Sven Slootweg (soft-deprecated) @joepie91@pixie.town

poly people 🤝 businesses
"our partners"

If you have seen people adding content warnings to #AIart its because sometimes the subjects can be distressingly uncanny or represent some type of body horror with misunderstood anatomy. #TheMoreYouKnow #feditips

Note to white hacker culture, we can't keep brushing racism under the "this was built by Communist trans furries" rug. We aren't beyond question just because we are left wing. That's precisely the sort of attitude that has domestic abusing union leaders running a lot of the Brit left.

Radical politics is a community of practice not a moral position you assume via some identity signifiers.

As a trans person I'm ashamed of the way transness (white transness) is being drawn on to pinkwash.

i guess this explains why all of infosec twitter (which i am not a part of) suddenly started following me on fedi all at once

handing around exports of your follows for people to just import certainly is an interesting idea especially if you follow locked accounts (like mine) and i am not sure if literally any thoughts about ethics and consent went into that idea (probably none)

thanks for completely flooding my follow request queue i guess and i hope you enjoy your gay shitposts filled timeline

"photo id verification for social media" is an idea that should never be parroted uncritically

i've seen it going around, mostly from one user who came here in the last ten days, that #fediblock is "not to be used for individual users" and that it is "only for admins"

don't listen to that person, they are *wrong.* #fediblock was started by @CaribenxMarciaX as a way of letting communities protect themselves from bad actors and instances.

i repeat: #fediblock is EXPLICITLY open to everyone to report red flag users AND instances, and it has been for YEARS

see this is why, as someone perpetually broke and historically poor, i don't like listening to rich people for budgeting advice. musk spent 44 billion dollars to get humiliated in public but everyone knows that there are people who can do that for like £2000

Yesterday I posted about worries re. moderation at mastodon.social/.online:

https://www.patreon.com/posts/on-our-policy-74436103

And asked the community to feed back about whether they'd like to do nothing, 'silence', or 'defederate'. The response was almost entirely in favour of silencing.

Today, especially after noticing that those two instances DO federate with an instance that hosts fascist content, we went ahead with that silence:

https://www.patreon.com/posts/74474015

Thank you, .art, for supporting a solid no-nazi stance. <3

why do we have a system for advertising instances unworthy of federating with but not one for advertising the good instances?

on a scale of 0 to 10 where 0 is fully blocked and 10 is fully federating, it sucks that our current system is to make every instance a 10 automatically and then only with manual admin action turn that to a 1 or a 3

turning it to a 0 requires direct firewall access because even with authenticated fetch mastodon still sometimes voluntarily sends information to instances you have blocked

some instances (this one included) have public blocklists, which is a nice way to gauge an instance's morals, but a very bad way to actually deal with instances worthy of being blocked because doing so would require admins to constantly re-read the block lists of instances they trust and play "spot the difference"

and if your software is creating work rather than making work easier, you've created a bad game rather than a good tool

so here's my multi-part proposal:

don't instantly fully trust new instances. it's too easy for a bad actor to simply switch domain names and become fully trusted again. we've seen it hundreds of times.
allow automatic sharing of instance trust information. basically, an admin would enter information they know about another instance into a box, along with evidence and a verdict on what level of federation should be done, and that information would be immediately shared to other instances that have established mutual trust.
trust for federation information could be anything from "ignore" to "notify admins" to "automatically apply", on a per-domain basis.
establish a better entry procedure into federation. why can't we have a setting to notify admins about new instances so they can proactively vet them?
and finally, as long as we're giving admins more control over federation, why don't we also extend that to reports? the current system where a user can report a remote post and has to choose at the time of the report whether to send it to the remote instance sucks. that should be an admin decision, and reports should be able to be forwarded to any instance, not just the reported one.

of course we'll never get any of this until website boy stops running the show, because twitter doesn't have any federation-based moderation features either

#mastodev