Skip to main content

Innovation dynamics in the metasystemic stack

2 min read

Joi Ito expresses some misgivings (far milder than my own) about "the Bitcoin community", and along the way provides this gem of a case-study:

One of the key benefits of the Internet was that the open protocols allowed innovation and competition at EVERY layer with each layer properly sandwiched between standards developed by the community. This drove costs down and innovation up. By the time we got around to building the mobile web, we lost sight (or control) of our principles and let the mobile operators build the network. That's why on the fixed-line Internet you don't worry about data costs, but when you travel over a national border, a "normal" Internet experience on mobile will probably cost more than your rent. Mobile Internet "feels" like the Internet, but it's an ugly and distorted copy of it with monopoly-like systems at many layers. This is exactly what happens when we let the application layer drag the architecture along in a kludgy and unprincipled way.

Historically, the application layer of a network system pretty much always drags the architectural layer, because the application (or interface) layer is governed by commercial incentives to innovate; those commercial incentives may result in improved functionality, but they are just as likely (if not depressingly more so) result in the appearance of improved functionality (which is a very different thing, and sometimes the exact opposite).

This isn't to say that the architectural (or infrastructural) layer has no influence in the other direction, of course, but infrastructure is by necessity a very slow game: big-ticket projects on the largest of geographical scales. The interface layer is inevitably more nimble, more able to iterate quickly; when the interface layer in question is pretty much pure software (as in the example of the blockchain), that is even more the case, because the opportunity cost of iteration and testing is so low, and the potential rewards so ridiculously high. (However, the infrastructural layer is far from innocent, as the battles over Net Neutrality indicated very clearly.)

As Ito indicates, and historical evidence supports, open protocols and shared standards between sociotechnical systems lower costs and open up the field for innovation to *all* players in the stack, not just to the interface developers.

That alone should tell you exactly why Silicon Valley dropped the Open Web.