Yep, I’ve been particularly irked with the way people keep calling Bluesky “decentralized”. Simply having multiple servers doesn’t make it decentralized, that’s like saying YouTube is decentralized because they have dozens of CDN centers.
If every endpoint on the network is owned and controlled by a single entity that manages white/blacklists, then it is not decentralized. I’m not sure what you would call it, maybe “pluralized”? Not sure, but definitely not “decentralized”. A decentralized network will have no such authority that controls who can and cannot join it.
It’s federated in name only. It would cost millions in storage and networking to set up another Bluesky node because each node is essentially a full mirror (not that you couldn’t code custom filters or something I’d guess?).
No one is ever going to bother unless conservatives want to set up their own server where they can’t be banned for trolling.
And you could ban it. And I could create [email protected] (or however it’s formatted).
And then this cat and mouse could happen, all day every day. Just a series of me creating accounts, you blocking them, but it then becoming a full time job making sure I’m currently banned, and me evading the system.
The thing is, when that happens, the mods/admins of the trolls new instance ban them there, without the original admin having to do anything.
Abandoned instances, or instances that simply don’t moderate, get defederated, so it’s quite manageable.
The only case where it’s not that clear cut, is where the troll is a big issue for one admin, but not another. Take trans issues for example. I have a zero tolerance policy on transphobia on this instance, but not all lemmy admins are as aware of the dogwhistles as I am, so I will block some users that other admins won’t. It’s not ideal, but it’s manageable, because I can stop their brand of transphobia from reaching my instance even if they’re not banned by the remote admin. And if that pushes the troll to create another account elsewhere to get through the instance ban, then that becomes harassment, and the other admins will act, even if they wouldn’t before.
Nah, you’re not missing anything, and that is already happening. There’s a few users who keep creating new accounts on alternate instances to post troll threads. But it’s on a small-enough scale that it’s trivial for mods/admins to deal with. And realistically, it’s more work for the troll than it is for the mods, so eventually the troll user will either tire themselves out or simply run out of instances to make new accounts from.
It’s definitely an intrinsic problem that ActivityPub doesn’t really have a reasonable solution for yet. It’s sort of a necessary evil in order for the rest of the Fediverse to function freely, and part of the compromise that users and communities need to accept and live with in order to keep things running. But, I don’t think it’ll become a major issue unless the Fediverse sees a massive influx of users, like the sort of scale that Bluesky saw with Twitter’s exodus.
I wanna add to what other users already answered that this problem is not created by federation, only exacerbated.
If I’m mod of a community and I ban your [email protected] account, I cannot stop you from creating, e.g. [email protected] and coming back. Most servers have some barriers against spam account creation in place, but I’d wager you could easily create a handful of accounts on a server until they start to grip.
Even completely centralized platforms such as Twitter and Reddit are the same. You can easily ban/block evade a couple times per timeframe.
This isn’t entirely accurate. Main issue is that this is all very new and moving very fast, so while it’s not fair to call atprotocol completely open it’s not fair to call it centralized either.
Right now you can run a relay independently and you can scrape bluesky’s PDS set with it (guy did that in July and it cost about 150 bucks a month, probably more expensive now. https://whtwnd.com/bnewbold.net/3kwzl7tye6u2y) there’s just not much incentive to do so. This recent policing of accounts might change the calculus on that thought process, but for now it’s not really that you “can’t” it just nobody really wants to. (Additionally there’s only like one client that lets you set multiple relays and/or appviews, so the front-ends haven’t caught up with this idea either.)
They are VC-Funded, they will inevitably screw over everyone!
I understand where this notion might come from, though, ATproto has already matured enough that it could be used without BlueSky PBC existing.
It could die tomorrow, and with little effort the ~650 non-bsky PDSs and a relay (which would have to handle much less data at that point) could be ran by volunteers and interested people.
If it were decentralised, it wouldn’t be possible for the platform to set a network wide policy on parody accounts…
Yep, I’ve been particularly irked with the way people keep calling Bluesky “decentralized”. Simply having multiple servers doesn’t make it decentralized, that’s like saying YouTube is decentralized because they have dozens of CDN centers.
If every endpoint on the network is owned and controlled by a single entity that manages white/blacklists, then it is not decentralized. I’m not sure what you would call it, maybe “pluralized”? Not sure, but definitely not “decentralized”. A decentralized network will have no such authority that controls who can and cannot join it.
Distributed is the term, and to be more precise - it’s not federated because ownership and control is not distributed, even if the servers are
It’s federated in name only. It would cost millions in storage and networking to set up another Bluesky node because each node is essentially a full mirror (not that you couldn’t code custom filters or something I’d guess?).
No one is ever going to bother unless conservatives want to set up their own server where they can’t be banned for trolling.
they couldnt anyway, because bluesky wont release the relay code or interact with foreign relays. theres nothing federating or decentralized about it.
e. i love their answer when you call them out on it… its ‘but our design takes too many resources to federate! but someday, we will release the code!’
Really? Can you point me to where they have said things along these lines?
Why are more people not pointing out they say federation is basically impossible, so that’s why they are not doing it…
So it’s basically as federated as a fediverse server with federation turned off lmao
And that’s basically what Truth Social actually is, and no one claims that Trump’s server is federated.
The relay code is here: https://github.com/bluesky-social/indigo/tree/main/cmd/bigsky
And according to this you can host everything except “AppView”: https://alice.bsky.sh/post/3laega7icmi2q
Ok…so then doesn’t that take ALL power away from every mod?
Let’s say you ban [email protected]
But then I could create [email protected]
And you could ban it. And I could create [email protected] (or however it’s formatted).
And then this cat and mouse could happen, all day every day. Just a series of me creating accounts, you blocking them, but it then becoming a full time job making sure I’m currently banned, and me evading the system.
Or am I missing something?
The thing is, when that happens, the mods/admins of the trolls new instance ban them there, without the original admin having to do anything.
Abandoned instances, or instances that simply don’t moderate, get defederated, so it’s quite manageable.
The only case where it’s not that clear cut, is where the troll is a big issue for one admin, but not another. Take trans issues for example. I have a zero tolerance policy on transphobia on this instance, but not all lemmy admins are as aware of the dogwhistles as I am, so I will block some users that other admins won’t. It’s not ideal, but it’s manageable, because I can stop their brand of transphobia from reaching my instance even if they’re not banned by the remote admin. And if that pushes the troll to create another account elsewhere to get through the instance ban, then that becomes harassment, and the other admins will act, even if they wouldn’t before.
Nah, you’re not missing anything, and that is already happening. There’s a few users who keep creating new accounts on alternate instances to post troll threads. But it’s on a small-enough scale that it’s trivial for mods/admins to deal with. And realistically, it’s more work for the troll than it is for the mods, so eventually the troll user will either tire themselves out or simply run out of instances to make new accounts from.
It’s definitely an intrinsic problem that ActivityPub doesn’t really have a reasonable solution for yet. It’s sort of a necessary evil in order for the rest of the Fediverse to function freely, and part of the compromise that users and communities need to accept and live with in order to keep things running. But, I don’t think it’ll become a major issue unless the Fediverse sees a massive influx of users, like the sort of scale that Bluesky saw with Twitter’s exodus.
Also, ban evasion is not exactly fixed in other contexts either.
I wanna add to what other users already answered that this problem is not created by federation, only exacerbated.
If I’m mod of a community and I ban your [email protected] account, I cannot stop you from creating, e.g. [email protected] and coming back. Most servers have some barriers against spam account creation in place, but I’d wager you could easily create a handful of accounts on a server until they start to grip.
Even completely centralized platforms such as Twitter and Reddit are the same. You can easily ban/block evade a couple times per timeframe.
This isn’t entirely accurate. Main issue is that this is all very new and moving very fast, so while it’s not fair to call atprotocol completely open it’s not fair to call it centralized either.
Right now you can run a relay independently and you can scrape bluesky’s PDS set with it (guy did that in July and it cost about 150 bucks a month, probably more expensive now. https://whtwnd.com/bnewbold.net/3kwzl7tye6u2y) there’s just not much incentive to do so. This recent policing of accounts might change the calculus on that thought process, but for now it’s not really that you “can’t” it just nobody really wants to. (Additionally there’s only like one client that lets you set multiple relays and/or appviews, so the front-ends haven’t caught up with this idea either.)
Theres about 760 or so independent PDSs right now: https://blue.mackuba.eu/directory/pdses and the list is growing.
You can host an independent blog using your own PDS and not even using bluesky’s Relay network at all: https://github.com/haileyok/blug
Or you can still have your data independent on your PDS but still allow the relay to scrape your blog update mentions: https://whtwnd.com/
More detailed is here: https://whtwnd.com/alexia.bsky.cyrneko.eu/3l727v7zlis2i
tl;Dr: