Administered by:
@rysiek@mstdn.socialOld blogpost bamp:
#BlueSky is cosplaying decentralization
https://rys.io/en/167.html
> BlueSky’s decentralization is a similar kind as with cryptocurrencies: sure, you can run your own node, but that does not give you basically any meaningful agency in the system.
> “Neutrality” and “speech” and “voice” and “protection from bans” is mentioned right there, front and center, in BlueSky’s overview and FAQ. At the same time moderation and anti-harassment features are, at best, an afterthought.
I wonder if BlueSky finally fixed the "DID: Placeholder" thing. 
(I would be surprised)
@rysiek - BlueSky is decentralizing the costs of running the business and centralizing the profits.
and, what about distributed protocols like scuttlebutt, cabal, ipfs or even bittorrent (when using dht/magnet)? are those not truly decentralized?
to be clear i agree that bluesky appears to have ulterior motives, especially with their "we don't know the absolute best way to do moderation, so atm the responsible thing to do is to not provide a block button". however i don't think that intentional hierarchy-building is the solution.
@ww I used to believe that federated systems like fedi are only a stop-gap solution until fully peer-to-peer systems emerge, and that fully p2p systems are what is needed.
I no longer believe that, especially after the whole blockchain/cryptocurrency kerfuffle. The p2p approach is a very libertarian one, and it works very well for individuals who have the capacity and resources to manage moderation etc on their own.
Most people are not like that. Most need a community to help out with that.
@ww instances (especially smaller ones) are a decent way to model communities. And being a part of a cozy, well-moderated community means that you can mostly avoid most of the bad stuff potentially flying your way — *somebody* will have already reported it, and moderators would have blocked it.
Every now and then you will need to report something yourself, but then everyone else in your community benefits as well.
Small community means your voice is loud enough, too, to make a difference.
@ww so, instances having agency means that communities have agency. They can defend themselves and their members.
If a member of a community strongly disagrees with some community decisions, they can move to a different one.
So, the individual gets *some* agency (at some cost — cost of moving, or putting the effort into influencing the community decisions), but in return they get the protective umbrella of the community.
And since there are different communities, there can be different norms.
@ww in BlueSky, node operators do not really even have the freedom of association, if I am reading the protocol right. Unless they enforce it out-of-band, via say firewalling rules.
BlueSky is talking big about decentralization, but the level it is decentralized on — nodes — is meaningless. All the magic happens in the "reach" layer, and that layer can and will be easily centralized, in a very similar way Google kinda sorta centralized (everything goes through Google Search) the Web.
however i think that you can achieve the same user experience community-hosted nodes provide regardless of your network topology by making all blocks and mutes public and subscribing to them. this creates a web of trust where anyone can be a moderator for their community, regardless of the server they're hosted on.
it's the same thing fedi does, but without the limitations that were seemingly accepted as the only way to implement it technically. and imo the idea of community moderation as sold by mastodon is somewhat dishonest. it isn't how most fedizens experience fedi. most instances aren't tightly-knit communities, and most people on here have probably never talked to their admins.
there're other issues with allowing anyone to become a moderator, and i do wonder if there's a better way altogether. but requiring technical expertise is definitely not a solution.
@ww moderators do not need technical expertise. Admins do. These are separate roles on a lot of instances.
Tying a community to a server means tying it to a domain name, IP address, and some infrastructure that costs money. Which means that there is a cost to burning it down.
Tying a user to a community (that is tied to a server) makes moderation… possible. Otherwise there is no cost to setting up sock-puppet accounts and making everyone play whack-a-mole.
@ww the result is that instance admins and moderators have a strong incentive to make sure people on their instance do not misbehave, and that spam and sock-puppet accounts are under control. Otherwise others will defederate, which will "burn" their domain, IP addresses, etc.
This in turn makes it harder for malicious folks to create loads of *effective* spam and sock-puppet accounts. Instances that allow this kind of behaviour get defederated, quickly becoming useless for their purposes.
dunno, i don't see it ever getting big enough to the point where you need google resources to index the whole thing. you're also saying that the chronological timeline, one of the "killer features" of mastodon, is an inherently uncompetitive approach. but if so, shouldn't we try to find a better option?
they are going to have an unfair advantage by being the default. i could see them making it as hard as possible to switch to a different implementation. but being the first to index the posts they have doesn't seem to be that important in comparison.
@rysiek Very clear. I remember reading an outline of the architecture soon after Elon Basin Day and thinking, at the part about indexing servers, "okay, that's not federation."
@rysiek in EVERY single case to date, moderation is either treated as a de minimus bolt-on (and usually handicapped stupidly, like in BlueSky), or initially boot-strapped well and left to flounder thereafter because it 'interferes' with the #monetization/#enshittification dyadic modality (e.g. Tumblr, DeviantArt).
It's just enervating.
@zeruch agreed. That's why I am a fedi person. Moderation here is definitely uneven between instances, no question about it, but instances do have the agency to manage their environment.
@rysiek I would never expect them to *not* be uneven, by the nature that there can only (at best) be instance by instance quality thresholds of content moderation, and even that will be limited by resources coupled with community adherence.
The idea of clear, discrete, consistent, policy-driven content moderation at scale requires immense staffing that is trained/equipped properly, and theres no 'model' in the current SV fluff-fest for that.
@zeruch I'd go further and say that the idea of a single policy that could be implemented in a clear, discrete, consistent way for hundreds of millions of people using a given social network is preposterous.
People have different expectations, come from different cultures, adhere to different norms, have had different experiences. There is simply no way to capture this in a single, coherent policy — which is exactly what SV-driven social media walled gardens are trying, and failing, to do.
@rysiek "'Id go further and say that the idea of a single policy that could be implemented in a clear, discrete, consistent way across hundreds of millions of people using a given social network is preposterous."
I don't entirely agree in the sense that just like real life we can generally globally agree that rape and murder are bad, but yes, your point otherwise stands well.
A key Fedi advantage is granularity of experience anyway.
/1
@zeruch I guess we vehemently agree. 
> just like real life we can generally globally agree that rape and murder are bad
The set of things we can globally agree on is so very limited that it is basically unusable as a "policy". A network that enforced only that set would be a cesspool, unusable for most people.
> A key Fedi advantage is granularity of experience anyway.
Which, ironically, is often mentioned as a bug.
@rysiek @zeruch at the risk of going off topic, on the universality of morality, consider that even rape and murder being bad isn't something universally agreed upon. At a minimum, there isn't even an universally agreed upon definition of what rape and murder are (see e.g. varying definitions of the age of consent, how far can self-defense be taken, or considerations on the death penalty).
Still, the granularity of experience of the Fedi *is* a problem, but IMO not in the way usually meant.
@rysiek @zeruch to clarify, the problem of the granularity of the Fediverse isn't that the experience and moderation are community-specific, but in the limitations of the management of the boundaries (re: inability to carry content over when migrating, impossibility to migrate cleanly across a fediblock, no possibility for individual override of admin decisions concerning federation, etc).
@rysiek ""SV-driven social media walled gardens are trying, and failing, to do.""
Yes, but that failure is driven not from trying to affix a clear policy up from on high (although thats the appearance they try to pitch) but from a total failure to actually administer ANY policy consistently due to resource and training/tooling limits. They seek "efficiencies" through a kind of Huxley-an Hucksterism that is patently inept and deeply self-defeating.
They just don't want to admit it yet.
@zeruch I might not fully agree on VLOPs not trying to enforce a central policy — nymwars, feeding mothers, nipplewars all showed they *do* have a central policy that is at least opportunistically enforced.
But even if your take is correct here (and it very well may be), my point is that even if they *did* try to enforce such a central policy, and even if they were *effective* at it, it would end in tears.
@rysiek thanks for reposting. Very well said!
@rysiek From the same people who gave us twitter.
@Smrki can't go wrong with that!