Why would they expect anyone to build on top of their new social networking protocol?
It is owned and will be driven by a private company and it’s desire for profitability, ActivityPub is an open standard that has been around for ~5 years and battle tested and already in use across different platforms.
Maybe their efforts will be co-opted, but judging by the team's background -- André Staltz with Secure Scuttlebutt, Paul Frazee with dat/Beaker/Hypercore Protocol -- I think it's pretty clear the driving motivation for much of the team behind Bluesky is not profit.
As for why people might want to use it instead of ActivityPub, I think it's just... different from ActivityPub. Even though ActivityPub is an available option I don't see why it should be the last word.
Bluesky's answer is this
> Account portability is the major reason why we chose to build a separate protocol. We consider portability to be crucial because it protects users from sudden bans, server shutdowns, and policy disagreements. Our solution for portability requires both signed data repositories and DIDs, neither of which are easy to retrofit into ActivityPub. The migration tools for ActivityPub are comparatively limited; they require the original server to provide a redirect and cannot migrate the user's previous data.
> Other smaller differences include: a different viewpoint about how schemas should be handled, a preference for domain usernames over AP’s double-@ email usernames, and the goal of having large scale search and discovery (rather than the hashtag style of discovery that ActivityPub favors).
>judging by the team's background -- André Staltz with Secure Scuttlebutt, Paul Frazee with dat/Beaker/Hypercore Protocol -- I think it's pretty clear the driving motivation for much of the team behind Bluesky is not profit.
Not to be overly cynical, but I just don't care. Things change. When mega-corps and investors come knocking with 10-figure checks... 99.9% of humans will cash out. It's human nature. I've invested my time and energy in many things that have been taken over and destroyed by bad/incompetent actors, virtually all of which had promising starts, or I wouldn't have gotten involved.
The organization is a future adversary. You build the technology knowing you won’t be staying at the company forever, and you signpost the things that protect users / your-future-self. This includes open sourcing everything, moving specs to standards bodies when they stabilize, and building the network around low switching costs.
I can’t predict the future. We may screw it up. I’m trying to protect the community from us if we do.
That's not a cynical take for anyone who's observed the past 20 years of the web.
Page and Brin explicitly spelled out how advertising could change the incentives of a search engine. That didn't stop them from selling out immediately and letting the VCs install a CEO to 'businessify' the company.
And besides the current people themselves changing, there are also many other ways in which incentives can change: new people get into the mix (investors, board members, c-level), people leave etc. I recently formulated it like this:
As long as an experience of something that I want to have in my life can be influenced by a single entity I have to hope the incentives of that entity do not change drastically.
As a small nit, André Staltz is not directly affiliated with Bluesky. Paul and a couple other folks on the team have a track record with similar open projects.
> We consider portability to be crucial because it protects users from sudden bans, server shutdowns, and policy disagreements
I think this will have the opposite effect. Account portability means that being banned on one site can ban you on another if you try to use the same ID. Being persistently recognized across accounts seems like a net-negative for all of these things.
Same with mastodon. Your id is tied to the server so if you get banned you have to create a new id on another server.
Mastodon is worse because your data is in the server owner’s hands. At least with Bluesky your identity and server are separate so you still have your data.
Exactly, it's just removing the pointless (and terrible UX) step of having to sign up again.
I imagined trying to explain to my mom why she'd have to signup for Mastadon again because the server she signed up with is gone or she got on one of the massive blacklists for engaging with the 'wrong' accounts.
More than that, it puts your data into your hands. If someone kicks you off a server for whatever reason you still have your account and you can visit other servers. With mastodon you’re at the mercy of whoever owns the server.
The OStatus crowd made similar arguments when Mastodon added ActivityPub support.
"We already have a protocol."
"I don't know about this new upstart's motives."
I think it's good to have competing protocols at first. Platforms will eventually consolidate on one for a while like they did AP until something new comes along. I know the BlueSky people considered building on AP, so
I'm curious to hear what it lacked that they think they can improve on.
It could be that some new platform enabled by that new protocol drives a transition away from AP the way Mastodon did with OStatus.
> It is owned and will be driven by a private company and it’s desire for profitability [...]
I imagine that that's the exact reason why for example advertisers will opt for bluesky over mastodon immediately, under the assumption that the reach would be the same.
Mastodon’s UX with regards to following people on other servers seems to be really broken. Is there a simple ”browse to profile, click ”follow”’ experience on any client?
All mastodon clients I have seen involved logging in to the foreign server and/or copy/pasting handles. If this can just solve that it already removes critical friction.
Yes it's weirdly broken, but it seems to be hard problem to solve. Servers aren't connected to each other unless someone follows someone on other server, thus it's a bit hard to build a full graph in one server to browse profiles.
Mastodon team should have made own browser extensions a long time ago, now the community is working to make this easier. I use one called StreetPass [1], but it still is a bit barebones, and doesn't yet help with following accounts accross servers. It does make discovering accounts nice.
I found that the federation aspect of the fediverse breaks the things I most like about Twitter. Sure, there are advantages to it but they're outweighed in my mind by the disadvantages.
What you bring up is a big one. Another big one for me is search. I used Twitter search all the time. For local things like "what're those sirens I hear in my neighborhood", to cultural zeitgeist like "who is this artist performing on SNL", to using it as a quasi-Google.
Search on Mastodon is broken on purpose. Anyone running a server sits in a tricky gray area of potential liability from hosting user-provided content. Specifically for content that is considered illegal to poses or distribute in some countries.
Couldn’t somebody write a simple service that just scrapes profile info for following? Really doubt anyone would care if that aspect is centralized if it’s just a UX improvement. Same with search, although you’ll always have a vocal minority angry that their public posts are being scraped.
The extreme backlash against this in the fediverse makes that approach a non-starter; you'll get defederated as soon as anyone realizes what's going on. Having said that, AP/Mastodon not only have poor search and discoverability experiences, but also suffer from serious performance issues relating to federation and fanout that can have outsized and unexpected effects on servers (or, in Mastodon's case, sites hosting URLs linked in posts, since every Mastodon server will repeatedly bash the site to create thumbnails from the URL). Bsky's core approach of making indexing and "algorithms" an explicit and decoupled part of the overall stack is the kind of thing that I think should make it into the fediverse, but I feel like the way the community works in that space means that either it ends up in Mastodon (which de facto sets the standards now) or in a working group that never ends up delivering anything.
There’s no legal issue though, and I think most users are willing to accept the tradeoff if it gives them search. The ones who it really bothers always have the option of joining a private instance. Unfortunately posting public data by definition always means somebody is going to do something with it that you don’t like, you can’t opt out of the world.
The issue is thinking that this is simply a bolt-on ux improvement when it's really a core part of the ux. Such a service would quickly rival the value of mastodon itself and soon overtake it. It might even eventually become what non-technical people refer to when they say "mastodon", kind of like how many older non-technical people think of google as "the internet".
I don’t know about that since it’d really just be a read only cache of public mastodon data. Account ids, follower counts, most recent posts, etc. but it wouldn’t support any of the write operations, you’d still need federation for that.
I'm confused by this, this only happens if you discovered the user outside of your normal server (like if you opened a link to their profile/post from an external source). If you discovered them within your feed just click their profile within your home server and click follow. "logging in to the foreign server" is almost never the right move. Am I missing a particular situation you're describing?
Gotcha, yeah it is more difficult if you got there from a profile link on a website, and browser plug-ins can help with that. I was just confused about the idea that 'browse to profile and click follow' doesn't work, that works just fine in most clients as far as I'm aware. What the OP is talking about is 'remote follow' which is different.
Yeah the weirdest thing to me is abstracting away copy/paste seems so easy to me but every client I’ve tried actually requires the selection and copy and paste of handles… why not just a button to follow once I’m at a profile?
I know Akkoma has this with the 'Remote Follow' button, I'm pretty sure Misskey does too. It's really not that hard but you have to enter your home server domain for it to create the redirect.
The author asked why you’d choose this over Mastodon - the answer is that the AT protocol is designed to be a much better solve for decentralized social media that avoids the scalability and centralization problems of mastodon. That’s the entire point, it’s weird you would have access to this without understanding the entire motive for its existence.
There is literally no option no export / import your data from one instance to another. It’s a real issue and the the opened issue about it has been up for years.
Post migration is very important and is a part of the decentralization.
If the differentiation is better scalability, it only matters when there is scale. That is not a reason for a user to choose AT over Mastodon.
I don’t think Blue Sky’s app is more decentralized that Mastodon currently.
So I interpret your point as being: the reason to use Blue Sky’s app (or another new app with AT protocol) is that the protocol, in the future will be better. I agree with the author that I see no reason for a user to choose this over Mastodon at the moment.
I listed two reasons: scalability and centralization. You ignored the latter, which is relevant to users today from a pure immediate value proposition standpoint, but people do actually adopt technologies sometimes not based on their current form but because they want them to succeed so they can lead to a better overall outcome.
> That’s the entire point, it’s weird you would have access to this without understanding the entire motive for its existence.
Not sure if the author posted it here or not, but it seems like at least somebody decided to share the article with the entire community here, and I'm not sure why you'd assume that anyone who happened to click the article here would already know the details.
The Bluesky team and Jack have failed to publicly acknowledge ActivityPub to the point of me wondering whether they haven't done any research or are ignoring it on purpose.
So somebody being confused about the differences doesn't surprise me at all.
When Jack first started speaking about bluesky on Twitter, years ago, ActivityPub was a lot more niche and he ignored any message related to it. It definitely looked like a NIH as he could have started to discuss activitypub specification. I was myself heavily invested in decentralized protocols (from blockchain to activitypub) and really tried to understand what he was trying to achieve.
My feeling was, in the end, clear: he wanted a "decentralised protocol" on which he had full power (aka "VC-style decentralisation").
You have to keep in mind that those successful in the valley know only one kind of thinking: raise money, get users, sell off. They can’t grasp decentralisation other than as a nice marketing term to add to their product (and, as Ripple demonstrated during the Cryptobubble, they are completely right when it comes to make tons of money with shitty tech).
To my knowledge, acknowledgement of ActivityPub came after the huge Mastodon burst caused by Elon Musk. It’s more a "oh shit, we are not the first" kind of reaction.
Also, you don’t simply design a decentralised protocol behind closed doors then expect everybody to adopt it. You need to be transparent, to discuss in the open. People need to know who is in charge and why. They also need to know every single decision. Decentralisation cannot be done without open source paradigms. That’s the very point of it.
I think we can safely assume that Jack has simply hit his mental glass ceiling. He doesn’t get it. He doesn’t have the mental model to get it. He will probably never get it (he became a billionaire by "not getting it" so there’s no reason for him to change). The whole project is simply him throwing money at a few developers telling him what he expect to hear in order to get pay.
When you've already answered a question in your FAQ yet people keep asking you "what about ActivityPub" every day for years you either have to tune it out or go insane. It looks like NIH from the outside but it's probably not.
Also, you don’t simply design a decentralised protocol behind closed doors then expect everybody to adopt it. You need to be transparent, to discuss in the open. People need to know who is in charge and why. They also need to know every single decision. Decentralisation cannot be done without open source paradigms.
So... that's pretty much what Bluesky did.
The whole project is simply him throwing money at a few developers telling him what he expect to hear in order to get pay.
Ironically, it looks like the Bluesky people are not telling him what he wants to hear (Bitcoin Bitcoin Bitcoin) so now he's backing Nostr. And if Jack has really cut Bluesky loose we should stop blaming them for the (considerable) sins of the father.
Hate speech doesn’t depend on character count. The reason you see a better experience on Mastodon is that all the political flame wars are still on Twitter. If/when the platform grows and starts attracting those audiences, it will have the same problems.
I doesn't depend on it, but it encourages.
Algorithms have no chance to pick better articulated longer posts, since the content is shared over multiple posts. It is hard to get attention for those.
You probably get larger audience on a single post, which causes bigger emotions in people.
There's these weird semi-accepted articles of belief/truth that are all over Mastadon. I really have never understood, nor seen evidence, nor even really explanation. Believing a stream of short posts promotes hate speech is one of these articles of faith. (Right now Im pretty skeptical.)
When your word amount is strictly limited, you cannot really discuss about things.
Algorithm-based platforms pick usually posts based on extreme ends.
People are interested on posts which rise emotions on them.
If someone says something you don't like, you are more likely to like or share a reply, which somehow outsmarts and counters this original post. And as your word amount is limited, you will need to use stronger words to get attention and rise emotions on people. You are rewarded for expressing extreme opposite in a smart way.
Soon you will be in vicious cycle, since things get quickly personal, as there is no room to articulate anything properly, and you need to win this conversation with few words. Writing well articulated multi-post replies won't get much attention, as they usually don't rise emotions for masses, and algorithm won't pick them anyway as attention is not in single post.
In general, my high level note is kind of non-specific, that I am extremely hestiant to govern by fear, that too often it generalized specific niche behaviors while ignoring very positive healthy aspects of the subject. Fear without demonization is difficult to pull off, but I can confess, this is a bit of a slippery slope argument. Still, I far prefer regulation based on positive behaviors, based on better lights.
There's a lot of really good comments, challenges & issues you'be raised here. Im not trying to refute here, but there's a lot of things Im not certain on or have comments on.
> When your word amount is strictly limited, you cannot really discuss about things.
I've never felt this limit. I quite like having a series of replies, where I can break up my responses into different pieces. And I like how people can reply to individual parts of my reply. It makes it clear how people align with the various pieces of a discussion, rather than writers and readers having to figure that out for themselves.
The constraint on size creates a much richer structure of information that greatly aids discussion, in my view.
> Algorithm-based platforms pick usually posts based on extreme ends.
Im not sure how longer posts changes the algorithms fairly blind desire to drive engagement, which yes, can exacerbate/amplify extreme positions.
> If someone says something you don't like, you are more likely to like or share a reply, which somehow outsmarts and counters this original post. And as your word amount is limited, you will need to use stronger words to get attention and rise emotions on people. You are rewarded for expressing extreme opposite in a smart way.
Feeling very dualist on this. I heavily agree & disageee. This is a pretty interesting section to me, good challenge.
I think we overascribe how much people tailor their toots to get amplified.
Brevity is the soup of wit, because it feels good: the simple appeal of something small & compact that makes sense. We're not writing short & hot because we are forced to. We're doing it because we can reduce the situation down, provide a neutron-star dense distillation of perspective, a place to see the dielectic from. That is enormously rewarding to compact the idea & concerns & reply, to pull off a real view, briefly.
But it of course risks reductionism, and it doesnt leave a lot of explanation, or acknowledge the subtleties, or meet in the middle.
I still have hard time seeing post-size as having a strong relation to the specified problem here. We can make longer posts, with threads, and do... when the situation calls for it.
> Writing well articulated multi-post replies won't get much attention, as they usually don't rise emotions for masses, and algorithm won't pick them anyway as attention is not in single post.
I dont think allowing longer posts will change this dynamic in any way. Long posts are less appealing & less popular, broken up across multi-replies or not.
The way I can see myself being wrong about this is that upvotes are distributed over a wider range of posts, rather tham concentrated into a single post.
But there's two other counter-factors. Often on HN I run into comments where Im absolutely loving the post, so happy to see something expressed so well. But then it takes a turn. Something goes off the rails or it uses a bad argument or comes to a bad conclusion. Being too big is huge risk.
Also, for a long thread, many people will have multiple upvotes. There wont be a single comment with massive upvotes, but in aggregate the thread probably has many more upvotes total. If the algorithms arent properly looking across threads to promote & surface, that doesnt seem like it should mean changing the medium to adjust to the algorithm, it seems like changing the algorithm to understand the medium should also be a considered possibility.
Last I checked, Mastodon defaults to a 500 character limit which involves a bit of work to charge (it's hardcoded in 2-3 places). I think both the character limit and the default image crop are design flaws.
The problem is that political combats and campaings have been made on the same platform, and those might decide the future of the country and many people.
E.g. Twitter is not for really for general people to provide content, it is mainly for those with large audience already. And they want attention.
> “Provide content” you say this like there’s some objective measure for “good” and “bad” content, but there isn’t.
There is. The algorithm decides what you see and not see.
If someone has already larger follower list, it is likely that their posts will appear to others feeds whom are not following them, if the post causes strong interactions.
One of the largest complaints about Mastodon is the complicated (yes it can be complicated) sign up process. You have to pick an instance and if that one is full, pick another, then another, etc.
By allowing everyone to register on their app to their own servers? Mastodon is now prioritizing large servers with open registration in joinmastodon.org for this same reason. They re-ordered the servers some weeks ago so that open registration servers come first.
Hahahaha, apparently DeepL forged a new word! It's (supposedly) a translation of “mambembe”, pt_BR adjective that means something ordinary, mediocre, half-baked, average (in a bad sense). (Plus, it's sounds amazing, although I acknowledge that even in Portuguese it's an uncommon word.)
To avoid (more) confusion, I replaced it with “half-baked”.
Easier account discovery/follow/porting is the surface. The core differentiator of Bluesky/ATP is the ability to have a global view for its network, as compare to Mastodon's limited local view. Things like search engine is not only possible but first class citizen in ATP design.
Obviously depending on your attitude towards ad tech, you may like or dislike it. However, if everything still (eventually) follows money as it used to be, and Twitter's 5 billion ad revenue (or more if you believe FB is shrinking too) wants to go elsewhere, ATP is offering a viable solution while Mastodon does not (at least yet).
The article mentions running separate servers to be able to claim your own domain. However, in AT I believe you can simply add a TXT record. There is no need to run a server unless you want to.
Also, you can be @example.com if you want. You don’t need to be a user on a server like @me@example.com.
Tbh, none of them really attract me right now. They are too clunky and don't have many people to interact with. I'm slightly annoyed by Nostr's proposition and total lack of moderation controls. Even if its developers are well-meaning, I see this as a dangerous, undesirable position to take.
I'm really comfortable with Mastodon/ActivityPub. My personal profile is in a public instance[0], and for my Portuguese-written site I'm hosting a Microblog[1]. It's just great.
> ActivityPub has the advantage of being a de facto standard and not being owned by a private company.
It should be clear by this point that "being owned by a private company" is not a disadvantage. If the Bluesky team nails the onboarding experience, then Mastodon/ActivityPub will remain a niche service.
Interesting, I wonder how heavily censored it will be considering it's run by a regime of people formerly attached to old Twitter. My guess is "very to almost completely."
Hi, Jay from bluesky here. Actually nobody on the team, including myself, has ever worked at Twitter. We're designing for flexible and composable moderation systems driven by user choice.
Good deal! I just can’t trust anybody from old Twitter after their abuses. Looks like I’ll have to wait and see how your product develops rather than just writing you off as a sort of zombie mode Ole Twitter.
I don't think anyone inherently wants censorship. It comes down to it when popularity rises and certain standards are expected of it. Would I mind if I had a 8 year old and he saw "fuck" on a website/social media? No I wouldn't but some people decided that we should have a specific linguistic moral code, so the capitalists try to walk the path of least resistance on this topic (just complying)
And yet, I'll bet that your version of "THE MEDIA THEY CONTROL" doesn't include whichever publication has politics that you, personally, agree with. How mysterious.
It is owned and will be driven by a private company and it’s desire for profitability, ActivityPub is an open standard that has been around for ~5 years and battle tested and already in use across different platforms.