1. 7
  1. 4

    Advocacy is boring and it doesn’t work. It’s the OSS equivalent to voter canvasing, but without the sanctimonious halo of “your vote counts”, and with a much greater maintenance cost per-user (“visit the polls once every 4 years” vs “use only GNU software”). “Decentralization by volunteer vigilance” won’t work any better than “code hygiene by vigilance”.

    Incentives work very well, and decentralization enthusiasts (in which I count myself) should focus on that problem. The technical problem is much less interesting than the incentive problem. (By the way, that’s also why bitcoin was interesting, and the people who complain about its energy use are very much missing the point.)

    The most interesting incentive-based distributed approaches known to me are brave and urbit.

    1. 3

      I’m a huge fan of decentralization, and there are some really exiting options out there right now such as PeerTube, Pixelfed, Mastodon, and others. Having said that, I’m not convinced the author of this particular piece means “decentralized” in that particular manner. It seems like Storj is saying “Decentralized” when they actually mean “Geographic Diversity”.

      1. 3

        Geographic diversity is decentralization of physical equipment, and that’s perhaps not a bad place to start. But, yeah, by that definition Google, AWS, and Cloudfront are all highly “decentralized”.

        It’s hard to advocate for decentralization of ownership (or more operatively, control) without coming across as ideologically motivated. But by avoiding discussion of ideologies, we give tacit support to the dominant one.

      2. 2

        The first link in the article points to a source that misrepresents facts. The link title is “Today, Google and Facebook control over 75% of all internet traffic.” The article pointed to confuses Internet traffic, Web traffic, and a statistic of referals as measured by a set of online news publishers. This news article points to a random blog on the Internet, that is dubious in terms of its many claims, since there are no clear arguments that separates fact from hypothesis.

        1. 3

          Further, the ones controlling the traffic are ISP’s. They can literally disconnect people from Internet. Google and Facebook are services people vountarily use which have alternatives.

        2. 2

          “A decentralized system has no central point of failure and is less likely to fail due to power outages or hardware failures.”

          Whereas, it’s more likely to fail due to complexity, other users’ actions, and/or bad business model. The quoted examples supporting centralized as risky are a chat app that’s not high-uptime design and a massively-complex, cloud platform. That’s misleading. Compare to designed-for-reliability solutions like VMS or QNX clusters that stay up for years to decades while running efficiently on commodity hardware thst improves each year. Decentralized offerings dont even compare.

          “collusion resistance found in a decentralized system also makes it harder for participants to conspire against other participants for self-benefit”

          They just rob the exchanges instead which currently fail security at higher rate than centralized banks.

          “the state has exercised its right to target and incarcerate ethnic and religious minorities and political dissidents. “

          Which it can do with most decentralized systems, including to their operators or users if they become a huge problem.

          “Decentralization allows for market forces to optimize the network at a greater rate than any single entity could afford”

          I dont know if this one is or isnt true. Gotta see what people do in general: use decentralized networks or traditional ISP’s. The latter. Also, oligopoly and monopoly pricing got them tens of billions in profit. Im not sure there’s much they cant afford in centralized model.

          “At Storj, we are building a decentralized cloud object storage network to replace traditional centralized data centers.”

          Now that sounds interesting and maybe even doable for some chunk of the market. I look forward to the benchmarks on speed, space, energy, and cost vs using multiple, centralized providers with regular protocols.

          1. 1

            I’d like the benefits sans the inane ideology, thanks.

            1. 1

              Thank you, company selling decentralization