This is... the least charitable possible interpretation of the current state of Bluesky and the AT Protocol. It's not wrong, but it's not clear that it will be right either.

Personally I would have preferred if they implemented federation before going into beta because that is the whole point of a decentralized network. But OTOH there is an opportunity right now to lure people away from Twitter and that opportunity may not exist in a few months. I don't think progressive decentralization, over a span of months not years, is inherently evil.

From what the CEO said, they were giving out invites from the start and no one was using them and then suddenly the momentum hit and everything was gobbled up and lot more than they anticipated joined.

> The DID itself is derived from the sha256 hash of the first operation in the log. It is then base32 encoded and truncated to 24 chars.

https://atproto.com/specs/did-plc#how-it-works

They could be using public keys as identifiers, why are they using nonsense?

Would public keys be recoverable without having the private key, one of their objectives?

They also say: “We're actively hoping to replace it with something less centralized.”

They could replace it with allowing you to hold onto your own keypair.

because they want the ability to perform key rotation

well exactly. If they hold onto your private keys for you, then they can keep you from losing them and also give you a new public key on a set schedule.

but they're also holding into your private keys for you, which comes with some risks.

> because of that they are tricking a lot of people into using them

This may come as a surprise.

But your average cool, interesting person doesn't care whether Bluesky has a decentralised protocol or the intricacies of its identity model.

The reason most people are using it is because it's the closest thing to a Musk-less Twitter.

Which is funny because this demographic values blocking, banning, and censorship very highly while these are exactly the sort of things Nostr and Bluesky are designed to prevent.

Irony is lost on these authoritarians who are anti-free speech.

Totally, see the other reply thread - not even the engineers building it seem to understand the irony.

Bluesky at least is not design to prevent "blocking, banning, and censorship" at all, not sure where you get that idea from.

Blocking/moderation/"censorship" will happen from people's own actions, or subscribing to someone else's actions. If you want, you can just not participate in blocking/moderation at all and have a raw view, but most people will probably opt-in to a moderated view, as the internet is usually (not always) viewed with a moderated perspective.

Huh? But blocking someone on a distributed protocol seems impossible? to block someone, or censor them, you would have to get every node in the network to cooperate. Sure you can “mute” someone and not see their posts personally pretty easily. But if you want to make it so that others can’t see someone’s posts, like Alex Jones or Trump, and generally block the public from seeing them, this action seems much harder.

Yeah, if what you want is to prevent others from seeing other-others content, you're gonna have a hard time implementing that in a decentralized/distributed protocol, just as you say.

You basically have to chose between that or allowing the platform to censor anyone, which these "new breeds" of social networks seems to make the choice of doing "personal blocking" rather than "network wide" blocking.

What I think most people want though, is not "blocking" in terms of "no one will be able to see this person" but "blocking" in terms of "I don't want them to be able to interact with me, nor do I want to see their content".

I think they do want to censor people like Trump and Alex Jones, and prevent others from seeing them. That’s “deplatforming” and the people that are leaving Twitter are very pro “deplatforming” and are upset Elons messing with it, but then deplatforming is impossible on distributed protocols. Thus the irony, and my first post. Do you understand now?

You're gonna have to start citing something from folks joining bluesky saying that that's what they want, because inside the platform, no one is echoing that sentiment, nor does either bluesky or AT protocol websites suggest an approach where people are being blocked in that manner.

bluesky/AT protocol is pretty clear how they see moderation to work.

> Our approach to moderation is three-fold: automated filtering, manual admin actions, and community labeling

> Labels can be automatically or manually generated, and can be applied by any service or person in the network

> For example, an organization like the ACLU could create a “hate-speech” label, and services or users in the network could subscribe to their labels to flag, filter, mute, or ban content. The Bluesky reference app we’re building has custom controls that lets users hide, warn, or show content, and will add more controls over time. Through this approach, we aim to prioritize user safety while giving people more control and transparency over how moderation is done

https://blueskyweb.xyz/faq

Plus https://blueskyweb.xyz/blog/4-13-2023-moderation

If people, after reading that (or ignoring that) still end up bluesky, we can kind of assume that they agree (explicitly or implicitly) with the approach.

I pressed a button on Twitter to not show me anything from musk. Now it’s a muskless twitter. Meanwhile I’ve heard that alternatives like mastodon have half the user base taking about musk. That’s much harder to block.

And of course nobody on Twitter is talking about Musk.

Or posting screenshots of stuff he does etc etc.

Indeed, a few more button presses are needed to type "musk" and "Elon" in the mute words setting. Most users don't use or know about this function.

It's an optional thing to disengage from the hype but but not really obvious for normal users.

you can apply filter on Mastodon, not only on tags but on words themselves. simply filter "Musk" to have a Musk-free Mastodon experience.

This is wrong:

1. even average people do appreciate the fact that the thing is an open protocol, even if they do not comment about it or understand it or seem to care.

2. average people do follow some leaders when deciding where to move, and the leaders are people that mostly care about these things -- these are the ones being tricked.

The solution to the problem of centralization that Bluesky has been touted as tackling (and that Mastodon doesn't deal with in a satisfactory way) doesn't look anything like Bluesky or Twitter or Mastodon. Instead, it's just making RSS (Atom) suck less.

Crazy that no one in the relevant circles seems to have realized that microblog posts don't require anything more sophisticated than what a static site generator can produce. The heavyweight server-to-server protocols that we've seen are just way too heavy.

This is how I've been feeling but I'm assuming there's more to this that I just don't know.

I keep thinking, what if we: 1. Use domains/subdomains as usernames by just... having the content on them.

2. Follow a truly basic structure for index (feed), single post, etc. so that clients know how to consume them easily.

3. For interactions (replies, likes, etc), the client posts them to your own server (or service hosting you) and they're available at a url referencing the original post url on their own domain. ex) mydomain.com/replies/thepostsurl/1

4. It then posts the reply url to the original posts server endpoint, which can accept or not.

5. When a user loads the post, it lists the replies and interactions as a list of links. the client goes and gets their content directly, and renders it.

This would work without any servers actually having to talk to each other or store more than a link for interactions. If you want to confirm the interaction url on receiving a submission, your server could check but it doesn't need to store or cache it.

Am I crazy or would this not work just fine? or, is this what activity pub already is?

That's still too heavyweight. My static site doesn't have an "endpoint" for people to post notifications to. And it doesn't need one.

A like is nothing more than a way of saying, "I like this". Think. You can come up with a way to do this that doesn't require you to abandon RSS (read: Atom).

If Alice posts something and Bob likes it, then he can say so—from his perspective, he clicks "Like", and in turn this just ends up as another entry in his own feed. He doesn't need write permission to anything on Alice's server. If Alice is subscribed to Bob's feed, then she's already lined up to get it. If she's not subscribed to Bob's feed (and maybe even doesn't know he exists) but is so neurotic/insecure that she craves validation from strangers on the Internet after years of conditioning on Facebook, Twitter, Reddit, and HN, then she can subscribe to Marge's feed. Marge is Alice's friend. Marge casts a wide net (follows a lot of people) and keeps an eye out for anyone saying they liked Alice's shitposts. When she notices, she lets Alice know: Marge squirts an entry into her own feed saying, "hey, Alice, look over here at Bob saying he liked your shitty Beetlejuice tribute" (which is pretty much exactly the mechanism behind boosts/retweets—except these would be boosts/retweets not of content but of what is known at least in the XMPP vernacular as "presence" information). Also, Marge is actually Facebook/Google/whoever, once they realize how lucrative it is to have this kind of influence and mandate in the next generation of social media.

Previously: <https://news.ycombinator.com/item?id=30862612>

You are basically describing Nostr: https://www.youtube.com/watch?v=ssNkOMx2E5Y

I've looked at Nostr before, and I'm confused by how this response came to exist.

I'm describing RSS (read: Atom) with better UX. Nostr is a from-whole-cloth new protocol that's fundamentally incompatible with everything I've laid out.

(The video you linked to is unwatchably obnoxious, by the way.)

The biggest hurdle is the vast VAST majority of people do NOT want to host or run anything, which means there is varying levels of centralization, which introduces the issues people have had.

* * *

This is already way past the "average" user; for millions even the configurability of MySpace was a bridge too far.

There is an argument that people who can't be bothered to figure it out probably have nothing interesting to say, and that may be true in the vast majority of cases, it's not universally true; I'm sure we can find quite interesting social media accounts posted by someone who doesn't have the inclination or time to figure out "free static file hosts".

That's what Nostr is designed to fix: https://github.com/nostr-protocol/nostr

My thinking is that it would be a bit like email: run your own server if you want, and the other 99% can use any of the zillion hosting services that exist or would crop up to do it for you with the ease of Gmail.

I like this a lot. It also addresses (maybe even solves?) the moderation issue, which is the harriest problem to solve when content is centralized.

Even as a dev, I am not going to take care of 3 and 4. I just host my site via a static generator on GitHub.

Who's going to do it for normal users? That's what bluesky is doing and anyone who's interested can theoretically host their own servers.

> Use domains/subdomains as usernames by just...

Username should be owned. Domains are leased.

True, but from what I've seen the usernames on every social solution, decentralized or not, are essentially leased too aren't they? Either they're based around a domain you control or they're under someone else's control ultimately.

It’s fine to have a rational disagreement about systems design and deployment, but this borders on paranoid screed. Just because you don’t like something and other people do doesn’t mean it’s a scam.

note that fiatjaf is the creator of nostr, which is a competing product

But are they really competing products?

Also recipient of a ~$250k grant to develop NOSTR from Jack Dorsey himself, who is also behind Bluesky...

Biting the hand that fed him.

Dorsey also has stakes in Musk's Twitter.

I'm fairly sure Jack is currently just trying to get any social media effort off the ground, and throw money at a bunch of different efforts, hoping at least one of them work. He been pretty open about that process on Bluesky at least.

basically he's mad nostril isn't gaining traction

It may be a scam technically, and it may succeed commercially

Bluesky's innovation over Twitter is to try and free itself from onerous regulation and scrutiny by governments. Their defense, over and over again, will be "just run 'govsky'", "just run 'chinasky'", "we can't censor you, no need to start Truth Social (lol), also no need to give us super burdensome regulatory requirements."

They're doing things a little backwards, building a gmail before there's a big SMTP network out there, but that's smart. The default state of this positioning now is that they'll slowly drain Twitter of their attention capital ( this is on purpose; their UX is the Twitter UX) so Elon has to respond somehow.

That's not an innovation. The EU itself and at least one EU country's government already have Mastodon instances. Quite a few organizations have as well. Truth Social is a fork of Mastodon.

Where is Bluesky's innovation?

"Innovation over Twitter" is what I said.

I think ActivityPub and Mastodon are alright, but I think a thing that looks and behaves more like Twitter will be more successful at succeeding it than something that doesn't.

I get that Mastodon people are salty about Bluesky. Personally I was hoping Elon Musk buying Twitter would be taking the Jenga piece that collapses the social media tower. But this is how nerds vs capitalists goes, unfortunately, and my evidence is all the walled gardens that sprung up in the fertile pre-web-2.0 internet.

Being the second to do something (without significant improvement) isn't what I could consider innovative by any definition I know. Mastodon already did the things you listed. Can you clarify what you mean? It sounds like you expect Bluesky to do what Mastodon already did, but better somehow.

I think the miscommunication here is that you think I'm talking about Mastodon at all.

Edit: cool music btw! I like the sophistication of your percussion.

Why does everyone forget Matrix? There have been multiple truly decentralised Twitter clones on top of Matrix.

No aggressive promotion, no VC-esque backing. Which is very good for Matrix.

Very cool and very normal to crank off about decentralization/moderation and then have people in the author's comments section threatening people with violence. It's almost as if there are a lot of intricacies around making these networks work well for people and there's more problems to be solved than just raw unmoderated decentralization if you want to foster a healthy community that doesn't devolve into a tiny fraction of the worst people on earth.

Not necessarily endorsing this (I haven't looked into the details of Bluesky myself) but one thing I don't understand is - how does Bluesky intend to make money? Have they announced a plan? It's a corporation, so obviously they have something in mind. But googling about it, I can't find a good answer. Centralized social networks bank heavily on the network effect, but you don't really have that if you make a genuinely decentralized network like Bluesky says they're trying to.

name is important, agree or not. this name, bluesky wont fly.