A mental model for decentralization
jacobobryant.com
jacobobryant.com
That said, most people can pick up the meaning of centralization based on the context without every discussion having to spend a paragraph defining it. A discussion about a centralized Internet isn't too hard to understand, nor are discussions about finance, media, markets, governments, etc etc... The idea that people can't figure out what decentralization means in the context of a discussion about the Internet or finance unless someone spells it out in precise detail is kind of obtuse.
All I can say then is that empirically, if I talk about a centralized Internet (700,000 results), people are able to understand what it means... if I talk about a centralized monetary system or centralized banking system, people understand that too. These all appear to be commonly discussed and researched topics that people are aware of. Someone might need a precise definition of "centralized psychology" or "centralized chemistry" since those topics do not seem to produce much of any academic research, but centralized finance appears to be well understood.
If you don't believe it, you can see for yourself by doing a search and going over the abstract of several academic papers that use the term without having to spend paragraphs precisely defining it.
Do you not know how to use Google? "Centralized finance" returns 685 results, and if you go through the results you'll quickly see that most don't even talk about the financial system.
There may not be one central planner but there are definitely power centers that exert far more influence than rest of the participants combined.
There’s fed which walks a tight rope of managing inflation while keeping low unemployment rate. And then there’s federal government, by far the biggest spender in the market. You also have Fannie Mae etc selling mortgages which is huuuge.
There is and it is called the Board of Governors of the Federal Reserve System. By setting the EFF rate (or QE policy if EFFR ~= 0%) of the global reserve currency, they control many facets of the financial system and broader economy.
It's not very fine grained control ("bake X loaves of bread"), but it is centralized control and planning.
I doubt blockchain is a silver bullet. I really liked Kleppmann's work on local-first software / CRDTs. It seems a lot closer to something practically usable at least.
I'm not anti-crypto, and in fact I work in the space and am primarily focused on some areas I feel are in need of being "decentralized" in the CDN and streaming video spaces. And while a ton of technical progress has been made, I can't shake the feeling that we are in many cases implementing this stuff wrong, or haven't even bothered to address the "how" or "why" in as much length as one should before going out and building it.
Keep in mind that this is a somewhat large space, and depending on what parts of the decentralization space you're looking at you might see players who are all-in on blockchain tech, or you might see people who view it as a giant waste of time or even actively campaign against it. In particular when I look at projects that are using federation (Mastodon, Matrix, etc), I see a lot of people who are fairly skeptical about cryptocurrency and outright hostile to NFTs.
I have my own biases there, I don't think that NFTs are in practice particularly decentralized, and I've slowly come to believe that cryptocurrencies have both systemic problems that make them less valuable as a decentralization tool than they're often advertised as being, and community problems that make them difficult to use as a basis for any kind of political movement (and decentralization is inherently a political idea/movement).
But my biases aside (I'm not trying to make a pro/anti blockchain post here), just remember it's a big space without a single uniformity of views; from the circle of advocates I'm paying attention to and interacting with the most, I would have almost guessed the opposite, absent better stats I'd have guessed from my experiences that at least a plurality if not the majority of the decentralized "community" is against this stuff. Interesting to run into people who seem to be more in touch with parts of the community that are outside of my bubble.
This isn't my experience. These types of solutions may be generating a lot of interest in some places (perhaps because the people talking about them hope to profit?), but they're not all that exist.
Every system that is lacking this is doomed.
I use scuttlebutt because it's fun and nice. I'm not sure what more incentive is needed.
And I know that it works for you, but if we want something that can reach more than a few thousand people, there needs to be some kind of economic incentive for it to be viable.
The problem is that when you start measuring on "reaching more than a few thousand people" instead of "is creating a lot of economic incentive", cryptocurrency experiments around building communities also start to look pretty bad in practice.
I have yet to see any general-user social platform built on top of cryptocurrency amass the same amount of mainstream success as Mastodon. And Mastodon is comparatively tiny when stacked up against Twitter/Facebook, it should be the low-hanging fruit. But at least Mastodon is a platform that seems to have somewhat evolved past the initial "the only people using this are the people who built it or who are invested in it" phase. On the protocol level, look at stuff like Matrix; it's got very little incentive built in to get you to run your own platform, and yet... it still kinda looks like it's beating blockchain networks at being actually useful.
A lot of crypto spaces/experiments have a lot of money moving through them. But the ratio of money to users/disruption seems to be really low. So sure, crypto has an interesting approach to incentivizing economic investment, but at a certain point I start to question whether those incentives actually result in generalized, successful platforms that tons of everyday people are using. It unfortunately is starting to seem like the incentive structures attract a lot of capital and a lot of eyes that are either hostile to the overall goals of decentralization, or at least that the attention doesn't seem to be providing the kind of value/structure required to get non-invested, non-capital-focused general users to join in.
Mastodon (ActivityPub in general) is growing, but I'd say they are still far from being mainstream. When we start having companies using their own instance instead of Twitter/FB to promote their content and when we start having youtubers leaving Google to broadcast their content from their own branded Peertube, then I'd say it has hit mainstream.
> A lot of crypto spaces/experiments have a lot of money moving through them. But the ratio of money to users/disruption seems to be really low.
That comes with the territory in any hot market, and crypto is no exception to Sturgeon's Law. The main point that I would argue though is that is not just about the "money", but rather if these new systems can be sustainable and if they can create new alternatives to the status quo. The best example that I can think of is Brave and BAT. The BAT "economy" is still ridiculously small, but on a bet about the longevity of Brave vs Mozilla, my money is (literally) on Brave. Mozilla can have all the wishy-washy nice words about how they are protecting the web, but it doesn't take much to realize that they can not bite the hand that feeds them. Meanwhile Brave is already growing the user base on the browser and its search system will be a way for them to fight Google at their core. You will have soon lots of people making coordinated efforts to get more people off Google and into an alternative where they can benefit.
And it turns out that system (for all of its problems) is working better than blockchain for shipping actual decentralized software that ordinary people can use.
> Mastodon (ActivityPub in general) is growing, but I'd say they are still far from being mainstream.
And again, "far from being mainstream" is still miles ahead of any blockchain social network.
> When we start having companies using their own instance instead of Twitter/FB to promote their content
We do have this, Trump's recent venture into the social media space was built off of a Mastodon fork. It wasn't built off of blockchain.
> Mozilla can have all the wishy-washy nice words about how they are protecting the web, but it doesn't take much to realize that they can not bite the hand that feeds them. Meanwhile Brave is already growing the user base on the browser and its search system will be a way for them to fight Google at their core.
Brave is literally built on top of Chromium, it's literally Google's tech stack with some extra stuff pasted on top.
It makes sense why not needing to take money from Google would matter, but in practice, Firefox still ends up having better privacy controls and more power to oppose certain web standards purely because it has a different engine than Chrome. It turns out that in practice, architectural independence matters more.
This is another good example of the finances ending up mattering less than people think they do. Mozilla has money, that's how it's able to fund projects like Servo/WASM/Stylo, projects that Brave (for all its financial independence) doesn't have the developer resources or the architectural independence needed to helm. And the trajectory of those projects is worth its own conversation, but I don't think dependence on Google's money is the reason Mozilla had to eventually drop Servo, and I don't think that Brave is in a position to pick a project like that up.
----
You don't really have to convince me about an idea, you don't have to argue theory to me. The theory of financial independence is fine, I don't have a problem with that, I just don't see the theory playing out in practice. The blockchain doesn't need a theory for why it matters, all the blockchain needs to do is to launch projects that are better than the alternatives. So far it hasn't, which is a really low bar to clear. You're attacking Mastodon, how embarrassing is it then that there isn't a blockchain-based Twitter alternative that's bigger than Mastodon?
If you're right, and if a blockchain social network springs up that out-competes the other projects on the metrics that actually matter (not just money), then I'll start taking the space more seriously. But so far it just hasn't happened; and a lot of the promising projects that are closest-related to blockchain (even the really promising projects like IPFS, or the poster-child for this stuff, Handshake) have ultimately ended up being a lot more disappointing than the projects that have taken root in other spaces.
Sure, it makes zero sense to have a social network on a blockchain. No one really wants to have their social graph on an immutable public ledger that costs a lot of resources to keep around. But I wouldn't be surprised to see a project where crypto is used as a way to finance a Mastodon (or Pleroma, or whatever) powered constellation of servers.
This is the premise I'm questioning — what does success look like?
I suppose for me it's that I don't have to create an account with any particular company to participate.
So for example: Matrix displaces Discord and Twitch; ActivityPub displaces Twitter, Facebook and GitHub; podcasts displace iTunes and Spotify; Flatpak displaces Google Play; something(?) displaces Amazon (as a shop front) and PayPal.
> there needs to be some kind of economic incentive for it to be viable
There's no economic incentive for people to publish blogs or run Gemini capsules, yet they do.
By “viable” I mean “useful for some real use cases”. (Fun is a use case.)
By “viable” do you mean “people can make money off it”? Profitable technology is a very different goal from useful technology.
And frankly, get rich quick schemes are the reason the net is full of so much garbage.
So, success is quite measurable and it is certainly not achieved yet. This is not to say that we should just give up and be assimilated by the Borg. It is just a reality check and a reminder that there is a huge world out there.
I'm curious: why not? Is it just name recognition of WhatsApp? Are people unwilling to create a new account? and if so, what made them willing to create a WhatsApp account?
I know that the onboarding process from the Element client still is quite painful, and I didn't want to play tech support for a group of 30+ German parents and teachers.
> What made them willing to create a WhatsApp account.
The easiest authentication/onboarding process ever. That was the 19 billion dollar problem they cracked.
I like Andrew Yang's core policy proposals including Ranked Choice Voting, and most notable I believe is his Democracy Dollars policy - wherein every eligible voter gets $100/year voucher they can allocate to the political candidate of their choice, which would wash out the corporate lobbyist money of roughly $4 billion with 8 times that by citizens (including giving those who can't afford $100/year for an additional layer of "voting with money" to fuel/funding the political system), meaning $4 billion to up to $32 billing from eligible voters; https://andrewyang.com - Andrew Yang, ran in last Presidential race as a Democrat - polled higher than Kamala Harris and raised half that of Bernie Sanders, starting as an unknown - then ran for Mayor of NYC, and author of The War on Normal People, and most recently Forward: Notes On The Future Of Our Democracy - which in part was to announce the new relatively centrist political party he started called the Forward Party.
I feel another noteworthy policy proposal of his is Journalism Dollars, similarly where you give every individual say $50/year voucher to fund/support the journalist of your choice - so then journalism and news media isn't adversely incentivized by the advertising dollars coming in from industrual complexes (including from the political establishment/duopoly of the arguably captured Democratic-Republican parties).
There will not be one single standard of policies that will work for all people, no matter how "democratic" the process to establish the rules are and no matter how good the ideas you think they are. What we need is to reduce the scope and reach of governments, get their leaders to act on a localized level and to be directly accountable for the results of their policies. Anything else will take us to more centralization, more forced conformity, less diversity and more fragility facing global-scale events.
We need more Switzerlands and less USAs and even less European Unions.
No, it's not.
> History has shown time and again that the best we can do is some local maxima that works only for a part of the population at the cost of everyone of the periphery of power.
To the the extent that’s defensible, the problem with your recommendation:
> What we need is to reduce the scope and reach of governments,
Is that the evidence does not show that that produces better results for more people than than alternatives. You've jumped from a defensible argument about the limitations of government to an indefensible response.
> Is that the evidence does not show that that produces better results for more people than than alternatives
I am not defending localism as a way to "produce better results" for more people. I am defending it on the premise that it is a more robust idea and more likely to sustain long-lasting societies than any progressive "one size fits all" government.
You seem to have come to your conclusions referencing history, and not leading from a conclusion starting at founding principles.
The United States of America has been an experiment, with the basis of the Constitution by the Founding Fathers. It's done quite well in its first iteration. The next framework now needs to be laid out, but part of that is educating the population (local and globally, as each geography is going to be prone to the actions of bad actors) so then they can vote rationally, reasonably, and ideally not following or indoctrinated into ideology.
You seem to have a misconception or confusing what "government at a global scale" is, perhaps you're once again presuming the current status quo of generally captured and incompetence elected general officials (like the establishment/duopoly in the US) will forever be and so then yes, if you do the same thing over and over again (with incompetent and captured politicians in positions of power and others' captured leading our government institutions) then yes, that would be insane - but the status quo can change - and new ideas, the truth, is making its way to regular folks because of the reduction in capture of attention that the internet has allowed, and focusing on creating online platforms and systems to further strengthen education-learning, communications and reaching like-minded and like-hearted community is the key.
Yes, even if you don't realize it.
UBI, "Journalist Dollars", "Political vouchers" and pretty much every proposal from Andrew Yang are all starting to sound like variations of the old joke about the farmer getting in trouble with the inspector about the feed given to the pigs until he decides to just give each pig $10 so that they can eat at the local buffet.
They all involve blanket solutions without looking into any context or idiosyncrasy of any of the separate regions. It reeks of this coastal elite mentality that believes that people on the periphery want to live like those in the big economic centers.
> The USA has been an experiment (...) The next framework now needs to be laid out.
Read again and tell me how you are not implying that one framework is to be established and implemented. How are you not advocating a top-down solution?
> focusing on creating online platforms and systems to further strengthen education-learning, communications and reaching like-minded and like-hearted community is the key.
Yeah, that has been the technocrats wet-dream about the Internet. I've been hearing how the internet would help educate people all across the world since I got my first 14.4k modem. But instead of this utopia you are painting we got Donald Trump, Orban, the Chinese "social credit" system and a continuous detoriation of "Western democracies" that benefited only but the tech elite.
No, thank you. This is not the world that I want to my live in and my children to grow on.
Having a central government isn't inherently top down. You need an organization to do things like collect taxes. Are you trying to suggest that, for example, a central government collecting and redistributing tax funds say in the form of UBI is somehow top-down? In fact giving UBI is more consersative, less liberal, and more hands-off, as it isn't the government deciding what "you" do with that money.
I think you're missing some key understandings which is skewing your view, and you're certainly projecting hate based on your comments - "elite mentality", ". We got Donald Trump as a consequence to regulatory capture, MSM being captured, etc; for example, 10 mins. video of Noam Chomsky explaining how the Republican party was captured by corporations/bad actors - https://boingboing.net/2019/04/20/useful-idiots-r-us.html
Your argument too about "my utopia" I'm painting got "Donald Trump, Orban, [etc]" makes no sense either - I stated a new, evolved framework is needed, using what we've now experienced based off the Constitution, to avoid and counter the pitfalls that lead to a longer list than you presented. Maybe my responses could be more useful if you spoke specifically to why you dislike different policies of Yang, and how you understand them to work or not work? E.g. Don't you think washing out $4 billion of lobbyist money with $32 billion of eligible voter money would shift politics to getting more citizen-wants oriented politicians elected into power? If you say no, then why?
You've made/applied a lot of assumptions/inferences as to what to my comments suggest, your understanding seems completely off the mark.
Why would people even want to give this money to politicians? ANY of them? Could I simply get the money and pocket it? Why not?
More seriously though: why do you think is more likely to happen with such "journalist" money? Do you think people will suddenly promote rational discourse or they will just give this money to whoever confirms their biases? Imagine Alex Jones getting millions of dollars of government money to fund his lunacy. What will be the reaction on the liberal side? Will they just say "it's all fair" and try to compensate by over-funding NPR and PBS, or will they try to find ways to rig the game in their favor?
> You need an organization to do things like collect taxes.
That organization does not need to be centralized. As an example: You could have tax collection at the city/municipality level, and then have these municipalities paying to theirs state according to their pre-agreed budget and commitments, and having that bubbling up.
> Maybe my responses could be more useful if you spoke specifically to why you dislike different policies of Yang
It is not about the policy, it is about its scope and the idea that you can apply the same kind of thinking on a country as large, diverse and uneven as the US.
UBI in itself is not a bad idea. Thinking that the way to implement it is by pushing it at the federal level is a catastrophically bad idea.
I would give Yang a lot more credit if he went to run for mayor of a small town in the Rust Belt, tried his ideas there first and then went on to promote them. But when he goes on to think that he can run a country in the same way that people run a SV startup, big red lights should be flashing in our heads.
> Don't you think washing out $4 billion of lobbyist money with $32 billion of eligible voter money would shift politics to getting more citizen-wants oriented politicians elected into power? If you say no, then why?
"Because populism" is the short answer. The long answer is "if voters were rational, people would be able to filter out politicians who are not aligned with their interests and lobby groups wouldn't exist. Adding more money to the system will not solve the issue."
If we want to get politicians that do what the citizens want, we need to reduce the distance between the citizens and their representatives. How about we get rid of the $4 billion lobbyists AND the $32 billion "voter money"? How about we got rid of all forms of Federal taxes and started favoring again a state-centered union? How about people started to be more interested in discussing their own town budget plans (where they can actually SEE the results of their choices) instead of giving them money to only pretend they have any say in these multi-trillion sausage factories?
No, thinking that the way to implement it is by pushing it at any lower level is a catastrophically bad idea, because the 14th Amendment, the Commerce Clause, and other federal Constitutional and legal constraints prevent any lower level from establishing any viable definition of the covered population universe.
> How about we got rid of all forms of Federal taxes and started favoring again a state-centered union?
Absent any proposal for solving the problems that occurred the last time we tried that, which is why we have the Constitution (the second federal structure of the USA), that's a decidedly no serious proposal.
> How about people started to be more interested in discussing their own town budget plans
Proposals about what people should decide to be interested in are non-actionable.
As for "the problems that occurred the last time we tried that": one of the links on the front page from HN is exactly about how the US is closer to a civil war now than people think. If stability is the main function of the Federal Government, it looks like they are not doing a good job.
No, pointing out that we tried the extremely loose federation of states without federal taxes model, it produced real problems, and the model under the Constitution was an attempt to solve that, so if you want to argue for a return to that older model you need to acknowledge that, propose a fix within that model to the problems, or make an actual argument as to why those things that were seen as problems are actually desirable is not "the bureaucracy needs grow to satisfy the needs of the growing bureaucracy" circular logic.
History exists, your breezy offhand suggestions need to acknowledge and address it if you want then to be taken seriously.
Neither is pointing out that UBI by a entity structurally not empowered to exercise boundary controlon the population to which it is universally applicable, and therefore non-federal government in the present US Constitutional system, is problematic "the bureaucracy needs grow to satisfy the needs of the growing bureaucracy" circular logic.
> If stability is the main function of the Federal Government, it looks like they are not doing a good job.
Not that anyone has argued “stability is the main function of the Federal Government”, so that’s kind of a non-sequitur, but based on, and compared to, what exactly?
Again, no, it's not.
Global scale, one-size-fits-all government is the right-to-center-right, neoconservative/neoliberal colonial capitalist project.
The far left these days (Leninism having lost ground with the fall of the Soviet Union and it's sponsorship or allied groups abroad) is dominated by anarchist and libertarian socialist groups; and left-to-center-left groups are often milder versions of the same or differ from the center-right in the priorities of government, not its scale or scope. The far right may favor slightly smaller scale (large nation rather than global) government coordination, but tends to favor fairly intrusive scope of government and doesn't tend toward localism.
Equating maximum support for scale and scope of government with the Left may have made some sense when Leninism (and, despite “Socialism in one State”, it's Stalinist descendant, given observed Cold War behavior) were the most visibly dominant, notionally Left philosophies (though it still required ignoring the integrated colonial globalist nature of global capitalist ideology), but it's just plain looney-tunes today.
They generally support a similar level of scale and scope of government as the parties the same distance on the opposite side of the local center, with different goals. There's nothing particularly left, even by your definition of “left” that excludes pretty much everything past the center-left, about it. On scale and scope, but not purpose, of government there’s a fairly broad centrist common position.
As vague rhetoric, sure. When you get to brass tacks, those same Republicans tend to also to want federal control to prohibit state actions that they disagree with across a wide spectrum of things. The Republican Party, no different than the Democratic Party, favors state control on issues it feels it can't rally a federal legislative majority on, and federal control where it can and there are states that would otherwise defy it.
Are we defining "people" here racially, as a language group, or as the humans who live within the particular borders of an old colony or the territory of a long dead conquerer?
Why is Switzerland the ideal size? Does something about that size reduce the proportion of people at the "periphery of power"?
The USA was built on the same idea, where states have a lot more direct power than the federal government, but that idea is basically lost because people on both sides of the political spectrum have been pushing for more centralization and federal control.
"Determined" to be ideal? Found the totalitarian.
Nowhere did I state who's determining what's ideal, but you seem to have a need to demonize or put down - so you assumed it must be what I believe is ideal. I suggest you stop making assumptions, and in part based on your comments in this thread, that means broadening your perspective.
Ideally who's determining what's ideal is a democratically elected nation of people who have elected politicians to create policy that will reflect the wants, desires, beliefs of their population - from the local level and all the way up.
I didn't downvote your comment by the way, that was others - I don't downvote anything as I think downvotes in this context are overall harmful and a useless and lazy signal at that.
Network architectures are means, not ends.
The question is what are we trying to do?
Based on that context, we can determine whether a hammer, screwdriver, centralized architecture, decentralized architecture, or federated architecture is most beneficial for us. They all have tradeoffs (e.g. easier to remove screws than nails, easier to secure centralized than decentralized systems, easier to power federated systems than decentralized systems, easier to trust decentralized and federated systems than centralized systems) which can make them more or less useful for us in any given context. The point is, they are tools to help us realize our desired outcomes -- not the outcomes themselves.
There's some weird psychology of tech circles that keeps us focused on decentralization. Often it just seems like bikeshedding rather than problem-solving. My guess is the Social Media generation has become wary of people controlling their lives, preferring autonomy. They hear about network decentralization, and then come to the conclusion that it must be better than the alternative, because the alternative is centralization, which is Google and Facebook (but never Apple apparently, which is pretty ironic).
"A successful utopian society would most likely be one made up of 1000s of smaller utopias."
This strikes me as the valid mega-target that we should be attempting to reach collectively - space for all.
It should also be the foundational principle for software solutions. And neutral software solutions should also be able to accommodate people who don't want to be in the proposed solution!
That was the original rhetoric behind "Defund" IIRC. Break up the central police force to a rump used to coordinate and regulate police functions that have been devolved to exist under other government departments. Basically a bid to put a domestically-operating military under more thoughtfully-organized civilian control instead of simply having city cops, state cops, and country cops whose only points of oversight come respectively from the mayor, the governor, and the president.
Do not initiate harm. In reverse, do not roll over when someone is initiating harm against you.
* Security is wildly different in a decentralized model compared with a client/server model.
* Decentralization is about one thing only: maximum autonomy.
* Federated systems are semi-autonomous. They are a move towards decentralization but not completely so. With decentralization you only need an application that send/receive agreed upon instructions.
* Anonymity and privacy are opposing qualities. Decentralization is not about either but enables one or the other.
As I have been exploring this subject I have formed some criteria for decentralization:
* Address based (IP addresses or identifiers that resolve to an IP address without third party consultation, no DNS)
* Protocol agnostic (what ever two parties can agree upon)
* No third parties (no servers or message proxies unless you own them)
Think about it in terms of snail mail. The post office moves packages around without opening or inspecting the contents. The contents are never stored in a database and could be broken. It is address to address communication.
Decentralization allows for personal or organizational devices to be remotely connected as a large physically distributed computer:
* A single file system, cross OS.
* Remote application execution, such as Steam’s remote game play.
* Content on demand. Push media out to your friends or access their computer and pull it off yourself.
Reminds me of dialing into a BBS way back when, that early on was just someone with a non-dedicated or dedicated telephone line hooked up to their personal computer's modem; you could connect to fun, non-real-time multi-player text-based games (ASCII art FTW!) hosted on their PC, and some had file servers open for download and sometimes upload.
However my database is address based, it reads and writes blocks of encrypted data, and it is protocol agnostic, it works over S3 and Google Drive, or could work over a pendrive sent in a letter back and forth, or IPFS. It just requires a single source of truth and optionally event notification.
So would this fit into an every peer-to-peer system is decentralized, but not every decentralized system is peer-to-peer picture, or is this not decentralized at all, because you would normally use it with Amazon S3, albeit in a maliciously-secure manner (the confidentiality, integrity and authenticity of your data is secured via authenticated encryption, the client keeps the key, the access pattern is obfuscated).
In the case of your scenario it could be considered decentralized as you are using cloud storage as merely a storage point, as in a dedicated redundancy, but I am hesitant to consider a dedicate server as a point of decentralization regardless of its intention or access limitations.
https://en.m.wikipedia.org/wiki/9P_(protocol)
Here is the approach I am taking: https://github.com/prettydiff/share-file-systems
maybe instead of mp4 files no one bothers with, upload your demo on peertube or similar?
I think, that hits the nail on its head.
Many people are happy with MANGA basically owning the Internet and many people aren't.
If we look at companies bending the knee for China or shutting down accounts, simply because they have different moral than the account owners, it's clear that we need much more decentralization than we have right now.
How much? I don't know.
Data mobility could also perhaps be considered the other side of the decentralized coin.
The key is that these companies are very conservative when blocking users and giving others special privileges. Your service won't be blocked by Cloudflare and IPs unless it's particularly bad or your government requests it. Whereas e.g. YouTube has very strict rules and often bans people incorrectly or for debatable reasons.
Idk if the issue is actually decentralized vs. centralized. I think it's "do I decide the rules or do I just enforce them?"
This is a static view. A dynamic view would recognize that systems have a lifetime and tend to go through centralize and then disintegrate. Decentralization often exists prior to centralization. Catch systems on the rise.
5.2. Encrypt, Always
When deployed at scale, encryption can be an effective technique to reduce many inherited
centralization risks. By reducing the number of parties who have access to content of
communication, the ability of lower-layer protocols and intermediaries at those layers
to interfere with or observe is precluded. Even when they can still prevent communication,
the use of encryption makes it more difficult to discriminate the target from other
traffic.
Note that the benefits are most pronounced when the majority (if not all) traffic is
encrypted. As a result, protocols SHOULD be encrypted by default.
All systems that scale HTTP-based applications require a lack of encryption on the backend. The encryption is terminated at the load balancer and the entire connection is then inspected in plaintext. Sure, the "best practice" is to encrypt between endpoints, but they must be decrypted at each hop so that load balancers, API gateways, and applications can inject or parse headers, perform rewrites, forward traffic, inspect requests (WAF), create sticky sessions, and cache media. If all HTTP connections were end to end encrypted, the web could not scale.
Another example is DNS over HTTPS. The encryption centralizes control of the protocol at the server. The encryption removes the ability of intermediate LAN/WAN DNS resolvers to cache records for a large local network, or to serve their own custom zones transparently, or do things like ad-blocking. As there are valid and necessary use cases, it means that now new "hacks" will have to be developed to accommodate the old use cases, that the protocol design intentionally ignored. The result is more centralization.
Distributed decentralized applications also suffer when PKI is used for encryption. It is very common for backend applications to stop working because they were using a private CA to encrypt backend connections and the certs expired without warning, killing all the connections. Refreshing the certs then involves re-deploying not just the servers, but all the clients that were reliant on an old client cert, and possibly an expired CA cert. This use of a private CA also breaks decentralization compared to the public CA model.
Decentralization fundamentally trades inefficiency for long-term robustness. Centralization trades long term robustness for short term efficiencies.
Under our implementation of capitalism, most people don't value robustness on a timescale longer than 20 years unless there was recently a crisis to remind them of it.
So every aspect of human infrastructure oscillates on the centralized-decentralized axis over time. Thinking of "decentralized" as something that will stay that way is wrong, over time people will ever centralize the system in pursuit of short term returns until the perceived "decentralized" system might as well be centralized in practice (see internet infrastructure), once a crisis happens due to over-centralization we build a new system (or re-brand an old system) that is more decentralized in response. That system eventually decays into centralization the same way.
I value changing society to think in the long term, so I think it is a lot more valuable to change the cycle versus attempt to participate in just one swing of it.
https://www.unturf.com/let-us-define-a-permacomputer/
I avoid writing prescriptively to avoid setting a bias on the talent.
I AM growing my own "YouTube". Not open source yet but I'm planning big things.