(#zbsrabq) In other words, choosing to follow an actor does not follow the same model as Yarn/Twtxt does, where we then decide to "fetch" their feed. In fact it's the other way around, you end up telling server X that you're interested in actor Y, and servers X will then "push" stuff back to you. You can see how this can lead to some of the infamous DoS behaviours that can occur in the wild 🤣

(#zbsrabq) @marado Well, the reality is that you cannot actually control this too well. Because of the push model of Activity Pub you have to a) trust that the servers you are interacting with are both correct and not malicious and b) have to send `Undo+Follow` activities to actors you no longer want to follow (missing right now) Problem that I've found in practise however is that even for a tiny pod (used for development) I'm seeing interacting with servers I never interacted with (ever).

👋 Hey y'all yarners 🤗 -- @darch and I have been discussing in our Weekly Yarn.social call (_still ongoing... come join us! 🙏_) about the experimental Yarn.social <-> Activity Pub integration/bridge I've been working on... And mostly whether it's even a good idea at al, and if we should continue or not? There are still some outstanding issues that would need to be improved if we continued this regardless Some thoughts being discussed: - Yarn.social pods are more of a "family", where you invite people into your "home" or "community" - Opening up to the "Fedivise" is potentially "uncontrolled" - Even at a small scale (a tiny dev pod) we see activities from servers never interacted with before - The possibility of abuse (because basically anything can POST things to your Pod now) - Pull vs. Push model polarising models/views which whilst in theory _can_ be made to work, should they? Go! 👏