I keep seeing people today say "these labeler implosions just prove that Bluesky needs a T&S team". Folks: Bluesky has a T&S team. @aaron.bsky.team runs it. They've gotten incredibly fast at handling issues. I've seen sub-5-minute responses on some of my reports.
The Bluesky T&S team are the only ones who can remove accounts from the network entirely. They also handle labeling for the official Bluesky moderation service, @moderation.bsky.app -- that's the labeler you can't unsubscribe from.
When Bluesky started, they were hoping they could minimize the number of "account removed from the network entirely" actions and use labeling to allow people to decide what they do and don't want to see, so there could be multiple options for people to choose which moderation viewpoint they wanted.
I had criticisms of the paper when it came out and I've had criticisms of Bluesky's approach to it; I've said all along that they're going to discover that there is *a lot* of content posted to a service that shouldn't be there & a lot of users extremely unhappy with a pure protocol approach.
My assessment all along has been Bluesky's experiment in moderation-as-a-service will only succeed in creating a viable community to the extent that they *do* also perform a degree of centralized moderation, and that the biggest question will be how they approach determining what that degree is.
The statements made by the team during the early days of the network left me fairly pessimistic they'd ever be willing to arrive at that conclusion, but that pessimism has changed *significantly* over the last 6-8 months or so.
Quickly, from my 6/23 blog post:
...hypothesizing that ranking can more or less fully subsume other forms of automation & include manual ranking and label inputs. ...Rankings can be based on many dimensions of attributes - so rankings could take a hybrid form that includes label attributes.
Oh, totally. There's plenty of stuff we leave up that people are very upset that we do (and vice versa). We start from "is it US-legal" as a baseline, but there are some deviations from it we make for the overall health of the service.
And even before the AEGIS implosion there were a lot of labels that people blamed AEGIS for that were official!
Like, Great Replacement Theory stuff *absolutely* gets hit by the official service of it gets reported- including account level tagging!
I mean, sure, but "better than Twitter" is not exactly a bar to aspire to. Taking whatever fraction of Twitter posts draw a block and doing that at random would be better than Twitter, in that occasionally you'd end up blocking a Nazi or transphobe.
I do think it's an important clarification that even they can't remove accounts from the network *entirely* - it's really more the network *that most users see*, but in the future when there's other services competing with Bluesky's, this could look very different than it does now...
The number of people who advocate for "we must know the exact identities of every single person who is doing moderation" is just bonkers. Being the public face of a T&S team is fucking miserable. My wallet name Google search results are still poisoned from my stint at it, which ended in 2007.
If I had it to do over again, I'd have done all my professional work, even the public-facing bits, under a pseudonym like "Sam Smith" that was ungoogleable. There's zero reason for any individual agent's name to be public, and many good reasons they shouldn't be.
The best move I ever made while I was running the LJ team was to start all my agents using persistent pseudonyms. If you were talking to Susan, it was always the same person using the name Susan, but her name wasn't Susan and she might not have even been a woman, etc. Saved so much agent harassment.
Yep. I never used my real name as a crisis counselor and I asked my supervisors not to put my face/name on social media while working in gun policy. Too many wingnuts.
Yes! And there was the advantage that if you did a case that blew up and everyone was screaming at "Susan", you could rotate your name and be Jessica for a while!
bit of a tangent, but just noticed it, but you're totally right (and it sucks that you are) but is also absolutely correct that the pseudonym should also be a male-presenting western-presenting name. Hate that it is correct, but it is.
Yeah, I landed on "Sam" because it's plausibly gender-neutral but most people read it as male. (Ditto "Alex", and I had a few others on the list that aren't immediately coming to mind.) But I think it's more the genericness (and thus non-Googleability) that's important
I’m not in that space, but I use my initials vs first name in these spaces on purpose. I don’t include pronouns for the same reason. I get much less garbage dumped on me.
I was about to yell that I'm quite googleable, actually, but even under my full name *I* rarely appear. Rightwing commentators, Supernatural actors, non-binary singers, but very seldom me. Good choice 😉
Ha! I like the thought of "put your deadname up for adoption", lol. It's pretty much academic for me by now, my wallet name is so thoroughly connected to me and all, but yeah, if I had it to do over I'd do it way differently.
The only privacy I've got on Google Search is that my name's fairly common, and the first five or so pages are about a jazz drummer who shares it (if you've heard me drum, you know that's not me.)
Between the sports player, the actor, and the alleged murderer (he was convicted but one of the key witnesses may have been lying), no one ever finds me on a basic search. I gotta say, it's not bad.
Maybe we need to turn your name into a common noun, to make it ungoogleable: any moderator is henceforth called a rehaeli [insert your real name here] & ppl list it in their job descriptions & CVs.
Or would that just compound the aggravation?
