It NEVER fails

⸸ commander ░ nova ⸸ :~$

@KimCrayton1 I’m curious what kind of moderation policies dair-community.social has, because this shouldn’t be happening. Any good instance should have the racist/bigoted instances blocked and suspended

⸸ commander ░ nova ⸸ :~$

@KimCrayton1 I looked at your instance and it appears to have a very bare bones moderation policy and no suspension list to speak of (unless it’s private). I don’t know if this is the main reason these things are happening but it’s probably part of it (there are quite a few nazi instances that are routinely evading blocks and it requires active administration to keep under control)

Charles U. Farley

@cmdr_nova I saw this explained in another thread. They're not coming from "racist instances". People just keep signing up for new accounts on big instances to evade suspension, so its a game of whackamole/death of a thousand cuts.

@KimCrayton1

Can-crisociality 🦀〰️🥫

@cmdr_nova This is not on DAIR (Kim's instance) at all. As @freakazoid said the posts are coming from shitty people on instances that are not widely known to be a problem, and an example Kim gives illustrates that: https://dair-community.social/@KimCrayton1/112853496319229676
It's just people being racist and shitty.

Solarbird :flag_cascadia:

@inquiline @freakazoid @cmdr_nova I'm kind of wondering if this is partly fallout from the _lack_ of algorithm.

rdgit did a NotAllMen/NotAllWhites, and it is 4000% not Kim Crayton's job to educate him on that shit.

But another person stepped in and talked to him about what he was doing and he ended up posting an apology a couple of hours later.

TO BE VERY CLEAR: I do NOT imply Kim shouldn't've blocked him, NOR do I think she should've done that work. I am NOT saying that.

But if you look at his posts in general, he seems more ignorant than ill-intended, which makes me wonder if he'd ever have found Crayton with an algorithm filling his timeline and leading him to new accounts to follow.

I have a point, bear with me (1/2)

Solarbird :flag_cascadia:

@inquiline @freakazoid @cmdr_nova The algorithms tend to point you at people either politically opposite you (to hate), or politically & socially very closely aligned with you (to like), which means he'd likely've been herded along with other older white men who have his same blind spots.

Again, THIS IS NOT KIM'S PROBLEM. But I do think it is _kind of_ a problem, because you _don't_ get the allied-but-not-the-same interactions that _can_ educate.

The problem with _that_ is that it puts the burden on BIPOC, queers (hi), religious minorities (hi), etc. And that's intolerable eventually because _no one_ has the infinite patience required for that.

I don't have a answer for this problem. But I think it might be worth thinking about.

(fin)

Charles U. Farley

@moira @inquiline @cmdr_nova This would be a lot less of a problem with a "limit replies" feature. It would also be less of a problem if replies to a thread were visible to everyone who can see the thread. Right now, if someone replies to a post, it will only be visible on instances where people already follow that person and to people who were mentioned. If someone replies to *that* reply, it may become visible if someone clicks on the reply, but they may not see the reply to the reply.

This all boils down to the simple fact that replies should not be treated like regular posts that just happen to have a "references" field. This is not email or Usenet. The OP should "own" replies. They should be able to limit them and delete them, and they should be distributed with the original post, to guarantee that everyone who can see the original post can see replies.

It might even be a good idea for replies to others' posts to ignore the reader's block and mute settings, so that harassment of others isn't invisible to you, but maybe that should be a setting itself.

Jon

Yep. And GoToSocial is implementing this! Also I think discussion forums that are adding ActivityPub support are as well -- I saw a post from @[email protected] with some interesting implementation thoughts recently. But in a discussion a few weeks ago, Renaud said it's not going to happen on Mastodon in the short term.

A couple of other things that would help deal with the "shitty replies from new isntances" problem:

- a new "bubble" visibility level, limited to specific instances (broader than local, but narrower than public/unlisted)

- making federation opt-in, or controllable by policy, as opposed to today's "federate by default with Nazis and everybody else". Of course that's got complications as well, but from a security perspective it's really hard for me to see how to make things work long-term without that.

And @[email protected] great point about how not relying on algorithms here changes the tradeoffs. Twitter also filters out "low-priority" replies (or whatever they call them), which on the one hand is problematic but on the other hand does limit some of the annoyance of these.

@[email protected] @[email protected] @[email protected]

Charles U. Farley

@jdp23 I suspect that if the Fediverse is to thrive long-term and not be replaced by something built on Veilid (or just die out in favor of silos), Mastodon will end up being replaced as the premier Fediverse server. The current devs just don't have the right set of priorities.

@moira @inquiline @cmdr_nova @julian

Jon

Yeah, IMHO, there are three ways Mastodon needs to be decentered:

- Decentering microblogging withing the fediverse as a whole. This is happening!

- Decentering Mastodon and its forks within microbloging. This is also happening but not necessarily in a healthy way -- threads and flipboard aren't Mastodon but take the fediverse in a very corporate direction. More positively though GoToSocial and Letterbook are very encouraging here as well.

- Decentering Mastodon relative to forks. I talked about this at great length in https://privacy.thenexus.today/mastodon-hard-fork/ Reply-limiting wouldn't be easier in a fork; but, a fork that's focused on safety would prioritize it, so it would be more likely to happen.

@[email protected] @[email protected] @[email protected] @[email protected] @[email protected]

Can-crisociality 🦀〰️🥫

@jdp23 @moira @freakazoid @cmdr_nova @julian Adding @hipsterelectron as we were chatting about this some last night

d@nny "disc@" mc²

@freakazoid @jdp23 @moira @inquiline @cmdr_nova @julian how does veilid address harassment? it's a very neat but also nascent project, and i thought anonymity was one of their priorities: this absolutely can be ensured while also providing safety against harassment, but it would require breaking some new ground and i didn't realize they were that far ahead yet

Charles U. Farley

@hipsterelectron Veilid is just a peer to peer protocol. Any app built on top of it would have to address these things.

I don't bring it up because I think it's anywhere near being able to provide a replacement yet, but because if it accomplishes its goals a lot of the people I've been hearing talk about about ways to improve the Fediverse are probably going to move over and start building on it.

@jdp23 @moira @inquiline @cmdr_nova @julian

d@nny "disc@" mc²

@jdp23 @moira @freakazoid @inquiline @cmdr_nova @julian is the decentering of microblogging related to user safety re harassment, or is that a separate goal which you're framing as a prerequisite to decentering mastodon-the-codebase's power to enact social policy by developer fiat?

d@nny "disc@" mc²

@jdp23 @moira @freakazoid @inquiline @cmdr_nova @julian there are forks focused on safety via moderation but i'm not going to give the name of the most well known one here because (1) i believe they haven't made much progress on that last i checked (as told to me by the dev who was working on it a few months ago) (2) i do not believe the admin-maintainer of that fork and its flagship instance exercises good judgement on moderation decisions for reasons i will not elaborate on here. maybe my personal experience with that admin colors my judgement, but imo it may also indicate that feature development of moderation tooling should be proposed/discussed in a more open context to avoid again setting social policy by developer fiat.

Charles U. Farley

@hipsterelectron Open discussion would be good, but having that discussion not get dominated by the loudest, and ensuring it produces policies that actually work for real people in practice, would require a level of leadership/governance which does not currently exist in the Fediverse. And then people have to actually write the code. This is why "developer fiat" has been the norm in the open source world.

I suspect it'll end up working out the other way around, where some developer who happens to have some good ideas and care about the right things puts something out there, and a community forms around it. Then that community adopts some more open/democratic method for deciding changes moving forward.

This is pretty much how Mastodon got started in the first place: Eugen had better ideas than the GnuSocial etc folks, so his thing ended up becoming the norm. But "better" was a pretty low bar, and while it's been good enough to get us this far, it's still resulted in a system where Kim Crayton gets called the n-word more times in one day than on any other platform, and is the only platform where she sees racist comments on her posts.

@jdp23 @moira @inquiline @cmdr_nova @julian

Jon

Decentering microblogging isn't specifically safety-related, it was in response to @[email protected]'s framing of what the fediverse needs to do to thrive long term. Lemmy's a case study in how a non-microblogging platform can have worse moderation and safety than Mastodon! (And PieFed is the link aggregation space's analogue of GoToSocial: more attention to safety, but at a much earlier stage so with a much smaller installed base.)

And yeah there are various existing forks that are better from a safety perspective than Mastodon (including having local-only posts), but almost all the resources still go to mainline Mastodon. Agreed that a culture change in how development gets done is needed, as I say in that article

"a more diverse project with a community-oriented focus and more consensus-oriented decision making (as opposed to the BDFL style) also offers significant opportunities for improvement."

@[email protected] @[email protected] @[email protected] @[email protected] @[email protected]

Amber

@[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] @[email protected] Does this admin have a w in their name anywhere by chance or am I thinking of the wrong admin

d@nny "disc@" mc²

@freakazoid @jdp23 @moira @inquiline @cmdr_nova @julian i'm aware that developer fiat (a term i adapted from @ariezra's book "industry unbound"; he said "engineering fiat") is the norm in this space, because most open source software is solving technical and not social problems, and because tech is a highly monopolistic global industry which inculcates this exact winner-take-all horserace mentality in the engineers who contribute to it. i'm proposing that one way to avoid this outcome yet again is to explicitly decouple protocol development from the whims of individual implementors and directly include users as stakeholders, just as you said. i'm not proposing trying to convince the main mastodon codebase to accept this decoupled protocol development process! i'm saying that this needs to be built in from the start to the governance processes of whichever codebase wants to actually serve this community, instead of just using the fediverse community to advance its own personal goals as others have.

d@nny "disc@" mc²

@jdp23 @moira @freakazoid @inquiline @cmdr_nova @julian regarding the tech debt of the mastodon codebase itself: rather than a hard fork, a complete reimplementation (which itself implements an explicit spec derived from activitypub and negotiated with users-as-stakeholders) seems like a great way to achieve multiple goals at once. i believe there are already mastodon-compatible codebases not written in ruby on rails (not sure of this), and these would be very good to get on board with the protocol negotiation process or even potentially use as a basis instead. but my impression (i may be wrong here) is that nobody has yet established such a formal spec (derived from activitypub), or an appropriate governance process for it, because "governance" and "the spec" have been effectively defined by "getting mastodon dot social to pull in your changes". any previous work in this regard would be really good to build on as long as the people doing that work accept the terms of communal specification negotiation.