I don’t understand how ActivityPub works that well and I haven’t used Threads so maybe I just doing get it.
But why would we have to worry about defederating Threads when it’s a Twitter clone? It doesn’t have communities and such so I would think it’s not compatible. We’re not federated with Mastodon instances right? This seems like something for Mastodon to worry about. How is a Lemmy instance refusing to federate with Threads relevant?
Do we even know is Threads will federate at all? I figure meta is using activitypub simply because it’s cheaper and easier than coding their own, not because they intend to federate. Zuck wants private data so he can sell us ads and propaganda. Allowing people to follow from federated servers both denies him user data and the ability to push ads.
The Threads app itself clarifies that the intent is to federate down the road.
I signed up for Threads, and there were 3 upsides of Threads on their welcome screen, one of them was that it will be part of the fediverse someday. It doesn’t mean they’ll do it, but they definitely intend to use ActivityPub as more than just an internal protocol.
Could be that Threads will use ActivityPub to feverste with other Meta platforms in the future too, assuming they want to add such capabilities to Instagram and Facebook.
It could be, but they specifically used the word “fediverse” in their marketing. It doesn’t make it sound like they want to create a new thing, that would sound like they want to integrate with the existing fediverse.
The Verge has an interview with Adam Mosseri and he mentions ActivityPub will be coming soon™️. Whether or not it actually happens, we’ll see.
https://www.theverge.com/2023/7/5/23784870/instagram-threads-adam-mosseri-interview-twitter-competitor
It is a ready-made software that could let them quickly launch a product, and it can deflect the allegations already made of intellectual property theft from Meta hiring former Twitter developers. Of course they could develop their own, but that would take time and added risk.
What benefit would Meta get from federating with the likes of Mastadon? They want profit, and I do not see a way Meta makes money from letting people follow Threads users from Mastodon, KBin, and Lemmy.
activitypub is less ready made software and more a set of predefined constraints and protocols
definitely more work to federate than to just have a solution that talks to a database! especially at scale
Why did google implement XMPP originally for Google Chat? They originally did federate with XMPP servers. But eventually, they decided that federating wasn’t worth it. At the time they still used XMPP, they just refused to federate.
Why wouldn’t you use ActivityPub as a protocol if it has been proven out and does 90% of what you need? Just because they use ActivityPub, I don’t see them ever Federating in any meaningful way. Essentially, I think the protocol is well designed and Meta is just using it as their starting point with no reason to share any data with the fediverse.
Was that meant to rebut the idea that Meta wants to federate with other ActivityPub services? The fact that Google did the same thing with XMPP, gained user share, then defederated once they achieved critical mass - classic EEE?
They certainly could, but using something off the shelf saves development time and costs. Not only did someone else already do the base work, but they are fixing bugs and adding features as an ongoing task. And that all happens free, without Meta spending a dime. Meta only needs to add their customizations.
There’s been plenty of speculation on why they want to federate, which is much less clear. It could be an attempt to get around EU antitrust (etc) laws. It could be an attempt to usurp Mastodon as the primary destination for Twitter refugees. It could be an attempt to slurp up the data from people that refuse to give it to Meta. But this is all just speculation, and it’s unlikely that they will honestly reveal their reasoning.
Are they using an off the shelf ActivityPub implementation (if so, which?) or just implementing the protocol themselves? If it’s the latter, which I expect it is, then implementing the protocol does not save development time or effort. It’s just a set of specifications that they decide to conform to, rather than doing things some other way that may better suit their business goals.