Back to basics: what is the point of decentralization?
paulfrazee.medium.com
paulfrazee.medium.com
This is an important albeit abstract essay. At points it gets into Bitcoin although that's not the most interesting stuff; as someone who was closely involved in that space I'll comment on a couple of things before getting to the real meat of what he's writing about.
Paul says: "I have not yet been successful at enumerating the political systems of Bitcoin governance, but I strongly believe it exists because otherwise the protocol would never have changed" and also "It’s the wide decentralization of Bitcoin’s operation which produces the outcome they desire — which is not to change the protocol. (Or, at least, not often, and not without very wide consensus.)"
Paul finds it difficult to figure out how Bitcoin is governed for the same reason figuring out where power truly lies in a communist state is difficult. Or for that matter, the same reason figuring out who decides what goes into HTML5/browsers is difficult (can anyone name the Chrome team member who has the final call on this?). In reality Bitcoin is not governed by any form of distributed consensus, it's "governed" by an extremely small, murky and poorly defined clique of developers that systematically erases any discussion of any protocol changes they disapprove of from public forums, and keeps most discussion of those they do in private forums. That's why it's so hard to discover how it works.
This clique claim to be ideological defenders of protocol stasis, in much the same way as communists claimed to be defenders of the working class. When you look at what they actually do, this belief becomes untenable: the Blockstream employees who took control of Bitcoin around 2015 immediately embarked on complex protocol changes like segwit, in order to build the Lightning Network, an even bigger change which more or less replaces Bitcoin for actual usage by end users with not only a different protocol, but an entirely different design that bears no relation to blockchains at all. Indeed in my view it even fails to meet Satoshi's original goals for Bitcoin, but whether you agree with that or not, it is clear that LN is a very large protocol change from the practical perspective of people writing wallets, and any objections to the contrary rely heavily on definitional/semantic game playing.
Thus Bitcoin is really the leading example of exactly how badly we're failing to implement collective governance for notionally collective decentralized systems. Probably because that community ignored the lessons of history, it ended up collapsing into a Soviet-esque form of authoritarian dictatorship, complete with revolutionary myth making.
And it worked! Paul's belief that "If some group of people got together and conspired to push through a change to the protocol, they’d need extremely wide agreement by the users to accomplish it." isn't really true, but rather what Bitcoin propaganda claims is true. In 2015 when I left the project I described in great detail several such conspiracies which successfully did force through major protocol changes against enormous opposition, by exploiting the fact that there were very few miners, they were almost all in China and those miners were terrified of doing anything that looked like a "split" or "leaving the consensus", which they defined as "whatever Bitcoin Core ships". Thus whoever managed to take over the bitcoin.org domain name and release process could make any change they wanted on the majority chain regardless of what users or even the miners wanted, and combined with very large DDoS attacks on any firm or node that didn't run Bitcoin Core, and you have a successful takeover. [2] [3]
Moving on from Bitcoin, Paul starts by talking about how the web won partly because it didn't try to integrate backlinks to the protocol, allowing for more decentralised operation. But then towards the end he mentions that he didn't like P2P twitters because he was worried he might not see all the replies (i.e. backlinks). So I wonder what to conclude from this. If it were really so important, wouldn't Xanadu have won? Wouldn't pingbacks and backlinks be much more important than they are? It makes me wonder if maybe we're mis-understanding the success of the web over other systems, as apparently finding replies is important, but the web hardly supports this and even using search engines is kind of ropey for this use case.
Indeed from a quick check of the Xanadu wiki page we see that despite originating in 1960 it apparently never actually shipped anything until 1998 and "a working deliverable" wasn't implemented until 2014. According to a writeup from 1995 [4] "Even today, the technology to implement a worldwide Xanadu network does not exist". So it may have been that Xanadu's ideas actually were preferable, but Tim Berners-Lee and the people he attracted actually executed instead of theorizing all day.
The same contradiction seems to arise in the final paragraph where he says, "I think it’s meaningful that there aren’t any popular, traditionally-designed social networks which only show data from your local social graph like the p2p social networks do". But doesn't this describe Facebook? You only see content posted by your friends. The UX is p2p-like even though the implementation isn't.
I feel like this is an important type of thinking to engage in, but I'm left not entirely sure what to conclude.
[1] https://beakerbrowser.com/
[2] https://blog.plan99.net/the-resolution-of-the-bitcoin-experi...
So the same could be said of what we need in consensus: sometimes the right move is to turn to an authority to get an answer, other times we coordinate one out of a network. Twitter is doing some of both to be a useful service.
I have a creeping suspicion that ML and distributed ledgers go together in that one can be the flip-floppy complement of the other - ML as agent and oracle, ledgers as a mechanism to build shared knowledge among agents.
(I am also of the opinion that we are hopelessly clueless about how to apply these technologies well right now. But even if that means it is foolish to try when it results in so many failures, I think joining the fools is a noble idea.)
A lot of your observations on Bitcoin’s governance checks out. I stand by my theory about consensus-based tech driving consensus-based politics; the control over bitcoin.org, protocol discussions, and the core implementations might represent capture over the levers of that political power. If the Bitcoin community accepts that arrangement then it counts as legitimate political authority, but as you say the lack of transparency about that process suggests it’s dishonest. I get the sense that the btc community is ideologically inclined to pretend no such political system exists, and that’s the cloud cover that the developers operate under. All part of the reason I want to push at this topic.
Your observation that Facebook meets my description did occur to me, but I chose not to elaborate and instead let “like the p2p networks” do the heavy lifting for me (whoops). Facebook works that way for a lot of its interactions, but it leverages a lot of global collectivist resources as well like search, metrics, and anything community-oriented, while the anarcho-p2p networks use none. The difference between the product designs of (say) SSB and Facebook may therefore seem minor, but the philosophical resistance to anything collectivist in SSB is huge, and that’s what I think needs pressing against.
On the Web vs Xanadu questions you raise, the Web discarded all distributed collectivist features except for the ones I enumerated (DNS, IP addresses) and then used distributed individualism and authoritarian collectivism to handle the rest of its features. The Web did get those features in a “good enough” way though: in the authoritarian collectivism form. Decentralization is revisiting that approach and butting up against this challenge of solving distributed collectivism now, and I think the tendency of decentralizers to think in techno-anarchic forms might be causing them to miss this important area of progress.