So, what would “Better Safety Tooling” on the Fediverse look like?
And, what would meaningfully help the Fediverse to serve more people?
—
Tagging @laurenshof because you had some interesting insights I’d LOVE for you to expand on :)
Discussion
So, what would “Better Safety Tooling” on the Fediverse look like?
And, what would meaningfully help the Fediverse to serve more people?
—
Tagging @laurenshof because you had some interesting insights I’d LOVE for you to expand on :)
1. Shared blocklists, at the user and server level, configured by default in Fediverse server software
2. Content filter API servers, like Akismet
3. Fix followers-only on Mastodon
4. Private groups
5. E2EE messaging
6. CC licenses for content
We need to make public participation more explicit, and make it easier to participate privately.
@evan Blocklists have a flaw, and need to improve at all costs.
@benpate @evan @laurenshof If I understand correctly what you mean by "trust relationships", this relates strongly to something I blogged about lately:
All popular platforms I know of in the fediverse offer connection-roles in more or less the same way as X and Instagram do. The role-system adopted by LinkedIn and Facebook offers higher trust relations on top of the simple "follow" model: "connections" and "friends", respectively. (I describe this in detail in the post I link below) I believe this is an important distinction, which may be the reason why Instagram and X are arguably more toxic than Facebook and LinkedIn.
https://blog.gillesdutilh.com/2026/06/mutuals-are-not-friends-2/
@benpate @evan @laurenshof If I understand correctly what you mean by "trust relationships", this relates strongly to something I blogged about lately:
All popular platforms I know of in the fediverse offer connection-roles in more or less the same way as X and Instagram do. The role-system adopted by LinkedIn and Facebook offers higher trust relations on top of the simple "follow" model: "connections" and "friends", respectively. (I describe this in detail in the post I link below) I believe this is an important distinction, which may be the reason why Instagram and X are arguably more toxic than Facebook and LinkedIn.
https://blog.gillesdutilh.com/2026/06/mutuals-are-not-friends-2/
I've found @DePemig 's work very helpful for me in thinking about social trust.
"Mutuals are not friends" is a key insight that a "Follow" is a helpful signal, but not as strong as we often make it out to be.
This is why I'm hopeful we can make progress on "Endorse". It *is* a bit more fiddly, another command for users to understand so I do worry about making things more complex. But it is a significantly stronger signal, an explicit trust signal (vs interest signal that Follow indicates)
@benpate @evan @laurenshof
@benpate @laurenshof I'd also say: more implementations of 'indexable' and 'discoverable'.
I love this. Yes.
And, a generalized solution, if that’s even possible.
For consent on Bandwagon, I had to roll my own. When you publish an album, the last step is to check on/off any partners who are allowed to rebroadcast your stuff.
There’s a “select all” to be included when new partners are added. This is super important, because new things come along (like Bridgy.fed) that don’t match our pre-existing models.
It sucks to keep adding more Boolean flags to the model.
@benpate i think "better safety tooling" is a bit of a red herring, because the real problem is in figuring out who has the authority to set boundaries.
assume the current fediverse/mastodonishverse/whatever to be a "bounded space". exact boundaries and technical details dont matter yet. it is a shared "space", that has some boundaries and values that people want enforced within those boundaries.
(examples can be "no cloudflare" or "relays should be optin", or whatever)
/1
so now you want to improve this "bounded space" with some rule/tool/thing, that makes the space "safer" (for any definition of safer)
the core problem is that the functionality of this Thing are contested by its nature, but that there is no source of authority who can make the decision for the entire space (decentralisation yadiyada)
so any rule/tool/thing that matters cannot get community buy-in because it is contested and nobody can dissolve the contestation
/2
@laurenshof I love this line of thought. Thank you!
So far, I've naively assumed two boundaries: server (chosen by admins/mods) and user (my own personal blocklists)
But that doesn't really follow the way our natural communities cross-cut the Fediverse, does it?
@benpate yeah, my most radical opinion is that i dont think servers actually exist, in the sense that they dont actually meaningfully play a part in either how we understand the socially constructed object 'fediverse', nor in how this "fediverse" again shapes our social understanding (via the posts entering our feeds through which we create our social perception of the fediverse)
@benpate calling cloudflare fascist is very easy. figuring out a rule/thing/object that causes 'no cloudflare' in this bounded space is hard, and implementing that rule/thing/object is literally impossible
thats why were stuck in this endless cycle of discourse and yelling, because there is no structural way to break out of it
Cc: @oli I suspect you have thoughts on this. I hope you don't mind me tagging you in.
Thanks! And yes. FIRES looks like a fantastic part of the solution :)