Embed Notice

something i just realised is that in a way, 2 self-hosted XMPP or Matrix users talking to each other reveals more metadata to ISPs and law enforcement (and a man-in-the-middle)

rather than "this user is talking to people on [platform]" and (if they are law enforcement) having to ask [platform] for access to metadata (which could be rejected), they have immediate access to see "Hey, this user on [self-hosted single-user domain] is talking to this user on [self-hosted single-user domain]"

definitely something to keep in mind if that kind of thing is in your threat model