, and using it for a while I also got the ability to share invites with like 10 people or something.
I don’t even know if doing that kind of thing would work today though, given how many spammers and such trawl the internet. Probably some small number of people would grab every invite for themselves, instead of taking only one invite each.
You can include your email in your profile. I’ve done that and I get occasional emails from it.
I know of no way at this time to run one's own instance or federate.
There are of course significantly less users compared to the main instance, but it should give you some confidence that federation dev work is on the right track.
It has to be better somehow, and the whole selling point is decentralization to sidestep the possibility of recreating the current Twitter fiasco.
That goal is not yet achieved, which means it's below parity.
If you don't like Mastodon, there's Pleroma, Misskey, Calkey and others, as well as apps, or you can find RSS feeds from them if you just want to read, or you can use one of the alternate web frontends.
Originally people were extending feeds with the Atom ActivityStreams Extensions, in XML, for richer data structures which contain the social activities of persons. The protocol PubSubHubbub (now WebSub) was used for some realtime updating of feed subscribers (Push-then-pull, instead of Pull). Mapping persons or rather their accounts to resources like feeds was the impetus for WebFinger. And for notifying comments or reactions to upstream feed producers there was the Salmon protocol. Together those formats and protocols morphed into the OStatus protocol.
With the rise of JSON and with lessons learned from OStatus and pump.io those specs morphed into ActivityPub today. The underlying ActivityStreams vocabulary is still the same, I think, but the mechanism for following and responding now uses dedicated out- and inboxes.
I think microblogging and other social network stuff and long-form blog posts are different content, best consumed differently. There is some overlap, of course, but articles do not need shoehorned in. Sometimes people simply undervalue the simple link from a microblog post to the longform article.
Which is based on an open-standard, interoperates with Mastodon and is from a company that has the battle scars to prove that they can operate a global service in a mature and a reasonable way.
Given Jack Dorsey's support of the crazier positions of Elon Musk and RFK e.g. WiFi causes leaky brain it really does look like the best of a bad bunch.
[1] https://www.theverge.com/2023/7/1/23781271/instagram-threads...
> Which is based on an open-standard, interoperates with Mastodon
By which we mean: capitalises on work done to form an open standard, to parasitically drain users, and then deliberately break compatibility a year down the track. FB messenger literally did this when it started too - supported a bunch of different protocols to capture everyone, and then let them break.
That said, I’m impressed teams at Facebook have managed to stop working on stealing features wholesale from TikTok and Zuck’s anaemic VR disaster long enough to knock out an app to parasitise a different community.
yes, yes... let's cheer at Facebook doing things right...
yeah, I'm not sure why he's going all-in on a fringe candidate who's clearly never going to win