My mind just blew out the back of my skull.
Only took (2024-2007) 17 years for my original concept to be fully realized in code.
I mean, there you have it, folks:
One simplified way of thinking about labelers is that they let anyone attach "hashtags" (labels) to any post or profile on the network.
But only users who subscribe to a "hashtagging service" (labeling service) will get them, and client apps can do kind of anything with these "hashtags" (labels).
“let **anyone** attach "hashtags" (labels) to **any** post or profile on the network.”
Right now all I can think of is that implementation in the above specific scenario is going to be so heavily abused by bad actors 🫠
(I mean, yes, cool applications too, but, oof)
Which is fine. If a tree fall in a forest that you never visit, are you in danger?
The whole point is that a bunch of "moderators" can go to town labeling content, but if no one subscribes to them, their efforts will have been in vain.
This is good.
Nature is healing. 🌵
But then you also have the LibsofT*kt*c problem.
Small, organized yet loud nefarious groups will use this for targeted harassment or worse.
They very much are putting people in danger.
I fail to see how this can be abused in that way, but maybe I’m just not imaginative enough.
Each label contains the ID of the labeler, so you have to opt-in to subscribe to a bad actor’s labels. Mentions/screenshots are already a vector for pile-ons; labels don’t create any additional risk IMO.
That's my understanding as well; if there's some vector that isn't obvious here, it'd be helpful to understand given what I understand to to be intent of the Bluesky team.
