Post

A lot of people are concerned about the potential for abuse with this custom label feature What's the worst case scenario here, given that labels are only visible to the people who opt-in to seeing labels from that specific labeler?

Co-ordinated harrasment campaigns can already share links to Bluesky profiles that they plan to target using off-site mechanisms like shared Google Docs, WhatsApp groups etc Is the threat from custom labels purely one of making things more convenient for those groups, or is there something else?

My opinion right now is that the features to watch out for are features that make some form of harassment possible that isn't possible at all if the feature doesn't exist - and I don't think custom labelers qualify for that I'm very ready to be convinced otherwise

0 reposts22 likes

... and here's the inevitable US politician labeler showing their top corporate donors! bsky.app/profile/hips...

Andrew Lisowski 💻@hipstersmoothie.com

My latest labeler is finally working! What it shows for each politician - 3 labels for the top corp donators - 3 labels for the top industries that donate to them Subscribe to this labeler to be more aware of the interests our politicians represent 🚀 bsky.app/profile/us-g...

Ah, you mean you can't opt out of being labeled, is that it? If so, I misunderstood your point there, I thought you meant "into your face" as in "shown to you no matter what"

From my experience actually setting this up, noone is going to bother with using the label system as a way to target harassment or vector people while the list system still exists - labels basically only exist for this very limited use case, and are a bit of a royal pain to actually set up and use

- you need a server to run the labeller, a domain name and public IP to link it to, and then the actual process of applying labels is in a moderation-style report-queue system. The labels are then not searchable by others. Lists, on the other hand, are baked into the app and really easy to set up

Bluesky has an open API (as does Mastodon), so if someone wanted to build a custom app for targetted harrasment there would be nothing to stop them from doing that So labels is less friction for building something like that - but I agree that less friction does matter

On Twitter you could build that as a Chrome extension, which these days isn't necessarily that much harder than building a custom labeler - but of course Chrome extensions aren't an option for the Twitter native mobile apps

I think it's orders of magnitude less likely that people will subscribe to a bunch of chrome extensions (or bother to build and support them) for this than they are to do the labeling thing but just because it can be done now doesn't make it good?

It’s info that’s already published, just a new way of organizing and delivering it for consumers. As long as the labels are accurate, I don’t see the issue, but unsurprising some public figures may be uncomfortable with the factual record

Labeling can totally become harassment. I run a nostr labeling bot for offensive behavior and language. The folks getting labeled definitely feel harassed by my bot. It’s ironic.