This is the thousandth time I've clicked on a "crypto" article expecting it to be about cryptography, and instead it's about blockchain/cryptocurrencies.<p>Is this a losing battle?
I've lived through a few eras, where open and/or decentralized systems were expected to win (Internet, crypto today). For most users, the critical things are usability and utility, not decentralization.<p>If we care about decentralization, we have to care about these points even more, to make products competitive with what users are used to. Even though many engs I know value decentralization, most people won't choose systems for this reason alone.<p>Just like with Github, there was a bunch of anger before when Slack was displacing IRC. I wrote some thoughts about what we/I could learn from that example:<p>What Open Source Can Learn From Slack<p><a href="https://www.nemil.com/musings/oss-and-slack.html" rel="nofollow">https://www.nemil.com/musings/oss-and-slack.html</a><p>If you really want to get more cynical, Tim Wu's book "The Master Switch" is a masterful look at 20th century technologies (radio, telephone, telegraph) going through the idealism of the early days, to the inevitable frustration when it creates new anti-consumer behemoths.
I expected this to be dumb, but it's a very perceptive article that makes a very good point: There are a lot of benefits to centralization, and people who want to create decentralized systems need to recognize that and take the necessary steps to ensure their systems do not become re-centralized.
"With this new set of powers, we saw a Cambrian explosion of open source that coincided with git's adoption. [...] That's where GitHub stepped in, providing elegant, centralized solutions around all of these new problems."<p>Did the author live in a parallel universe? There was an insane number of small open source projects way before git was created. And there was a clear place where to find them: SourceForge.<p>The cool things about git are the ability to sync repos and the ability to handle merging branches way better than SVN.<p>Everything else we had. Sourceforge dropped the ball and that allowed github to take its place.
This article would make more sense if crypto currencies were actually decentralized... that ship sailed maybe five years ago? Today, only a handful of corporations control the majority of the proof-of-work hashrate [0]. What’s more, development teams represent another centralized group (ie. you can count the devs that contribute to consensus code on 2 hands if you’re generous).<p>So my question is: in what way are crypto-currencies not already in the state the article is warning about? If I capture the lead commiter to your chain as well as the lead miner, don’t I effectively own the chain?<p>[0] “Both Bitcoin and Ethereum mining are very centralized, with the top four miners in Bitcoin and the top three miners in Ethereum controlling more than 50% of the hash rate.” <a href="http://hackingdistributed.com/2018/01/15/decentralization-bitcoin-ethereum/" rel="nofollow">http://hackingdistributed.com/2018/01/15/decentralization-bi...</a>
This was actually a great article. So much crypto centered content nowadays is garbage. Very pleasantly surprised by this. I think there are even a lot of interesting related questions that were not explored in this relatively concise post.<p>Though I don't have any problem with them today, services like Infura are definitely analogous to GitHub and may eventually become weak points in the same way.
Very insightful article, thank you.<p>So the cycle is:<p>- problem exists<p>- quickly centralized solution to problem appears<p>- new problems appears<p>- slowly decentralized alternative solution to old problem appears<p>- repeat<p>I would add that it seems that the big winners are the ones coming up with centralized solutions. The inventors of decentralized solutions don't have nearly as great rewards.
I am really upset about this acquisition.. And equally or more concerned with centralized power points, maybe to an extreme that is too radical. But I do believe at it's core, crypto is different because it's core ethos is to reject that notion. So while the internet disrupted centralized broadcasters with a decentralized protocol for content distribution, it wasn't baked into the technologies ethos. So in my opinion, this difference of core mission, along with the ability to launch dApps (anonymously even, if you'd like) without fear of being shut down, will be a differentiation that protects this technology's fate from creating new centralized choke points.<p>Ie. if Sean Parker was driven by the mission/ethos of crypto, Napster couldn't be taken down, people would still be using it today, and the traditional music industry would be dead (as well as traditional tech companies like Spotify).
This happens because
`distributed < distributed + centralized`
nearly always.<p>For example, git is great, but git plus an (optional) issue tracker has more value, so that's what people use.<p>Coincidentally (or perhaps causally?), FOSS < FOSS + closed source.<p>The only way I can see this changing is if the distributed (or free or open-source) benefits of the system fail early and painfully when centralized (or paywalled or close-sourced). I can't think of any such systems off the top of my head.