1. 15
  1.  

  2. 4

    In case anybody’s curious, I’ll happily name things that, unlike Diaspora, are actually a threat to the adtech industry:

    • Anti-trust regulation. This is the most important one.
    • Ad blockers and neutralizers like Tor and Archive Today. This is less important than anti-trust regulation, but still, directly attacking the revenue stream is obviously a threat.
    • Communities like Something Awful Forums and Lobsters, that use barriers-to-entry that ad sites won’t replicate because it would cut into their profits. This is the least important part, because the barriers to entry that make them impervious to takeover also prevent them from actually competing for attention, but still, if you’re looking for a value add, there it is.

    I specifically don’t list federated systems like email and gnu-social, because an ad-based business can just join the network and dominate it with their money. Which is kind of the point of the article.

    And I don’t care about invite-only groups that Facebook and Reddit have. I’m sure the companies benefit from having that feature, but they don’t push them because there’s not enough network effect in offering small forum hosting. Whoever’s in charge of the group can just decide to switch, and the other members will go along with it because it’s the leadership, not the platform, that delivers all the value in a group like that.

    1. 3

      The author takes a ‘soft’ perspective on what counts as decentralized – one that I consider inherently doomed. Specifically: compatibility with existing web standards locks us into at least one particular kind of centralization – the address of data contains information about its hostname, and this name is either permanently associated with a machine or must periodically be reserved by a maintainer entity, who is on the hook for the content. In other words, accepting HTTP means locking into a situation where somebody needs to keep paying money to prevent links from breaking.

      A fully decentralized hypertext system (I refuse to call it a “decentralized web” because “web” implies web standards and web standards are inherently centralizing) is basically incompatible with economic-centered analyses: so long as using it costs anything, people will use it in to produce economic monopolies, incentivizing flooding with low-quality content, and trying to disrupt other kinds of use.

      Luckily, software is uniquely well-suited to post-scarcity situations: it has near-zero reproduction cost and so marginal (economic) value of software trends toward zero. The use of hostnames enforced by the url scheme of web tech constitutes a kind of artificial scarcity: an identical piece of data hosted elsewhere is considered non-authoritative. Content-addressing solves this problem, making expensive centralized hosting totally unnecessary while also removing the possibility of centralized analytics and arbitrary modification.

      Having a properly decentralized hypertext system involves throwing away a lot of things that seem like common sense to most people: hostnames, mutability, centralized corporate control, monetization, data centers, and the possibility of deletion. I don’t think any of these things are actually as valuable as we make them out to be: having become unnecessary, we (as users and developers) won’t miss them, although they (as entrenched nodes in a hierarchy of power) will miss us.

      1. 2

        hostnames, mutability, centralized corporate control, monetization, data centers, and the possibility of deletion

        BitTorrent showed how to build a storage system that gave up all those properties. It was actually pretty successful, until NetFlix and Spotify completely out-classed it for usability. As the article itself alluded to, a centralized corporate thing can be distributed-under-the-hood, but a decentralized system cannot be centralized-under-the-hood. From a technical standpoint, anything we can do they can do better (that is, they can do anything we can do, but we cannot do anything they can do, so presumably they’ll exploit that increased freedom to build a better product).

        Also, data centers are valuable because, while bandwidth and storage are relatively cheap, uptime is expensive and valuable. If you want people to form habits, you need things to work most of the time.

        I don’t think any of these things are actually as valuable as we make them out to be

        They don’t have to be base-breakingly valuable. They just have to be valuable enough to tip the scales on what are otherwise similarly good products. When Popcorn Time requires you to install an app, while YouTube only requires you to visit a URL, and the experience is otherwise similar, people will pick YouTube.

        1. 3

          data centers are valuable because, while bandwidth and storage are relatively cheap, uptime is expensive and valuable. If you want people to form habits, you need things to work most of the time.

          The uptime of a fully-distributed fully-replicated chunk of data is always going to be better than the uptime of a piece of data that exists in an authoritative way only on a handful of centrally managed racks, no matter how expensive the racks are.

          When Popcorn Time requires you to install an app, while YouTube only requires you to visit a URL, and the experience is otherwise similar, people will pick YouTube.

          Youtube essentially installs an app (running in the browser sandbox). If downloading complicated code sight-unseen is necessary to gain users, there’s no reason that we need to use hostnames to do so (and if we use proper DHTs, we get correctness checks for free, while web apps remain vulnerable to injections of arbitrary code).

      2. 0

        I am not pro or contra centralization per session but the author misses the point that economics is largely a function of technology - specially if we view institutions as social technology