Email or username:

Password:

Forgot your password?
Top-level
Erin Kissane

@anarchopunk_girl I think the WaPo article was poor, but I think David and Renée found some obvious weaknesses in the system and had a lot of excellent advice for fedi admins. The Observatory gets shit from all sides for basically everything they do, including their research on centralized networks.

If you want to get into methodological specifics, that's fine, but I haven't seen anything that especially concerned me to date.

4 comments
novatorine 🏴🏳️‍⚧️

@kissane I'll be upfront about the fact that I haven't read the report bc I don't have the energy, but from what I've heard, the vast majority of the servers they found the CSAM stuff on are already defederated by everyone, they're on the Tier 0 blocklist that instance admins use basically verbatim to get started. And there's nothing anyone can really do beyond that, because Mastodon is FOSS. We can't actually stop horrible people from making horrible instances. So I don't really know what the point is of saying that you can find a small number of instances that are mostly walled off from everyone else that have CSAM on them. It's like complaining about the internet as a whole bc some people are hosting bad shit on it.

@kissane I'll be upfront about the fact that I haven't read the report bc I don't have the energy, but from what I've heard, the vast majority of the servers they found the CSAM stuff on are already defederated by everyone, they're on the Tier 0 blocklist that instance admins use basically verbatim to get started. And there's nothing anyone can really do beyond that, because Mastodon is FOSS. We can't actually stop horrible people from making horrible instances. So I don't really know what the point...

Erin Kissane

@anarchopunk_girl Most of the CSAM is already walled off, but quite a bit wasn’t, even in their tiny sample, and they had pragmatic technical suggestions for strengthening protections. Especially with fedi’s rapid growth, that seems crucial to me. I think the report itself, which I’ve read, has been misrepresented here bc the press was uninformed and just bad, but a lot of good admins are paying attention.

novatorine 🏴🏳️‍⚧️

@kissane okay I read the methodology section, and yeah my fundamental problem with their methodology still stands. They just took a list of the top 25 servers by population and scanned them for CSAM content, but that doesn't take into account the fact that that's not really representative of what people actually joining the network will be able to see or the network as a whole, because the top 25 servers could include bad servers by dint of bots or there just being lots of bad people, and they could be completely defederated from the rest of the network, but it would still show up as that bad stuff supposedly being "on the network," so I don't think it's really representative of the actual network itself or experience of using it or danger to people on it.

@kissane okay I read the methodology section, and yeah my fundamental problem with their methodology still stands. They just took a list of the top 25 servers by population and scanned them for CSAM content, but that doesn't take into account the fact that that's not really representative of what people actually joining the network will be able to see or the network as a whole, because the top 25 servers could include bad servers by dint of bots or there just being lots of bad people, and they could...

Erin Kissane

@anarchopunk_girl I…don’t think it’s perfect, but I also think it’s more usefully considered as a free diagnostics pass and recs set from the pros than as an unfair gotcha. Any CSAM is too much, obviously, so patching holes seems to me like an important priority for admins and devs.

Go Up