>An instance admin shouldn't be allowed by the protocol to decide who you are allowed to read or follow. I didn't downvote your comment but how would it even b

>An instance admin shouldn't be allowed by the protocol to decide who you are allowed to read or follow. I didn't downvote your comment but how would it even be possible using computer science principles to design a protocol that prevents admins from deciding what bytes go in and out of the machines they pay to run? Maybe I'm misinterpreting but a simple reading of your complaint seems like you're asking for the impossible. EDIT reply to : >One method is by end-to-end encrypting those bytes, and optionally some of the metadata as well. If there's a hypothetical new ActivityPub version 2.0 protocol that encrypts everything so that the any and all instance operators are forced to accept messages and images of sexually abused teen girls, those admins will choose to _not_ to run that encrypted protocol. In the end, you still haven't created a protocol that removes the ability for voluntary admins to decide. Yes, a limited scenario of 1-to-1 communication where encryption hides a conversation between 2 people (e.g. Signal) from the instance operator can happen but a 1-to-many followers broadcast model like Twitter/Mastodon/etc with federating/forwarding messages like USENET is different. How would a protocol for 1-to-many force admins to accept those bytes? EDIT reply to: >Yet people are still using freenet. Yes, of course _some_ are running Freenet. But that doesn't address my point: you can't design a protocol that _forces_ humans (aka "admins") to accept any and all bytes. There are many that choose to _not_ run Freenet.