[danpalmer]

I agree, but I feel this is a bit rich coming from Mastodon which is notorious for depending on a load of custom ActivityPub extensions for server to server operation, and just having an entirely separate REST API not based on any open standards (beyond JSON/HTTP) for its client to server comms.

I used to run a non-Mastodon Activity Pub server and it never really worked. There’s a reason why everyone just uses Mastodon, at which point who cares if there’s an open protocol.

[verdverm]

Have you given ATProto a look or try? Thoughts?

[danpalmer]

Yes, I believe ATProto is a much more mature approach, that is solving much more fundamental issues. The criticism of "it's not federated/distributed" feels unfounded, the argument against VC funding seems unnecessarily zealous (if the protocol works it doesn't matter, so critique on this basis is in bad faith). I think Bluesky also seem to understand that the protocol will at least for a while, live or die on the content that is there, and Mastodon doesn't come out well in this regard.

Activity Pub itself is... a bit boring? And maybe that's a good thing? I tried implementing it though, and it's very unsurprising to me that a) Mastodon is the only game in town, b) why they use custom extensions and semantics, and c) why none of the other interesting content types are being used in any substantial way.

I think Mastodon/ActivityPub is a local maximum based on a 2010-2015 understanding of what social networks are/need, and that ATProto and Bluesky represents a realistic 2025 era understanding of the social web, with plenty of scope for advancement and staying current in a way that I doubt Mastodon can ever do.