Post AxtfliCBDjdl3JefMu by ricci@discuss.systems
(DIR) More posts by ricci@discuss.systems
(DIR) Post #AxtfjqDaQBq9M5Q9SK by jwildeboer@social.wildeboer.net
2025-09-05T08:09:14Z
3 likes, 5 repeats
"But #Bluesky is also federated, decentralised!1!!" Well, no. https://arewedecentralizedyet.online by @ricci
(DIR) Post #AxtfliCBDjdl3JefMu by ricci@discuss.systems
2025-09-05T13:13:50Z
0 likes, 0 repeats
@jwildeboer I certainly hope that the technical mechanisms we see in the #atprotocol for decentralization will get deployed more widely in practice, and we'll get to watch #bluesky get more decentralized over time
(DIR) Post #Axtk7n8xu5eeaPw2e8 by Flo_Rian@norden.social
2025-09-05T12:43:15Z
0 likes, 0 repeats
@jwildeboer off-topic re: visualization The gauge might be easier to interpret if divided by labeled areas with colors and the number being the same color as the section it's in. Something like 0 "perfect competition"1 - 1499 "competitive" (blue/green)1500 - 2499 "moderately concentrated" (yellow/orange)2500 - 9999 "highly concentrated" (red/purple)10000 "monopoly"@ricci
(DIR) Post #Axtk7oX6joYotaEpuq by ricci@discuss.systems
2025-09-05T14:02:32Z
0 likes, 0 repeats
@Flo_Rian @jwildeboer That's an interesting suggestion, and I'm divided on it. On the one hand, I don't think it's 100% clear that the breakpoints used in antitrust are necessarily the right ones to use in this context, so staying fuzzy on the actual gauge might avoid suggesting sharp cutoffs where they don't really exist. (I did feel the need to include them in the text to give *some* context.) On the other hand, I do think it would be good to show that there is still a benchmark for the fediverse to hit, that it can still get better
(DIR) Post #Axtl2bw0FdaiKPxw3c by iju@mastodon.social
2025-09-05T14:06:01Z
0 likes, 0 repeats
@jwildeboer @ricci Do I read this right that Bluesky is about 12x bigger? And I assume we're counting active users?
(DIR) Post #Axtl2dKV42mSegR0sa by ricci@discuss.systems
2025-09-05T14:12:49Z
0 likes, 0 repeats
@iju @jwildeboer This is complicated.What I'm counting is the PDSes that the bluesky relay reports knowing about, and how many repos (user data) they report having.The number this returns is different from what you may see in other places. You'll see 38M for the number of total accounts. You'll see a few million listed for Monthly Active Users. My understanding is that what the relay reports is *some* measure of active users (it's the number of accounts the relay is currently maintaining data for) but it's not exactly MAU. The Fediverse numbers are MAU, but it's also notorious that you can't get a fully accurate number for this because not all instances choose to opt in to such data collection, often by misreporting numbers.
(DIR) Post #Axtm7ZCrvAX9GmmW9I by iju@mastodon.social
2025-09-05T14:25:00Z
0 likes, 0 repeats
@ricci @jwildeboer Thank you for explaining this to me. As might be evident from my profile, I'm not technically more that passingly literate.
(DIR) Post #AxtnL48SL7N8udiaem by Flo_Rian@norden.social
2025-09-05T14:38:34Z
0 likes, 0 repeats
@ricci I just had a hard time reading the graphs at first. Gauges (in reality) usually have a scale or zones, so we expect to be able to tell if we are in a certain range or hit "the red zone". In this case, one seems to have zones (green area, with the needle just on the edge), while the other one is all red.To avoid sharp cutoffs, it could also be one large gradient. Alternatively, not using a gauge graph would be an option, maybe a partially filled bar chart.@jwildeboer
(DIR) Post #AxtnbvVHgZYN43R4oy by stinerman@mastodon.social
2025-09-05T14:29:07Z
0 likes, 0 repeats
@jwildeboer i agree with the sentiment 100%, but to be clear Bluesky isn't any more or less decentralized than mastodon.social. Bluesky is a service that runs on atproto. Just as mastodon.social is a service that runs on ActivityPub.atproto is theoretically decentralized in a similar way to ActivityPub, but there's just no one else of any size building on it. atproto is decentralized in theory, but not in practice. Hence the terrible graph you posted.@ricci
(DIR) Post #AxtnbxFPCa2YTO1QLg by ricci@discuss.systems
2025-09-05T14:41:40Z
0 likes, 0 repeats
@stinerman @jwildeboer There are some initiatives going on right now that *might* make a difference in this, blacksky now runs its own instances (using its own implementation) of most of the pieces. It's still small in terms of the userbase on its PDS, but this could be an inflection point. Or not. Which is why I think this is an interesting time to start watching.
(DIR) Post #AxtntLWiBaPchEPSzI by jwildeboer@social.wildeboer.net
2025-09-05T14:44:46Z
0 likes, 0 repeats
@ricci @stinerman I am mostly saddened that the VC driven Bluesky folks decided that the NIH (Not Invented Here) syndrome led them to creating a new protocol instead of adding their weight to the ActivityPub/Stream protocol family. Now we have again a fractured ecosystem and factions. Le sigh.
(DIR) Post #AxuV0jpUWTBS7txt7w by ricci@discuss.systems
2025-09-05T22:47:59Z
0 likes, 0 repeats
@xerz @jwildeboer Oops! Thanks for pointing this out!Fixed
(DIR) Post #Axy0OwbUMOzmkRJjxg by jwildeboer@social.wildeboer.net
2025-09-07T11:51:34Z
0 likes, 0 repeats
UPDATE: @ricci has moved the code powering https://arewedecentralizedyet.online to codeberg (yay!) and added a third gauge to visualise how centralised the decentralised git based ecosystem is :)
(DIR) Post #Axy0Oxodqz6wVWTl6O by thibaultmol@en.osm.town
2025-09-07T12:15:18Z
0 likes, 0 repeats
@jwildeboer @ricci I don't know how feasible it is to do one for matrix but it would be really interesting
(DIR) Post #Axy0OyIQ4GvjzsPXW4 by ricci@discuss.systems
2025-09-07T15:23:47Z
0 likes, 0 repeats
@thibaultmol @jwildeboer Yeah, I would be very interested in doing this for matrix but so far I have not found a dataset that would enable me to do so - and I don't know matrix well enough myself to try to collect one.I am also interested in doing this for email, but that is thornier - in part because the data is valuable to marketers and therefore the sources I've found that *might* work are expensive, and I'd have to pay to even find out for sure
(DIR) Post #Axy0sTJNpdSeaVvdRo by ricci@discuss.systems
2025-09-07T15:29:09Z
0 likes, 0 repeats
@0x10f @Flo_Rian @jwildeboer I realize that is an issue, but I wanted to use the actual metric, where lower is less concentrated, because I didn't want to go off modifying stats. I also thought doing the gauges backwards was likely to be more confusing.
(DIR) Post #Axy1HbpVIatIFIwHom by ricci@discuss.systems
2025-09-07T15:33:44Z
0 likes, 0 repeats
@irelephant @jwildeboer In a recent interview with Rudy of blacksky, he described the PDS as the most complicated part (and they should know, they have re-implemented or forked basically all the parts now). Yeah this surprised me too.The anecdotal information I've found on the appview itself suggests that it follows a similar distribution to the PDS, but with far fewer of them so far. I'm looking for a data source, though. Above the appview layer, I do think feeds and labelers are likely to show interesting decentralization; labelers are probably possible to get actual usage data on, but AFAICT only bluesky can provide data about the use of feeds, and they don't (all that's available is the number of likes, which is not the same thing.)
(DIR) Post #AxyBe9pGa7k9azDcLg by ricci@discuss.systems
2025-09-07T17:29:49Z
0 likes, 0 repeats
@gabboman @jwildeboer There is a difference between the protocol enabling decentralization and decentralization actually happening in practice. atproto is designed to be decentralized (well, with the exception of did:plc); the question of course naturally arises whether it is getting used that way or not. Clearly, *at the level of where user data is stored*, that is not currently how the vast majority of its deployment has been done in practice. This very well could change, and the protocol is (again, with the exception of did:plc) designed to facilitate that change. Will it change? I don't know, let's watch and see; that's what the site is for. But I also want to say that eing new and big is actually a potential impediment to atproto becoming more decentralized. Most systems *tend* to trend towards more centralization, and network effects *tend* to keep people in the places where lots of other people are. So, while atproto is *designed* to permit decentralization that does not guarantee that it will *happen*, and starting as one big ecosystem controlled by a single entity makes it *less likely* that it will achieve meaningful decentralization. There are people trying, though - blacksky being the clearest example. Is this the start of an inflection point where more communities spring up and migrate off the centralized infrastructure? Maybe! Maybe not! Let's watch - that's what the site is for.
(DIR) Post #AxyBhBCnuOgQk8fYUy by ricci@discuss.systems
2025-09-07T17:30:24Z
1 likes, 0 repeats
@palladiumasteroid @eris @gabboman @jwildeboer There is an explanation of this on the page.
(DIR) Post #AxyBnxwNcqhFd7Dsqe by ikuturso@mastodon.social
2025-09-07T17:20:44Z
0 likes, 0 repeats
@iju @ricci @jwildeboer FWIW this tracks BSky monthly/weekly/daily active users and puts the MAU number currently at around 5.65 million: https://bskycharts.edavis.dev/edavis.dev/bskycharts.edavis.dev/bsky_users_total.html
(DIR) Post #AxyBnz0feNiswiEpAu by ricci@discuss.systems
2025-09-07T17:31:38Z
0 likes, 0 repeats
@ikuturso @iju @jwildeboer Ah, very nice, thanks, I may be able to use some of this data!
(DIR) Post #AxyDQAMjyeqP7lKeNU by ricci@discuss.systems
2025-09-07T17:49:39Z
1 likes, 0 repeats
@palladiumasteroid @eris @jwildeboer @gabboman FWIW:* I have updated the labels on the tables in response to this comment to make it more clear what is being measured* I am interested in other measures of decentralization as well. For identity, the currently deployment of identity in the atproto world is *much* more centralized that the PDSes (more like 99.99% on BlueSky PBLC's did:plc system), and unlike PDS, has no credible exit. Anecdotally, it seems like Appviews are about as centralized as the PDSes, but I am looking for hard data on that. Similar with relays. Labelers are probably more distributed, I think it's possible to get data on them, so I'm working on that. Feeds may be more distributed as well, but only Bluesky can provide information on their actual use, and as far as I can tell, they do not. So, basically PDSes are the measure of decentralization *in practice* that there is the most hard data on that is the most generous to the atmosphere, which is why I'm going with that.
(DIR) Post #AxyITZlOVP7oX7JD0K by hadronized@fosstodon.org
2025-09-07T18:42:42Z
0 likes, 0 repeats
@jwildeboer @ricci is SourceHut considered centralised or decentralised to you? To me it’s decentralised as anyone can contribute by just having any email address but I’m sure there are nuances.
(DIR) Post #AxyITawQ7tXUBbTWpU by ricci@discuss.systems
2025-09-07T18:46:22Z
0 likes, 0 repeats
@hadronized @jwildeboer In this context, the question I'm asking is whether the *ecosystem* is decentralized with respect to where data is stored.So, in this case, sourcehut is one participant, as are github, codeberg, gltlab, all the forgejo instances that my data source (Software Heritage) knows about. The question being looked at here is whether user data (in the git world, repos) are concentrated on a few participants or spread out across many.
(DIR) Post #AxyWJll6S29eCMOHfk by simmac@chaos.social
2025-09-07T20:42:06Z
0 likes, 0 repeats
@jwildeboer @ricci at least I can transfer the user data within the ATProto-Universe, while my ActivityPub data/profile is trapped at the instance I'm using
(DIR) Post #AxyWJnD93GBChcWC1I by jwildeboer@social.wildeboer.net
2025-09-07T20:47:08Z
0 likes, 0 repeats
@simmac Migrating profiles from instance to instance is simple. Moving old posts is a bit more involved, but also possible. As I treat my posts here as ephemeral and have only very short retention set, it wouldn't bother me personally, though. @ricci
(DIR) Post #AxyWJoKyrc2eCDBxs8 by ricci@discuss.systems
2025-09-07T21:21:27Z
0 likes, 0 repeats
@jwildeboer @simmac I used to feel this way about my posts, Jan, though over time I've gotten more attached to them and would miss them if I migrated. I appreciate that atproto considered that from the beginning, and I hope more fediverse software works on better migration.FWIW if you want to know what it looks like to move from one PDS to another right now, this video shows how to migrate from Bluesky to Blacksky. They've made it fairly easy, though I think we can also agree that migrating PDSes will need to get a bit easier still to be useful to a large set of users. https://www.youtube.com/watch?v=_SdmiCRYeZASomething else that needs to be considered re: data ownership is that atproto separates that from identity ownership. In some ways this is cool and good, but keep in mind that right now approximately all bluesky users are using Bluesky Social PBLLC's identity service, and there is no credible exit from that. And I don't just mean if you have a bsky.social handle - even if you use your own domain name for the handle, if you are using a did:plc identifier, which approximately everyone is (I checked yours simmac, you are did:plc), you can move your data, but if Bluesky decides to stop resolving your identity, your data becomes a basically unrestorable backup - your identity is gone and no one will see that identity on the network no matter where it is. Unlike the PDS, which you can move off of bluesky's servers right now, or wait until later if they go bad in some way, you can't move off their PLC server without losing basically everything except your backup, which is only useful to you. This is in contrast to the fediverse, where yes your identity does change if you move servers, but there is a mechanism in place to redirect to your new identity, preserving at least your social graph.
(DIR) Post #AxyWmNN0CsxKxKl5Oq by ricci@discuss.systems
2025-09-07T21:26:39Z
0 likes, 0 repeats
@nelson @jwildeboer I strongly agree that the issues that affect this are human first, technical second. This is why I find discussions of the capabilities of the protocols to be interesting - and yes important to how the networks *can* be shaped - but to be missing a large part of the point. I think that's why it's worth looking at numbers that act as an (imperfect) proxy for how people are actually acting, alongside discussions of how the protocols enable them to act.
(DIR) Post #AxyWs3IDlYHeQREEoi by jwildeboer@social.wildeboer.net
2025-09-07T21:27:40Z
0 likes, 0 repeats
@ricci Yes, I think the fundmantal question in decentralisation isn't storage of posts, it is identity ownership/sovereignty. In ATProto it is practically centralised. In ActivityPub it is instance owned but with higher levels of freedom. It is still not user-centric, though. Real decentralisation would *start* with user based identity ownership and instances being an attribute of that identity. Right now it's the other way round in both networks, essentially. @simmac
(DIR) Post #AxyXFhF3P2p0z91PPc by can@haz.pink
2025-09-07T11:33:52Z
0 likes, 0 repeats
@soop @jwildeboer @ricci very cool! Love it!
(DIR) Post #AxyXFiDflfJM19Notk by ricci@discuss.systems
2025-09-07T21:31:52Z
0 likes, 0 repeats
@can @soop @jwildeboer One of the interesting federating aspects of git is that it can be (and is :) ) in both places at once - I set it up on github because I figured that'd maximize the chance of getting pull requests, but I've now moved my primary copy to codeberg. I can accept PRs from either place and push to both, which is cool.git is also a good cautionary tale, though, of the fact that just because the protocol itself is radically decentralized doesn't mean the deployments will inevitably be so. A lot of us are (rightly, imo) worried about the dominance of github even though we can take our repos anywhere we want with relative ease.
(DIR) Post #AxyZ9SdczBbZrbfcqe by jwildeboer@social.wildeboer.net
2025-09-07T21:42:31Z
0 likes, 0 repeats
@ricci One way to get a bit closer to identity ownership in ActivityPub/Stream is to run your own single-user instance (which I do since 2018). I fail to see how to have a comparable setup with ATProto, as it consists of more dependent components which form the network while in AP it is more atomic as all needed services are combined in an instance. @simmac
(DIR) Post #AxyZ9TpMZ2aPYIAVmK by ricci@discuss.systems
2025-09-07T21:53:14Z
0 likes, 0 repeats
@jwildeboer @simmac I believe you can get relatively close to this in the atproto world today if you are determined enough. You can create an identity with did:web, you can run your own PDS, you can run your own relay (it's down to the cost of a modestly-sized VPS, not the huge resources it used to require), and you can run your own appview. You can also imagine that someone might package this up all neatly in the future, in the way that running, say, an instance of Mastodon actually packages up a webserver, a database, a queue runner, etc. We're talking mostly potential here, not current reality, but current trends suggest it *might* happen.