@twtxt (#5gb3dqq) I didn't know unfollow events were publicly broadcasted...

(#5gb3dqq) @mckinley Yeah they are, but to be honest I _think_ they're a bit pointless. I _might_ just remove them. Less code :D

@mckinley @prologic (#5gb3dqq) The same goes for follow events, isn't it?

@lyse (#5gb3dqq) The utility of follow events I think is important for discovery 🤔

@prologic (#5gb3dqq) I consider this a violation of privacy. Telling feed authors that you are following them is one thing, but disclosing this to the whole world is problematic.

@lyse @prologic (#5gb3dqq) That's an interesting point. It most certainly aids discovery, but at what cost? I hadn't considered that, up to this point.

(#5gb3dqq) Honestly, the entire follow system is flawed. Check my followers, #3 was a web crawler with a user agent that happened to fit the regex, and #17 was myself requesting my own feed with a simple curl command. Unfortunately, I don't see a real solution to the problem while keeping the ability for external feeds to show up as "following" a user on a Yarn pod.

(#5gb3dqq) The user agent regex was made a little more restrictive after my git issue, but I think someone could use this and really start breaking things. I want to poke around more than I already have, but I'm not doing it on a live production instance of Yarn.

@jlj (#5gb3dqq) Admittedly, once you start interacting with others, it's pretty clear, that you're following them. But if you just want to silently read somebodys messages, that's not silent then.

@mckinley (#5gb3dqq) Hahaha, great fun! :-D I reckon, the following stats need to reset after a specified amount of time (e.g. a week or whatever), so that they get cleaned up a bit over time. Not perfect by any means, but probably a whole lot better than counting a single request from sevens years ago. Unfollow events from other systems aren't propagated to yarnd, so that would take care of this.