Can we talk about the use of the word 'decentralization' in context of crypto? I think it's deeply misused because these systems are the opposite. They're global state, single sources of truth. Distributed, not decentralized. This is not just semantics but explains a lot of their features, like the reason they can be a currency, or why they're so slow. (pushing everything through a global state machine is a lot of work).
Decentralized systems matter but they do away with a single source of truth, that's what makes them responsive. They have local consensus, not global consensus. Torrent networks are decentralized, protocol based chat apps are (riot/matrix), Mastodon is, and so on.
When articles like this say that performance of crypto networks is a 'technical problem' to be solved I don't think they understand that it's not a problem but a necessary and intrinsic feature because to maintain global consensus in the absence of an authority you need to impose costs on the users of the network. That's why crypto systems aren't comparable to the decentralized web.
You may be thinking of “federation.” Ethereum is both distributed and run in a decentralized fashion. It uses something like the GHOST consensus protocol. Nodes don’t need to trust each other.
If they were just a state machine, they'd be useful.
They're meant to be a pseudo physical state machine where changing the state takes real work. If you decoupled the needless work aspect, you'd probably end up with a useful value proposition (which no one would adopt)
I appreciate a good argument supporting decentralization, but promoting cryptonetworks weakens it. From Adrian Colyer's paper [0]:
> The top four Bitcoin miners combined have more than 53% of the average mining power. In Ethereum the top three have 61% of the average power. Miners do change spots in the rankings over the observation period, but place is contested by only a few miners.
Well it's the only one that's really taken off. The fediverse started to get bigger until mid-sized nodes started making ban lists and the largest nodes left, meanwhile nothing new is happening with torrents and chrome killed off native peer-to-peer libraries when they started being used by YouTube competitors
As much as I personally dislike how crypto has been pitched it's the only field where decentralization has really entered the mainstream
The things you list seem actual examples where decentralization did take off, but faded once people recognized the tradeoffs of other features implied by decentralization and the mainstream audience chose that they do not really want decentralization - i.e. it's not that it did not yet get a chance to take off, but rather that we-the-peopled tried it out and practically found out that for most use cases we prefer centralization.
I’ve tried to get into the fediverse many times. Every time the thing that stops me is stupid little problems of the sort that keep me in the Apple ecosystem instead of using Linux. I just keep running into confusing time consuming little issues and while I certainly can figure it out I don’t have time.
None of the issues I’ve found are intrinsic to decentralization in any way. They’re just UI/UX design issues. Nothing about the federated or decentralized design would prevent them from being fixed.
Federation is more fun IMO than the antagonistic, “trustless” crypto ecosystem. It’s also generally cheaper to participate in, whereas crypto nodes and miners will only get more expensive as it scales.
Hear, hear! I see no reason why can't we go back to sharing information across sites on many domains hosted on a number of small servers, like in the 90s. For example, having a Raspberry Pi in every home. This sounds like decentralization to me.
Seems feasible. As Raspberry Pi 4 with 8GB ram, 64GB flash, ethernet (for a DSL or cable modem), and wifi could meet most peoples needs for file sharing (like dropbox), gmail like web email, chat (matrix or similar), etc. Trick is it needs to be robust, trivial to use, and pushed by a well recognized name to get decent market share.
Ideally they would find each other via a DHT and even offer to trade services (storage, backups, replacing a down service, etc.)
Having these in a decent fraction of households would help mitigate the usual downsides of P2P (slow transfers, slow startup time, etc).
> Centralization has also created broader societal tensions, which we see in the debates over subjects like fake news, state sponsored bots, “no platforming” of users, EU privacy laws, and algorithmic biases. These debates will only intensify in the coming years.
This doesn't make any sense to me. It sure looks like motivated reasoning. What's the connection? The fewer sites there are, the more fake the news there will be? Why? State sponsored bots benefit from there being fewer platforms? When everyone had a blog, were blogs ever known for keeping out comments from bots? EU privacy laws only apply to large centralized sites? C'mon. I can see “no platforming” of users and algorithmic biases being related to centralization at least.
Perhaps they mean the _debate_ over these subjects (fake news etc.) is exacerbated by the existence of a centralized authority to appeal to to remove it. Before Facebook I don't recall much debate over removing something from the broader, decentralized internet.
That makes sense, thanks. There were DMCA takedowns before centralization, and there was certainly debate over SOPA/PIPA though. Other illegal content was taken down in the past too. (Libel, exploitative media, etc)
I think the difference now is that we've widely experienced actual real-life impacts of unfiltered & unregulated but widely circulated speech. Attacks by incels and dis/misinformation about COVID come to mind. Many people now realize that they and/or society at large been exposed to and tricked by malicious actors, or at least seen their effects. There have been several protests stirred up by foreign operatives for example. IMO that shouldn't be legal, but in a way it's another form of yelling "fire" in a crowded theater that we have to grapple with defining and regulating.
Imagine your friend is running their own social media website, a workalike of MySpace or Facebook, and you're an active user. This is a small community space with a few dozen or a few thousand users. Everyone is having a great time, the application is customized to your liking, and you have access to the operator, meaning if your account gets accidentally locked, you have someone to contact.
Now, imagine your data on this community space is freely available for you to download and back up -- for the entire public data store of this social network, in a format which would allow you to spin up an exact clone of the existing space, including content, votes, threaded conversations, timestamps. Yes, even user accounts, because they are backed by PKI.
This type of setup gives everyone involved an enormous amount of agency, consent, transparency-visibility -- something no one today comes even close to offering. Imagine if you could download a zip file of your entire Reddit account, the full threads of each discussion you've participated in, all your submissions, and start your own Reddit, and everyone already has an account on it.
The benefits of this type of setup are too many to list in this one comment. It's a true revolution. Imagine your social network changing something without your consent, and you having the option to pack up and go!
Not only that, but at any point you can re-validate or inspect everything that's happening, run your own data analysis on the social graph, do retroactive bot detection, etc.
This sounds pretty similar to what we've done with Peergos [0]. It's a p2p e2ee social filesystem. You can migrate servers unilaterally without losing any data, whilst keeping all your social connections, unlike email, and with all existing links to your stuff continuing to work. The social media aspect is discussed here: https://peergos.org/posts/decentralized-social-media
We take adversarial interoperability with ourselves to the next level.
This was written by a venture capitalist who makes a tonne of assumptions about the future:
"The internet is still early in its evolution: the core internet services will likely be almost entirely rearchitected in the coming decades. This will be enabled by crypto-economic networks"
There are definitely a lot of problems with the current climate of tech monopolies, but the issue they describe is not related to them being centrally managed (see why decentralisation section).
I'm glad the author admits that web3 doesn't currently work due to fundamental issues with performance and scalability.
Fix that, and then lets see what you have to offer. Right now, people are being duped out of their money on a daily basis. Any gains made through specalative crypto investments are someone elses loss.
Decentralized systems matter but they do away with a single source of truth, that's what makes them responsive. They have local consensus, not global consensus. Torrent networks are decentralized, protocol based chat apps are (riot/matrix), Mastodon is, and so on.
When articles like this say that performance of crypto networks is a 'technical problem' to be solved I don't think they understand that it's not a problem but a necessary and intrinsic feature because to maintain global consensus in the absence of an authority you need to impose costs on the users of the network. That's why crypto systems aren't comparable to the decentralized web.