Post

Mastodon can't succeed. It doesn't have quote-reskeeting and has no good content discovery algorithm.

can't tell if you're being absurd or are incapable of understanding that there are different kinds of success.

"a place I don't want to hang out" is a failure to me. Mastodon users can disagree.

Is this a sarcastic troll account or...?

Anyone who has used Mastodon would know I'm being 100% serious.

I think federation is a potential future but the current fediverse as an implementation has failed to make meaningful progress on UX for too long now. IMO they’ve blown it beyond recovery.

It's worth considering what the goals are, and judging about whether a platform is succeeding on the terms of the goals it sets out, not what other platforms do.

I agree with that and I don’t want to conflate Mastodon with ActivityPub, but to Tierney’s point, I think a lot of the decisions involved in designing the most visible parts of the “fediverse” have arguably gone against their own goals.

i think that would be a great outcome tbh

Do you think that ultimately ends like email has, though, with a few highly concentrated providers and a handful of nerds running small instances? (I ask seriously and would still see this as success, too)

There are lot of small-medium size email providers, such as protonmail, fastmail. It’s not just a choice between DIY and GMail.

I can’t speak for Anil, but yes, I do. I don’t think that makes it any less successful tho. Open ecosystems tend to become centralized for lots of reasons, but if they are open, you’re not locked in the same way. To paraphrase Jim Barksdale, there are only two models, centralizing and decentralizing

I'm with you -- I think email is in a better place than Twitter. But email's also a good reminder that protocols aren't products, and for a lot of folks Gmail (or Outlook) == email, and those providers are happy to wall their gardens despite remaining nominally interoperable.

I love this "big tent" fediverse idea, it's a great vision. And personally, selfishly, I've really enjoyed implementing multiple protocols in the same systems and tools, I plan to keep doing that. No need to choose or fight turf battles more than necessary.

...one interesting quirk of the analogy though, email has only ever been multi-client for client/server, which is easier because adoption is much more isolated. Email's *federation* has always been single protocol, SMTP. I wonder if driving multi-client adoption in federation is harder.

Isn’t the “fediverse” a term already used for a different protocol than ATprotocol?

It’s targeted toward the ActivityPub protocol because that’s the first decentralised social protocol. However, I think as more protocols open up, then it will expand. https://fediverse.info/

well put.

Having contributed to Mastodon and being in the ActivityPub space for a bit, I really don't see there being direct interopt between Mastodon and atprotocol services. Mastodon's growth alone is the biggest threat to the fediverse imho.

Could you elaborate? What do you count as part of the fediverse, and what is the threat here? I never used or participated in mastodon, so I may be missing some context.

I think the fact that we're referring to it as "Mastodon" is kind of a big point in this conversation. Mastodon doesn't adhere strictly to the ActivityPub spec and for a lot of people, Mastodon, ActivityPub, and the fediverse are one and the same.

We haven't seen an alternate way to sign on yet so bluesky is essentially an invite only gatekeepered network for now? Albeit with many clients. Until it crosses the federation bridge it's mostly a promising protocol design. That is where mastodon UX stumbled so 🤞🤞

I’ve been told the invite system will stay. Only for Bluesky of course, the protocol AT itself doesn’t rely on invites.

Do you know more on how an affiliated network could be stood up? PDSs need to pubsub to each other? Is bsky.social a "provider" that handles the authentication? Who/where holds the private keys generated after invite is validated?

Corollary, if Mastadon fails (for whatever definition of "fails") it will be because ActivityFeed isn't a good enough protocol.