The 2025 Social Web Trust & Safety Needs Assessment Report is now available.

@iftas @admin sure. just letting you know that if a move develops to start implementing it at bigger more general-purpose instances, there will be pushback, and for very good reason. thanks for all your work.

@iftas @admin moving toward allowlist federation will radically disempower selfhosters and everyone on the long tail of small activitypub instances. no thanks. I'll do whatever I can to oppose it whenever it comes up outside of insular "safe space" communities.

@deutrino @iftas @admin I worry about the whole thing devolving into an self hosted email type situation where a self hosted instances are more likely to get banned from interacting with "big tech" instances hosting millions of mailboxes due to a lack of traffic volume from the self hosted side.

@deutrino @admin there will undoubtedly be pushback to any structural changes, but as the network grows, so do the vectors for harm and conflict.

It's a messy network and we all only see the parts of it we're aware of. As more and more companies, people, services join, some communities will likely seek to shrink their visibility or exposure, I think this is a natural outcome.

@iftas @admin what I am saying, I think, is two things.

1. any push for allow-list federation at bigger general-purpose instances will inherently further fracture & divide a fediverse which already has divisions (mostly based on language, but also other concerns)

2. this is my online home. I've been here nearly 10 years. I'm moving more toward selfhosting with each passing month. I will very vigorously oppose any such push, and I assure you I won't be alone in that. just saying.

@deutrino @admin I can't personally imagine a world where large services go to allowlist, we've not seen them be overly blocky with their denylists, it would be very odd to see any large service move to allowlist.

In general the larger they are, the more permissive they are, I think that's a natural and probably desired result.

Federation is a privilege, not a right.

@deutrino @admin To your second point, I find it odd when people start leaning toward requiring - or wanting to require - others to hear them.

There are already large fractures, eg global north to SE Asia/Japan instances. You have every right to publish your content, and everyone else has every right to see or not have to see it.

This is how civil society, and our freedoms of expression and association work. Those freedoms include the freedom to not express and to not associate.

@iftas @admin thanks for clarifying your mindset.

@deutrino @admin here's what I think might be an illuminating factoid:

The IFTAS CARIAD observatory monitors the domain blocks put in place by the largest servers.

5,868 domains have been blocked by at least one of the observed sources.

Not a single one of those 5,868 instances is blocked by all of them.

Only eight domains are blocked by over 90% of the sources.

I think that speaks volumes about this network's ability to govern itself.

@admin thanks for the feedback!

I think as with all things global, some things will work for some, other things for others. A mix of denylist, greylist, and allowlist options would let every community pick what's best for them.

And, even with allowlist, one probable outcome would be "allow what these other servers I trust allow" and greylist everything else - the possibilities are pretty wide.

@iftas I cant deny that allowlisting would, or could, solve many problems. But I dont want to go without questioning what problems come with allowlisting. On small and large scale.

I'm not a software engineer, so I dont want to pie in the sky to a pipe dream here. But I would wonder what all those tools you mention would look like in combination, if possible. But, of course, it depends how those tools are defined going into hypothetical development.

@iftas Also interested in the cost-benefit of one or combination of those tools, once defined.

@iftas I think what I would like the most is both the projected benifits AND problems of allowlisting. Then to compare them to the cost-benifit we have in denylisting.

Then, look at which problems we can best address.