So it turns out Twitter's new open source effort for decentralized social media (@bluesky) is NOT what we were hoping, unfortunately.
Thread
Thread
First, it's a "change of backends" for Twitter, so by design the new decentralized backend would have to support all current Twitter needs, and specially business needs.
This rules out a lot of current protocols, I'll explain why...
This rules out a lot of current protocols, I'll explain why...
Twitter is "big world" social networking. In other words, social *media*. It needs to allow the use case of an account being followed by millions, and these million followers getting push notifications nearly instantly.
"Big world social nets" connects the whole world together.
"Big world social nets" connects the whole world together.
Big world has upsides and downsides. Upsides are reach and discovery. Downsides are a ton of abuse issues: spam (look, that's reach!), foreign interference in elections (reach too!), randos (reach!), etc. In other words, scale, Silicon Valley's favorite word.
"Small world" social networks focus on simple friend connections and small communities. Scuttlebutt (SSB) is a SW social net. Mastodon, to some extent, too, since intra-instance communication is far easier than cross-instance. If I'm not mistaken, Matrix is SW too.
Small world means that the application or protocol does not aim to provide the user a worldwide complete view of all other activity on the social network, so it's partitioned either by predefined communities (Mastodon) or the nearby social (sub)graph (Scuttlebutt).
Jack's thread was quite vague, but Parag's thread has more details. And guess what. They want blockchains! Well, not surprisingly, Jack was already bullish on that.
They also mention a need for data-intensive processing, which confirms the big world assumption.
Blockchains fit all those requirements: decentralization, open protocol, big world.
Blockchains fit all those requirements: decentralization, open protocol, big world.
They want the protocol to support business needs, because the big world assumption means big data, this rules out local-first protocols where data processing is local. So you end up with "miners" or whatever that need incentives to process your data.
So while I applaud their effort to do this open source (compare this to e.g. Facebook!) this seems a lot more like "continuing Twitter-as-is under decentralized backends", then it seems like "let's alter Twitter so that it's based on decentralization".
Move on folks 👋
Move on folks 👋
