I'm not sure where to find clear "membership" of the PBLLC but I assume they have to file paperwork showing something where they are registered. I haven't been paying too close attention but I'm sure some people are interested.
The guarantee is that the AT protocol doesn’t work that way. They literally could not, technically. You own your data, the entire protocol is built around that basic fact.
It's like insisting blockchain transactions are immutable when Etherium issues updates all the time to reverse transactions that happened due to hacks.
This is a really, really bad read. The point is that the data is built on the foundations of an enduring protocol, right now, and that if BlueSky were to pull some shenanigans it would be trivial for a user to pick up and move. The protocol provides a built-in incentive not to pull shenanigans.
Honestly, so many of the comments here are, but what if?? The what-if scenario is built-in to the design and is in fact more resilient than other protocols like ActivityPub.
I think it's important to note that 1) The AT protocol is nowhere near complete, and as far as I can tell account portability is not actually implemented yet and 2) There are no third party sites that use AT! Even if account portability existed, there's nowhere to go.
But if account portability existed and worked, imagine this scenario: Larry Ellison wakes up one day and decides that Elon is having so much fun that he wants a social media site too. The minute the sale is announced, all the account portability APIs are disabled, and the website switches to a hard login wall. They start aggressively throttling and banning users that look like scrapers. They send a DMCA takedown to archive.org, who obliges and removes the wayback copies. They disable viewing followers and following so you can't even get your graph out. Now what?
The problem of trusting someone else to host your social media page is that you have to trust them. If your answer is "Jack wouldn't sell the company", then I can't help but note that he sure sold Twitter to Elon. If your other answer is "Larry Ellison would never do something unethical and customer-hostile" then uh...
That would have to happen soon, and if it did, you'd be right. The vision of a decentralized social web would die again, and one billionaire would make another even richer. Some moaning would be had on the Internet, but the world would continue to spin.
But, if the AT protocol even makes it another ~6 months without that happening, it will become impossible. You won't need to "trust" anyone to host anything, and if you want you'll be able to take your data and go somewhere else. That's the vision, and that's so far exactly what the folks at BlueSky have been building.
So no, the answer isn't "Larry and Jack wouldn't do those things" the answer is, "They'll have to act very quickly and in ways that are completely out of character for them, not to mention the staff at BlueSky would have to all sell their souls which is also completely out of character for them as well."
But yeah, if you're a cynic I guess everything is on the table.
they could change the protocol going forward if they wanted but all of the data currently posted on the protocol from the past would still be migratable to another PDS who supports the "old" fork of ATproto. I don't think they could change all the data you already posted with a future protocol change, that's not how it works.
The DAO fork is nothing like you describe, and was a singular incident in the first major smart contract.
BTC had a similar fork in its early days to work around a bug that actually minted illegitimate coins, yet maxis sweep that under the rug.
But you knew this, you knew what the DAO fork was. You know when it happened. You knew it was the first major smart contract. You know the context it happened in. You know the community overwhelmingly supported it. You also knew it hasn’t happened since, yet you still went out and lied.
I have not looked into it very deeply yet, but as I understand the Bluesky project, this would not be as easy for Bluesky as it is for Twitter and Reddit.
You can use your own domain as your identity. Even when you use a service like bsky.app, you can use @yourdomain.com as your handle. And switch to any other service to publish under the same handle.
Anything can restrict traffic. SMTP has no lock-in and you can use your own domain, but Gmail could decide to send all your email to spam.
Anything can just start ignoring traffic and not showing content.
I remember when Google Talk used XMPP federation. Lots of people ran XMPP servers. Then Google turned off federation and only Google Talk users could chat with other Google Talk users.
Unlike email, Bluesky essentially has one service. There's no culture of openness there. It wouldn't be hard for them to decide that they'd become the single service in the future. 99% of users aren't going to leave for a different provider and they could simply black hole stuff from other providers.
I'm not saying they want to do that, but saying there's no lock-in is wrong at this point. Bluesky could shut down their ATProto entirely and just close if off and 99.9%+ of users wouldn't notice a thing.
Bluesky can only be considered open once there are many other home servers and the Bluesky home server isn't the vast majority of traffic. Let's say it's 2025 and Bluesky has 500M users and there are 10 other servers with 5,000-500,000 users. Let's say one of the alternatives seems like it might be a threat to Bluesky's business model. Bluesky simply stops sending its posts via ATProto and then users on those other servers need to migrate to Bluesky. Total lock-in. Now, if it's 2025 and Bluesky has 50M users, ATProtoX has 100M users, SkyNet has 200M users, SocialSky has 25M users, etc., then it's hard to lock someone in. Bluesky couldn't cut itself off from 85% of the network. Bluesky could cut itself off from 0.001% of the network.
Just because Bluesky is using the ATProto today doesn't mean they'll continue to. As long as Bluesky is essentially 100% of the content/users/etc., it's easy for them to pivot away from openness toward lock-in. It's only open if they can't cut others off and as long as they're the source of almost everything on the ATProto, they can always pivot away. Yes, with the ATProto you can move servers. That doesn't mean Bluesky will continue to use ATProto and have openness around posting/viewing content.
the key thing though is that the data that people post with right now on Bluesky is on ATproto, which can be migrated to another PDS without permission from your current PDS (bsky.social). so if Bluesky tried to defederate you could take your data and fuck off with it to another PDS who supports ATproto and bam, Bluesky can't do anything about it.
the only thing keeping lock-in right now is the PLC directory, which I think has plans to be owned by consortium eventually.
That doesn't seem like a mitigation to what the poster above you said, to me. It just means that if I don't like it, I don't lose my posts when I move. That's nice but it offers no guarantees in the face of a corporate actor harming the overall federation of a network.
You and others in this thread aren't grokking what this means in practice: it means that users are transparently redirected to your new home wherever you are. If a corporate juggernaut tries to harm the network, folks just pick up and move their identity. The protocol bakes in treating the provider like a dumb pipe. The protocol puts the power in the hands of the users and at any time they can just up and walk away, defeating the stickiness of the provider.
No, I definitely get that. Where your data is, or is not, has no relevance to a network member choosing to behave badly or harmfully against that network. Nor does it actually impact what they can do to the network itself. It simply provides you with durability concerning your identity.
I’m not sure how your argument serves as any counter to “VCs invest in companies to make money” except that the shell game may be one deeper in this case.
Bluesky is VC funded and there’s expectations of revenue growth down the line.