[verdverm]

ATProto actually separates moderation from PDS or App View. Users can choose which labellers they prefer and can even combine them, separate from where they host their data or the UI they choose to use.

https://bsky.social/about/blog/03-12-2024-stackable-moderati...

They do the same for feeds, 4 core components, with user choice and interoperability for each

[jazzyjackson]

So what levers does that give governments seeking to get compliance out of an internet service - if multiple apps are hosting anti-party propaganda the government has to block the domains of each app ?

Or, perhaps the domains of the content itself is blocked so apps continue to work but fail to load content within certain borders ?

[numpad0]

Bluesky's whole moderation and decentralization setups were only devised, or at least implemented, after it blew up in Japan and bunch of Japanese artists immediately started hammering the platform with novel content they consider to be lawful and more kosher than normal but were officially felt platform threatening to their team.

So, not to undermine efforts from Bluesky team - I applaud their SoTA attempt at microblogging architecture and platform so far - but Bluesky definitely has not solved the messy question of legality, ethics, and speech, at theoretical levels. Only hypothetical and/or operational.

[jazzyjackson]

I can only imagine what this refers to but after googling for a few minutes for "bluesky japan controversy" I'm just going to let this exchange color my impression of what's going on at bluesky:

> Deleted Post

>> Katie Tightpussy: BLACKTHORNE: We are a moderation service for Bluesky with the goal of improving social media for progressive queer folks and leftists who wholeheartedly enjoy Japanese anime, manga, games, hentai, fan art, and doujin.

MARIKO: The Anjin incorrectly believes that he is an expert on Japanese culture.

>>> Sign in Required

>>> Sign in Required

>>> Yep, the "controversial fiction" thing is such a sad (yet hilarous) cop-out. They KNOW it's wrong so they employ this linguistic obfuscation to try and make people think they aren't pedos.

https://bsky.app/profile/did:plc:hslv64eax7d2lwrm7qtg44ud/po...

[numpad0]

yeaaaahhh, it can't get more appropriate than to frame this problem with an image of an angry short ethnic woman in front of a tall white male guardian, right... and it's totally fine that no one in said ethnicity interacts with that post over there, right...? At that point you might as well include topics like mercury content in Asian seaweeds and arsenic in rice... duh.

The reason why you aren't finding anything specific to Bluesky is because it's not a Bluesky specific problem. Every social media that goes big in Japan will have this Japanese pedo flood problem, if you prefer it expressed in that kind of vocabularies. Social media that do not experience this stays irrelevant in Japan, for better or worse(frankly said likely better for profitability).

It happens as a spontaneous flood of 50:50 mix tangentially labeled pedo:nonpedo mixed content stream consuming non-negligible bandwidth, increasing in volume exponentially until Japanese fraction reaches steady state of >50% by content, ~30% by user count, and >50% of top popular accounts. The mixture and fraction metrics show indefinite steady up trend.

Gargron, the Mastodon author and benevolent dictator of the Fediverse, famously gave up and went on to basically race filter Japanese from the European half of the system, which by the way I have no choice but to fully respect given his circumstances, options available, and value to be recovered. Twitter famously deleted trust and safety team, and according to Elon Musk himself with his tongue in cheek, Twitter usage in Japan is "growing", amid its worsening Indo-Arabic spam problem and tanking global popularity. Even literal pornography websites like PornHub had this exact problem, in whose case they were forced to nuke the website to get rid of so-called JAVs using unverified CP as an excuse(lots of JAVs feature easily CP frameable females). And Bluesky created the whole moderation framework and default enforced implementation in response to it.

Anyway, what I'm saying is just, only, Bluesky's whole moderation framework is a post hoc solution to this problem, so while strong resistance against oppressive evil radical totalitarian governments sure is considered as one of ultimate goals, it's definitely not the goal in their initial problem definition.

[pfraze]

Your broader observations match what I generally know, but the moderation system wasn’t created as a reaction to Japan or any other specific set of circumstances we were facing. It was a system we had been developing since before launch and was designed to resolve the tensions of different perspectives in what’s acceptable

[stale2002]

The answer is that the government won't have many levels to pull to censor content on the internet.

I think that people are are celebrating the ban of X, and moving to decentralized platforms, forgot that the whole point of decentralization is to make censorship difficult.

When you move to bluesky, you just support an even more free version of twitter.

[verdverm]

Freer from both govt and corporate controls

Feed (algo) and moderation services are choices at the individual user level

[verdverm]

apps don't host data, PDS (personal data servers) do

apps implement views on that data, and may or may not follow the rules, of gov't or users. For example, even though you can block a user, detach a quote post, or hide comments, apps have to implement this behavior, and nothing stops a person from finding that relation.

ATProto is federated, not centralized, and not something gov'ts and regulations have thought thoroughly about. Also, with DID, I believe DNS blocking will be hard because I can change the name and still get to the same content

[jazzyjackson]

> not something gov'ts and regulations have thought thoroughly about

Not trying to be combative but I find this mode of thinking is likely to backfire, governments don't have to think they can just act - has bsky thought seriously about what their response will be to the same laws that X is suffering from?

[verdverm]

Users have choice over moderation, it's not necessarily something Bluesky can limit, by design.

It's also worth seeing how it plays out with others, while you are still not on the radar. Part of the reason Xitter is getting harsher treatment is because Musk antagonized. That's not the best way to negotiate, especially since he said he'd abide by local rules, like how he said he'd be better for free speech.

[AlienRobot]

I'm not sure I understand. Are you trying to say if a judge says "you must stop displaying this content" Bluesky will argue "we can't do that"?

I think Bluesky would get banned.

[verdverm]

Bluesky could implement it in their moderation service, but that does not mean users in Brazil would be impacted by it.

Users could swap moderation service or swap interfaces

The one thing that could happen is the PDS deleting the record for everyone, everywhere

One thing to separate is Bluesky from ATProto. Bluesky is the default implementation of the 4 core pieces, but one could use alternatives for all of them as well and still have their content show up in the bsky app. Imagine if Twitter was open source and federated