Applications have largely become client-server distributed systems. Composed programs are actually two separate programs. This is effectively an impedance mismatch, which is the root cause of complexity in web development. Source:

Did you listen to the latest revolution.social with Mike Masnick about bluesky? I understand that rabble tries to keep it nice, and wants to discuss a variety of things. But they briefly every so gently went into the (lack of) decentralization of bluesky and only sort of discussed it by referencing or implying the discussion whilst not actually having the discussion. The episode for anyone that is wondering: Revolution.Social: Techdirt’s Mike Masnick on Growing Bluesky, Clueless Regulators & the Case for Optimism File: (Fucking hell when are Nostr podcast events going to be a thing, this is terrible)

Uhm, I just can use another “app” to communicate with others who use the Facebook app?

Can you do that on the current Facebook? You can as long as Facebook gives you API access or something like that. It's the same on ATProto.

Are you sure it's a good idea to listen to this? You think this Masnick guy is honest?

I could be wrong, but I think he's getting at most users won’t bother switching clients just to see one person’s posts and they’ll most likely stick with the mainstream app (in the current case of at proto, Bluesky) - and in this case they won't be able to see your posts from another app because fetching data from your pds is prohibited by the main app overall, no matter where it comes from. Your friends would need to use the client that pulls from your pds to see them not a knock on at proto but the concept of bluesky being a dominant app on at proto

Hmmmm, can't recall if i read his 'protocols not platforms' thing so i don't even know if i agree with his base perspective. He is honest that bluesky is not fully decentralized; its just the same old 'but we will get there eventually, thats the plan, truely, because we all have the intention to make it so' type of reasoning. So you could say he is honest actually, in so many words he just straight up admits that he thinks bluesky can get away with lack of decentralization for now, and also admits that bluesky's innitial succes is due to the short term benefits these shortcuts provided them. All these things are adressed, but only touched on ever so slightly. He mostly inb4's it all, probably because rabble would point to the elephant in the room anyway. At the same time rabble is ofcourse very friendly. Rabble does push back and hit him precisely where it hurts, but does not go for the kill lets say. So to directly answer your two questions: Yes i think it is worth listening to because he is honest enough, and rabble is critical enough.

I've listened to it. It's so vague when they say "Bluesky will be more decentralized". They have been saying this for years, and a bunch of their users read that as "federation", assuming they will become like Mastodon (because that's the only example of decentralized social network they know), but in fact Bluesky has never planned to be like Mastodon at all and all their future-planned-upcoming "decentralization" updates are irrelevant things like people being able to host their own PDS. At some point it becomes a little more clear that Masnick doesn't have anything in his mind about the "decentralization" other than what Frazee says in the article above: their entire vision consists of this competition between all-powerful "apps". Masnick says something like "if we have this in place then Bluesky won't be able to become evil because if they do then everybody will migrate to a competitor overnight", and I've seen Frazee also say this in the past, that the entire point of ATProto was that "if everybody realized that Bluesky was doing bad things they could all switch to Skyblue". I don't know how they don't see how that is impossible to work. A huge part of the current Bluesky userbase is already furious with them, but they're not leaving because of their network effect, it is an immense cost to leave, so the platform can continue to do a lot of bad things, exactly like Facebook and everything else. Maybe ATProto makes it 5% easier to setup a competitor that can reuse existing software, but software was never the problem. Right after Masnick said that Rabble started giving the example of how the existence of GitLab places a check on GitHub, an the example couldn't be more appropriate. Despite the fact that GitHub doesn't have anywhere near the network-effect that a social networking website has, still the little bit it gets is enough to prevent anyone from leaving at all, even if the overall hate for GitHub has been growing for years.

That's a good summary. Also zooming out they're going about it one of the very few ways you can go about it if you want to complete with the UX of centralised services and to scale as they have. How their experiment ends up is wait and see, but if the goal is to compete on some level with that UX then it's pretty clear the Nostr/ActivityPub path is not going to work. Nostr and ActivityPub are for users that are willing to let go of certain key aspects of that UX, to embrace the chaos as it were.

For the record, running a Primal cache relay is no small task. You have to get familiar with the Julia language first of all, which is an interesting language but quite niche. And it's also a pretty heavy lift on the VM side. I think their vision is (was?) that there'll be other "pro" operators running alternative cache relays, since your average person, even if technical and familiar with say strfry, is not going to be comfortable to spin one of those up.

I think I get it what you're saying. As today most data is stored with the the same entity who also controls the views infra (Bluesky controlling the PDs infra and the View).

I see, I think I get it :) In the current context of being banned from a view, this now also results in your PDS being made inaccessable for other views (because both are controlled by the same entity).

No, that's not what I'm saying. Where the data is stored doesn't matter because people are not reading from the PDSes, they're reading from the "app". So if you use another app no one will hear from you.

Another thing I forgot is that they seem to operate in this realm of "being evil", like an "app" can either be good or evil: if it's evil everybody will migrate, otherwise everybody will stay. But this is obviously not true, the same policy can be seem as good by one group of people and bad by another, but in the ATProto world they don't seem to realize that.

Of course, some ATProto lovers may defend their protocol by saying everything can be worked around by making new clients that connect directly to PDSes or something like that, which is stupid because the entire protocol makes that very difficult and you would just be recreating Nostr on a worse foundation.

But the apps are supposed to read from PDSes right..

Oh, in the beginning of the interview Masnick talks about "freedom of association" and the idea that each server should be able to have their own rules, much in the veins of so I thought he was going in the right direction, but later he forgets that entirely and reverts to the all-powerful "app" paradigm above. It could be that he had good ideas that could have lead him to something more like Nostr with sovereign servers that make their own rules, but them he got confused by the weird view presented to him by the Bluesky team that involves "labeling" and "custom algorithmic feeds". I've seen many people (although that is decreasing now) talk about those things (they often use the misnomer "algorithmic choice" which makes no sense) as if they were relevant to the discussion of "Bluesky decentralization", but they're not. They are very cool and elegant and I understand why people would like the idea, but they are just features provided by a centralized "app" (in fact they assume a centralized server and can only work with that), they're not principles of a decentralized protocol.

So you get this world in which all the problems of selective censorship are "solved" not by the app (i.e. the centralized server) censoring things, but by giving the user the tools to censor what he wants. This will obviously break because the censorship is never about what you want to see, but about what you want to prevent others from seeing, so there will be pressure for the app to apply absolute censorship. And the app owners themselves will want to censor some things, like they have done with CSAM from day one, and no one disagrees with that at first, but then as time passes their opinions will change too, just like it happens on every platform. The same applies to the choice of algorithms: users are free to choose, but only within the limits of what the app allows them and for as long as the app allows them.

E você é honesto? Você é uma piada, ninguém o leva a sério, do Bluesky ao Discord isso é unânime. 😆