bnewbold
commented
4 months ago This is close to things we have discussed in earlier iterations of the labeler / mod service iteration. Some related ideas are the ability to block (or label) entire PDS instances at a time, etc.
As a direct technical answer, we can't add this specific product feature easily with the system we have today. Labels and blocking are different systems. There are scaling issues trying to do "millions of people blocking millions of other people" that we would need to resolve if we need to support that from a product/community standpoint. Doesn't mean we can't, or that we are decided against this, just that it isn't a quick-sprint kind of thing to push out.
As a specific example, labeler subscriptions are intentionally private. But blocks have publicly-observable effects (can tell that there is a block relationships between two accounts).
I definitely expect us to iterate around and consider more interventions like this, but I would guess there are some other protections and interventions we will try first (these aren't specific features on a roadmap/timeline, just ideas bouncing around):
- group-private accounts and content: flips content availability and interactions from "block list" (everybody but these accounts) to "allow list" (only these specific accounts). this is a frequently requested feature, further down the road technically, and isn't a super-set of this request (aka, there are visibility and reach issues with group-private), but i'm guessing will have a big impact.
- ability to manage mod lists (block lists) from Ozone directly: would make it as simple to add somebody to a list as to add a label, from a moderator UI/UX perspective. there might be bots or other tricks to help automate this (using the Ozone APIs) in the near-term, without needing bsky dev work. I'm a bit skeptical of leaning too hard on mod-list-blocking, especially as things grow, but it is working to some degree and could be more ergonomic.
- iterations on "interaction controls" (eg thread-gates), like having those be automatic or account-wide, and possibly driven by lists. maybe in theory these could be label-driven as well, if some technical issues can be worked out.
monicaellerose
Currently with the labels we are able to: show, warn or hide content that has a label. Issue here: certain labels exist where people who post under that content are people [many users] want to block automatically. here the ability to hide content isn't enough because bad actors can still engage with our profiles, while logged in.
Im not sure the right place to post this 'bug' or feature request because because of where it spans.
Trying to fill in the gap where lists fall short, lists can't keep up fast enough, and not all label services have lists for users to subscribe to. Adding to the complexity that if a new member signs up tomorrow, its an additional step and confusion for them to have to also add or subscribe to lists
I hope this can get added or considered to help minimize current and future issues and minimize holes in moderation