A somewhat polemic argument on why distributed social-media platforms are not just simply going to solve all our...
A somewhat polemic argument on why distributed social-media platforms are not just simply going to solve all our problems, just by being distributed.
https://blog.kugelfish.com/2018/12/the-fallacy-of-distributed-good.html
https://blog.kugelfish.com/2018/12/the-fallacy-of-distributed-good.html
It'll be interesting to see what social sites are thriving in 5 years.
ReplyDeleteI like the concept of open source and user built.
... the contrapositive of the argument seems to be true: the fact remains, Facebook has demonstrated itself to be a bad actor - and the more power it gained, the worse it behaved. Whether or not anyone else might be a bad actor is irrelevant. Morality is who we are when nobody's looking - or through power over others, we can do as we wish. FB has failed, precisely because its centrality gave it such power.
ReplyDeleteThe reference of the American empire as a neutral/positive entity is interesting considering how much overt and covert disruption of governments it has been involved in since WW2. And the "dark ages" were only dark from the European perspective and likely didn't affect the commoners. One set of rules or another to follow. Perhaps, given the source and administration of Facebook it should have been expected that the network influence could be bought.
ReplyDeleteOpen source doesn't negate that ability, or even censoring, but the user should be able to identify the style of pod one wants to support. It's not a mystery; the Conservative Right in the US has been very successful at identifying and growing their numbers through creating these venues and normalizing their goals which would have seemed extreme not that long ago.
Cass Morrison At the risk of sounding like an obtuse pedant - the so-called "Dark" Ages weren't as dark as we suppose, nor was the Roman world so enlightened. Those conclusions were reached by historians from the British Empire. Someone ought to study the sorts of histori-diculous conclusions of the Famous Historians, matching them to the social norms of their own eras.
ReplyDeleteThe Roman world kept chugging along, especially in Hispania, where it continued until the Moors invaded - and even then! - the Moors considered it Roman. All that Moorish architecture? Roman arch, Roman pediment, just a few curves and wrinkles they brought along from Baghdad. Feudalism was essentially Roman in nature.
We still call Hispanic people "Latino".
In short, the concept of the Dark Ages is nonsense. What followed the last few Caesars in Western Europe and the rise of the Visigoths - has been wildly misinterpreted. Ditto the Renaissance, which only happened in a few rooms in Italy and France and the Netherlands. It was not chaotic. Feudalism was hugely centralised: Charlemagne ruled most of what now constitutes modern Europe via feudal structures.
But to the point about the Fediverse over and against the Empire of Zuck, even the Fediverse depends upon the ground state of the Internet itself. Contrary to its architectural documents, the Internet, a decentralised topology designed to survive a nuclear strike upon Washington DC has become hugely centralised , to the point where the NSA can coopt a handful of inter- and trans-continental fibre bundles - and PRC can build the Great Firewall of China without much trouble at all. That is the problem with all such efforts at decentralisation: they only provide a better grounding for the next Tower of Babel.
Dan Weese Facebook is a corporation which by definition protects its directors and its primary (and only) object is to make money for its shareholders and implicitly for its executive officers. So, no, Facebook isn't "bad" - under its governance it is doing excellent. Corporations are not "bad actors" as such but they could loose customers and good standing and thereby advertising revenue.
ReplyDeleteCaspar Helmer Facebook is a deeply immoral entity whose only commodity for sale is other people’s personal data. When an argument is opposed by use of dissecting quote characters- it is so much squid ink. Were I to sell your personal details in violation of any contract or agreement we might have - and privacy laws, as in Belgium - that is not merely immoral but illegal. FB stands convicted.
ReplyDeleteHere’s the TL;DR, Caspar Helmer : Google sells adverts. Facebook sells you
ReplyDeleteI think there's a whole lot of "meta" vs. "micro" decision-making being forced on a huge number of happy G-plussers. Micro: how will I keep in touch easily with my actual valued friends-acquaintances-communities?
ReplyDeleteMacro/meta: What will replace an established habit/lifestyle, where one stop offered everything, from mobile or desktop texting (Hangouts) to easy sharing/viewing of YT vids, to public posting with immediate indexing, to (once) control over "circles" (emulated by dia/pluspora) to a well-populated public feed, combined with chosen circles.
That said, there's been tons discussed (of necessity, and urgency in part) about "where to migrate" as one polyglot mass, which will not happen, IMHO. Lots to the Fediverse, lots to MeWe (which is doing a huge campaign both for its features and especially, as an "anti-facebook"). But in the end, some will need to be concerned with continuing a visible public presence, some will be shaken to lose regular friends and familiar discussions and feeds, and many are freaking out realizing our years of investment in making collections, communities, etc., are now all threatened (or guaranteed) with extinction. ("Evil Google!")
And finally that gets me to the point, the headline of distributed networks (Federation, Diaspora, Friendica, Hubzilla, etc.) being necessarily a good thing or not. Again, that depends on one's goal and perspective. Some would be happy to have a core set of familiar people and groups still around, in recognizable form, with similar options to interact, share, easily comment (with editing!) etc. Others are concerned with ROI, SEO, or how to sustain blogs and other publications/posts now hosted only on G+, or only publicized "here" as a main activity.
So is the distribution idea "good"? For Tor, for pods not subject to disruption, for a lot of things, technically and privacy-wise it makes sense, IMM. For many it is probably confusing, including many who simply just want to go to an app or desktop and "check Facebook" or twitter (still options). It may be daunting, even though possible and easy for many, to connect across pods, for some. Others won't be bothered. In the abstract, sure, it's "good". So is choice.
Agree with all the comments on being owned/marketed by FB, etc.
Sorry to go on (this is my area I research, lecture on, and "treat")!
One other perspective on how we are used as marketing fodder by mostly Facebook/Google (with TripAdvisor in close pursuit of evil) -
Here's tech guru/Microsoft engineer Jaron Lanier. A light read. ;)
pbs.org - Jaron Lanier's argument for getting off Facebook
Dan Weese the point I was trying to make is that Facebook as a corporation doesn't have morals. Your point about breaking privacy laws and selling data about its users being illegal is what we need to pursue. Unfortunately, FB is not even accused, let alone convicted. And yes, their Irish subsidiary which holds the information of their European (and I believe other non-USA users) is subject to the GDPR (European General Data Protection Regulations) since May 25th - the European Commission will have to take them to task. Write your MEP!
ReplyDeleteCaspar Helmer Corporations are not without morals: law is a subset of ethics. I have defined morality as I see it: now I must define Ethics it seems . Ethics is me telling you “That is wrong. Stop doing that.” Where morality is personal, what I will not do, ethics carries penalties.
ReplyDeleteAt any rate, lying offends everyone - and Facebook lies to everyone- you, me, governments, itself... Facebook has no scruples
And Facebook has been convicted.
ReplyDeletereuters.com - Facebook loses Belgian privacy case, faces fine of up to $125 million
Since i have some experience with this issue I'll weigh in. I do agree with part of your sentiment, just the mere fact of distribution doesn't guarantee good results.
ReplyDeleteHowever, once one lays out a few of the underlying theoretical principles, it becomes clear that decentralization can indeed have some clear advantages. A long time ago i wrote a book on the subject (see link in my profile if interested).
You carefully qualified your position, limiting it to "privacy or data protection". But, you see, the value of decentralization is much broader than that. For example, a key value of decentralization is adaptability, healthy responsiveness to the context. It's no fallacy to say that decentralization enables a system to be more adaptable.
I won't off the top of my head lay out a larger framework for you to examine the issue, but once you do lay it out, it becomes clear there are some critical advantages to decentralizing a system. So, my claim is that once you sort out the architectural and engineering trade-offs (including both machine and human factors), centralized ownership and management of data can be very problematic.
So, my reaction to your essay was twofold. Yes, decentralization is no panacea, but no, centralization is almost never the better way to manage a large system.
So, what i recommend is to put your issue in a systems engineering context, and then you can see where the advantages and disadvantages are. I'd love to give a more detailed answer, but it's not fresh in my mind at the moment, hopefully this much helps.
While this this is very much an exaggerated caricature in black and white, reality is more in shades of grey - at least 50 of them...
ReplyDeleteBill Brayman - at the technical/engineering level there is no doubt about the benefits of decentralized architectures. Even any of todays large social-media sites internally rely on some of the most sophisticated distributed systems technology - there just isn't any other way to reasonably build such large-scale services any other way.
On the organizational level, things are a lot less clear. Economy of scale, winner-takes-it-all patterns or natural monopolies seem to be strong drivers towards centralization and the only major counter-force seems to be regulatory or political action. Or that they become as insensitive to the needs of their constituents that their once dominant position is being disrupted by revolutions or new challengers.
In terms of data protection, one of the biggest risk with large centralized organizations is hostile takeover or change of control, as they represent valuable targets. But if the choices are only to put all our eggs in one basket, either a in a massive well secured basket along with most everybody else's eggs or in some more ad-hoc small basket - not sure what the better choice would be.
The smart use of decentralization would be to spread our eggs across many different and disconnected baskets, which is what many users are doing by having many different accounts with different providers under different pseudonyms for different purposes.
Bernhard Suter A good post, and I should probably answer it in a freestanding article myself. You're asking questions and raising points that need to be asked, especially as they go against the grain for many.
ReplyDeleteI agree with you that there are many who hail decentralised systems as the Second Coming, Silver Bullet, and Philospher's Stone. It's none of these things, and it introduces its own failure modes and vulnerabilities as well. But it also avoids those of the highly-centralised systems we're presently used to. It is a matter of balance.
I've had a back-burnered project to look at the arguments for and against both centralisation and decentralisation. Each has strengths and weaknesses. It seems to me that the critical analysis should be done in terms of risks and threats, and what each option presents or addresses in this regard. And that's another back-burnered project, the largely under-addressed section of #PlexodusWiki on risks and concerns: https://social.antefriguserat.de/
You raise the questions of privacy and data protection. Those are only two of numerous issues people should be looking at (and yes, they're two critical ones). Availability, system security and reliability, administration overhead, skillsets, competence, and ease, familiarity, interoperability, data ownership, community dynamics, organisational and/or institutional longevity and durability, operator trust, protection against hardware damage, loss, theft, or seizure. Those are all additional concerns.
Trying to put my finger on the critical distinction between distributed vs. centralised systems, I think it comes down to:
Distributed: resilient, multiple independent execution, low probability of overall destruction or compromise, alternatives in the event of administrative, commercial, legal, or governmental obstruction. Small targets and large numbers make evading direct assaults probabilisticly likely in many instances.
Centralised: efficient, strong central focus on critical issues, typically immune to low-level degradation when operating healthily, capacity to access and deploy resources (financial capital, skilled expertise, legal defences, marketing and promotions, business and other partnerships), capable of presenting a defence against administrative, commercial, legal, or governmental obstruction.
The strengths of each tend to be the weaknesses of the other.
There are also some common failure modes.
For both approaches, there's a strong tendency for early development to occur in halcyon days, where wind, waves, and tides are aligned, and sailing is smooth. The advantage is you make great time, the disadvantage is that when storms arise (and they will), you are untested.
Both centralised and distributed systems flail against spam, abuse, harassment, propaganda, misinformation, and disinformation, though they do so differently. The alternative-pathways advantage of decentralised systems in routing around damage becomes a liability when defending against information attacks, and addressing the latter decreases robustness of the former. Centralised systems can mount aggressive (though expensive) "trust and safety" operations, but become targets in the process of claims of unfairness, censorship, or incompetence -- both false positives and false negatives become highly visible and potent criticisms.
I think the merits and drawbacks are best addressed on the basis of goals. What is the purpose of your online presence?
What do you want to accomplish, with whom, under what terms, at what costs, both direct and financial, or indirect and abstract, including risk-based costs.?
ReplyDeleteDo people really need a public-by-default sharing of their quotidian observations and commentary?
What sliver of of the public attention do those who do seek it deserve, or receive? Attention is the ultimately rivalrous good. Both individually, yours to give, and collectively, the public's to offer. It is hotly contested for.
What's the balance between "sharing with a known group" and "discovering new voices?" Does that mean finding your tribe or breaking out of your bubble? When is the wisdom (or madness) of crowds appropriate, and when do you need true expertise? How do you recognise it, and how do you avoid charlatans?
*It's also probably the case that the answer is neither "centralised" or "decentralised" but "appropriately scoped". Yuval Harari is an interesting read, though we disagree on points. What human societies offer is operation at multiple scales. There are units of individuals, families, teams, neighbourhoods, communities, towns, cities, counties, states or provinces, countries, empires, the world.
It's not one system. It's a bunch of them. Often working at cross-purposes. In cases, it's possible to move between them, other times not -- either individuals are stuck, or the world, collectively, is looking for, and bent on finding, a specific person. And there's a lot of grey there.
The online world is likely similar. But going from One to Several should be viable.
Lovely, thoughtful comments you guys. Edward Morbius Yes you should flush out your ideas in a stand alone piece.
ReplyDeleteThere's both the psycho-sociological and the systems engineering angles to address.
I want to remind everyone about the role of technical standards. In a way decentralization is a natural evolution because at some point communication and semantic (think data structure and meaning) standards are developed and become the infrastructure, and then the inherent power and monopoly of a centralized system is displaced by many competitors who use the emerging standard building blocks.
Although this idea seems trivial and mundane, it's at the heart of technical and social evolution, and leads to decentralization automagically.
Add to that the notion that Edward raised about the hierarchical array of scale (group/city/county/state/etc) and you have a pretty good handle on how social systems evolve.
Is centralised or not, more resistant to Denial of Service attacks? Feedly has been knocked out by that a couple of times.
ReplyDeleteDiana Studer It somewhat depends on the nature of the attack and the acceptability of downtime.
ReplyDeleteCentralised systems can be very considerably hardened.
Distributed systems are, in theory, harder to attack. In practice, there are often only a very small number -- a Skil Saw handful as I like to put it, numbers less than five -- of actual redundant sources or seeds. And those are fairly poorly protected.
It may not be possible to take the entire platform off-line, easily. But any given work or document, probably yes.
Solidifying this by offering front-ending via CDN would tend to be a way around this.
Diana Studer Off topic, but I'd love to hear some hints and your experience setting up feedly for a good set of feeds. I've seen lists of feeds for popular news sources, but i'd like to go beyond that. Any comments?
ReplyDeleteI've set up Feedly and I really like it. It's easy to review posts and click through for more details. Almost any website can be added and there are a range of viewing options. It's pretty easy to group interests too.
ReplyDeleteBill Brayman I started by transferring from Google Reader. Each month I check thru the ones that have gone Dormant (out? Worth waiting to see if they come back??) and the New ones (not interesting after all, or add to one of my other folders?)
ReplyDeleteMy Feedly is partly blogs, and partly science (water, botany, weather). Good new feeds I find by using my existing sources to curate for me. Currently reading 73.
Bill Brayman The problem with all these technical standards is how un-standard they become over time. One of the longest-lived of all the comm standards is EDI / ANSI X12.
ReplyDeleteTrue story: I was on an integration project. One drug firm had bought another and they're trying to deal with the reorder process. I'd been dragged into this project after a huge mess had been created. They were trying to write a one-size-fits-all solution - and it was failing hugely.
Let's face facts, when Walmart has an EDI standard, its vendors and suppliers are going to adhere to it - or they'll be kicked out of the Walmart vendor list. You do it Walmart's way. The US military has its own version of EDI for AAFES. Albertson's. Food Lion. Everywhere you've ever seen a bottle of Listerine, that outfit is ordering stuff which passes through my software. And every one is different.
Everyone has their own little wrinkles. For all of the supposed "standardisation", everyone ends up having to do it everyone else's way. It's an enormous mess.
There's supposedly a standard for health care EDI in the HIPAA legislation but it's still screwed up. All of the Blue Cross / Blue Shield / Anthem claims go through my validators, too.
EDI has been around forever - and it will probably continue until the sun burns down to a cinder - not because the standard is any good - it's not. It's because of its ubiquity, in all its forms. Just understand this about any "standard" - that's an existential statement. The way I do things is true and good and adheres to the standard. Your bullshit implementation means I have to write a translator. Get it now?
Dan Weese Well, ok, some standards get all messed up. I think in your case you're dealing with a standard that was premature because semantic standards are just now becoming understood. I've dealt with data interchange standards and they are indeed a nightmare. But the underlying problem is very complex, and i would say that it's still evolving.
ReplyDeleteI'm more familiar with internet standards. HTML for example has gone thru some confusion too, but for the most part has been a brilliant result. The upcoming activityPub standard is now experiencing what you're talking about. There still isn't complete agreement on how to standardize online social activity. Consequently, diaspora, like your EDI example is bucking the standard because it doesn't address certain issues the diaspora team has. But, not to throw the baby out with the bath water, the take away is that standards work well eventually. Tough cases like data interchange or social activity take a while, a long while perhaps, to become successful. But success will happen and the value of standards will emerge.
Bill Brayman HTML has made a dog's dinner of the Internet, with its lack of standards. I am pretty sure you were old enough to remember the dawn of the IMG tag. Fun times, eh? While the standards bodies were trying to work it all out, the porn industry ( the big money maker back then, if you recall ) took off without them.
ReplyDeleteThe standards bodies are often the problem. Linux succeeded. not because its monolithic kernel was the greatest, it wasn't - but because it was the product of a single person, backed by lots of big corporations' money, corporations like IBM who saw its potential and Red Hat, who saw money in corporate support.
Standards, ecch. I'm all for 'em. But I never put any weight on any of them. I've been eaten alive more than once. CORBA... I still have scars on my ass from that one.
Dan Weese With your general dislike of standards i think you're missing something big.
ReplyDeleteWhew, i'm still shaking my head at the comment "HTML made a dog's dinner of the internet."
Conversation for another time perhaps, but not here. Over and out.
g I used to "dream in html", literally, and now,, like ability to write (cursive or otherwise) it seems a useful, if dated, skill. Still. As for "dog-fooding", I revert to that standard: Google+ should stand behind how much they love and care for G+ !! ;)
ReplyDeleteMichael Fenichel Weese's dog comment wasn't the "eat one's own dog food" sentiment, his use was to disparage it altogether.
ReplyDeleteSigh, here's the basic story of the birth of the web. Three standards - URL, HTML, and HTTP enabled the web. A way to uniquely identify resources, a way to mark up document structure, and a way for client/server interaction using such naming and structuring.
There was literally explosive growth of the web specifically due to those standards.
And before that a similar story with lower level internet standards, TCP/IP, which enabled the explosive growth of the internet.
Don't get me wrong, I love standards. I don't grasp why other people don't adhere to standards. They just don't.
ReplyDeleteI don't understand people who don't build tests for their software, either. Or write spec.
So I've learned to cope with it by accepting that the world is not my oyster and writing my own parsers where I have to.
Standards... ( smh ). There will be two moons in the sky when the so-called standards-based streams I read conform to the standards, so I can just post a nice error and stack trace and someone else can get yelled at, eh?
Other people's screwed up XML and HTML? Won't parse through standards-based mechanisms such as Jackson ? Flails around in your DOM? Do you think I have a choice about what my systems are fed? I don't. I have to cope.
Over and out.
I get it, Bill Brayman and many others are clearly super well-versed in systems, coding, APIs, etc. (My thing 20+ years is simple plain-vanilla html, web-wise). But I think it's important to keep in mind the overwhelming numbers of casual "social media" users who are either (a) "in the moment" and not at all aware or interested in the underlying code or platform structure; (b) focused on the looming end of a big aspect of daily life, with G+ imminently leaving a void which many are scrambling to preserve as much as possible; (c) tech-savvy, maybe developers or designers or content providers, but not seeing any clear "replacement", possibly shopping for ways to export existing G+ destinations, maybe willing to reconstruct, maybe seeing it as too daunting.
ReplyDeleteMy point (the dog-fooding reference being a point about Google's abandoning its own baby) - Many users are like FB and AOL lovers who are happy when "it just works". I think G+ refugees are unique by virtue of our numbers and a wide reticence to replace G+'s inclusive/intrusive platform with FB's. My mantra: "Context and perspective". Such a shame, this evil-made unforced displacement of happy Google consumers... Have a Happy! /done
It's not all that hard to figure out, Michael Fenichel . G+ was not perfect and attempting to recreate it is just another fool's errand. Tell you why:
ReplyDeleteG+ was not Google's first trip to the Soshul Meetya rodeo. It was, in fact, its fifth. But it had a curiously wonderful bit of code, Orkut, which it murdered so as to make way for what would become G+. I was on Orkut. Orkut was really good. But Google, true to form, couldn't find its own ass in the dark with a flashlight and Google Maps. It understood advertising. It did not understand people. And still doesn't.
G+ scared the shit out of Facebook, but it shouldn't have - Vic Gundotra was a very nice man and a well-meaning incompetent. Gundotra had no idea what sort of boat he was skippering. It could have beaten Facebook into the dirt.
G+ is a case study in horrible management. G+ was SUPPOSED TO be some sort of identity management thingie and this social media thing was SUPPOSED to be a way for companies and their customers to interact. Instead, people looked at it, delighted, and realised it was a damned near perfect ( for the time ) way of kicking around memes and photos from Picasa and later Google Photos and integrating with Hangouts ( which is slated for execution soon enough )
I'm not telling you anything you don't know. Google never understood us, never supported us and never once thought to do much except force us to use our real names for a while.
I've just fired up my Fedora box and have a Ruby toolchain established. I pulled down the diaspora code and I'm taking a look at it. I don't like Ruby, never have, but it's a good language for this sort of thing - so I'm going down that route. Because there are two kinds of people: those who complain and those who worry they're not doing anything to change things usefully. May I never be the former.
If there is to be a viable social media engine, I contend it will be build around what the sociologists are telling us. I'd like, at this time, to point out what happened with the advent of two media mechanisms, the printing press and television.
ReplyDeleteHere's the TL;DR : these technologies breed markets, almost immediately upon their creation.
The impact of Gutenberg's printing press cannot be overestimated. But it brought something into focus which had only been a dream since the Greeks: the widest possible dissemination of information imaginable at the time.
In the wake of the printing press, the Frankfurter Buchmesse == book market - the first information interchange mechanism for printed material emerged. We're pretty sure it was almost immediate: the Buchmesse opens in 1454 and Gutenberg begins printing in 1452.
Television arrives in fits and starts. But Philo Farnsworth transmits his famous USD $ sign in 1927, after an investor asks "When will we see some dollars out of this thing?"
Gentle reader, think about your life. The Iroquois people say a human being has five fingers on each hand because that's how many friends he can have: all the others are just acquaintances.
No social media engine understands that concept. Most of the people you know, however well-intentioned, are not your true friends. Oh, the others might be a bit insulted if they knew they weren't - but we all kinda understand boundaries, those of us with a clue.
Other people, our acquaintances and such, G+ had Circles, but that was a stupid way of handling all that access business. Trust is earned. Who has earned your trust? How would you gauge such a level of trust?
I'm not the idea guy. I'm the guy who says, with Da Vinci
"Although nature commences with reason and ends in experience it is necessary for us to do the opposite, that is to commence with experience and from this to proceed to investigate the reason."
Whoever models trust most effectively - and earns our trust in the process - will be the big winner. FB will lose. It was always a treacherous busybody.
Dan Weese Excellent. I'd add one big, 3rd (r)evolution to printing press and TV - the "interactive" aspect. Equally world-changing when suddenly people could, for the first time, talk back to the screen, or blog (nee "column"). Letters to the editor? Reactions are now tweeted, liked, faved nanoseconds after any post or event; There are many active "followers" for fun & profit. And yes, across every single one of the Big 3 (modern) Media eras, trust has been vital, from Walter Cronkite to "all the news fit to print" to "fair & balanced" fake news announcers and bots. Interactivity is now taken for granted, one person's "social marketing" is another's "social media" life.
ReplyDeleteMichael Fenichel Yeah.... and I don't mean to disparage or demean anyone, especially Bill Brayman in the process.
ReplyDeleteA "dog's dinner" is a British-ism. A cook attempts to make a souffle, the damned thing collapses, it can't be served in a restaurant, it's a "dog's dinner". The Interwebs were sorta inevitable, shadows of it were seen in CICS in the 70s and the BBS system in the early 80s.
But seriously, HTML was meant to be a footnoting system for scientific papers. Let's not forget that. It was not designed for security.
And more's the pity, because it could have been....
Dan Weese 1454? When I went to work in the library in Zurich, I wondered why we had to wait ... for our new books, until the Frankfurter BuchMesse. Nothing quite like that for the English language. Half a millennium? I am reeling in shock.
ReplyDeleteBill Brayman would you mind asking dan weese if he'd mind unblocking me? Thanks.
ReplyDeleteDan Weese Edward Morbius asked me to ask you to unblock him. "would you mind asking dan weese if he'd mind unblocking me? Thanks."
ReplyDeleteBill Brayman I'll unblock him, on your good graces. But he's incorrigible.
ReplyDeleteTo get somewhat back on topic - In the comment thread on diaspora, somebody pointed to this interesting blog post: blog.joinmastodon.org - Why does decentralization matter?
ReplyDeleteAlso an interesting distinction here on degrees of decentralization - federated vs. distributed: https://docs.joinmastodon.org/usage/decentralization/
ReplyDelete