[This is the first of a series of excerpts from my book-in-progress, tentatively titled Desktop Regulatory State (click the link for a free online version of the latest manuscript]
II. Distributed Infrastructure and Ephemeralization
The larger and more hierarchical institutions become, and the more centralized the economic system, the larger the total share of production that will go to overhead, administration, waste, and the cost of doing business. The reasons are structural and geometrical.
At its most basic, it’s an application of the old cube-square rule. When you double the dimensions of a solid object, you increase its surface area fourfold (two squared), but its volume eightfold (two cubed). Similarly, the number of internal relationships in an organization increases as the square of the number of individuals making it up.
Leopold Kohr gave the example, in The Overdeveloped Nations, of a skyscraper. The more stories you add, the larger the share of floor space on each story is taken up by ventilation ducts, wiring and pipes, elevator shafts, stairwells, etc. Eventually you reach a point at which the increased space produced by adding stories is entirely eaten up by the increased support infrastructure.
The larger the scale of production, the more it must be divorced from demand, which means that the ostensible “economies” of large batch production are offset, and then more than offset, by the increasing costs of finding new ways of making people buy stuff that was produced without regard to preexisting orders.
The society becomes more and more like something out of Brazil or The Feds in Neal Stephenson’s Snow Crash, and the distribution of occupations increasingly resembles the demographic profile of the promoters and middlemen in that crashed spaceship in A Hitchhiker’s Guide to the Galaxy, who founded the human race on Earth.
The only way out is a new standard of progress that doesn’t equate “growth” with larger institutional size and more centralization: scalable, distributed infrastructure, stigmergic organization, module-and-platform design configurations, and production capacity sited close to the point of consumption and scaled to demand.
Paul Hawken and the Lovinses, in Natural Capitalism, stated the general principle that when load-bearing infrastructures are built to handle the load at peak demand, about 80% of the unit cost comes from the added infrastructure that comes from the 20% increased usage during the tiny fraction of time when infrastructure experiences peak load. They gave the specific example of home heating, where enormous savings could be achieved by scaling capacity to handle only average usage, with additional demand handled through spot heating.
More generally, centralized infrastructures must be scaled to handle peak loads even when such loads only occur a small fraction of the time. And then they must amortize the extra cost, by breaking user behavior to the needs of the infrastructure.
At the opposite pole is distributed infrastructure, in which most of the infrastructural goods are distributed among the endpoints, relations are directly between endpoints without passing through a central hub, and volume is driven entirely by user demand at the endpoints. Since the capital goods possessed by the endpoints is a miniscule fraction of the cost of a centralized infrastructure, there is no incentive to subordinate end-users to the needs of the infrastructure.
The classic example is Bucky Fuller’s own: the replacement of the untold millions of tons of metal in transoceanic cables with a few dozen one-ton satellites. The entire infrastructure consists of satellite dishes at the endpoints commuinicating — via free, immaterial ether! — to the satellites.
Likewise projected systems which replace the fiber optic backbone with satellite connections and last-mile meshworks.
Also the enormous infrastructure tied up in the civil aviation system’s central hubs and batch-and-queue processing, as opposed to small jets flying directly between endpoints.
Another example is mass-production industry, which minimizes unit costs by running its enormously costly capital-intensive machinery at full capacity 24/7, and then requires organizing a society to guarantee consumption of the full output whether consumers want the shit or not — what’s called “supply-push distribution.” If consumers won’t take it all, you soak up surplus output by destroying it through a permanent war economy, sinking it into an Interstate Highway System, etc. — or maybe just making stuff to fall apart.
The opposite of mass-production is distributed production on the Emilia-Romagna model described by Charles Sabel and Michel Piore in The Second Industrial Divide, with the capital infrastructure distributed to the point of consumption and output geared to local demand. The transnational corporate model of outsourcing is an attempt to put this new wine in old bottles. It distributes the production facilities, but does so on the basis of local labor cost rather than the location of market demand. So it still relies on the centralized wholesale infrastructure of warehouses on wheels/containerships, scaled to peak load, to transfer goods from the distributed production sites to the point of final consumption. The pure and unadulterated distributed manufacturing model, on the other hand, does away with this infrastructure by siting production at the last-mile network of consumption.