The computer becomes infrastructure

In the new issue of the venerable New Left Review, Rob Lucas offers a thoroughgoing critique of my work, going back to the 2003 article “IT Doesn’t Matter” (which, incidentally, I began writing ten years ago this month). Here’s a bit from his discussion of my first book, which grew out of that article:

On an abstract political-economic level, the extensive argumentation of Does IT Matter? was a sledgehammer for a rather small nut: it is a truism that no individual company will succeed in securing for itself significant long-term advantage over competitors solely through the purchase of goods that are also available to those same competitors. But the burden of Carr’s book was to provide an integrated economic and historical elaboration of the dynamics through which IT had been increasingly commoditized, making the transition from a prohibitively expensive endeavour for most companies — something only taken on at great risk by particular capitals in pioneering efforts, such as J. Lyons & Co’s late 1940s LEO (Lyons Electronic Office) — to an increasingly standardized, widely available, mass-produced good with a rapidly deflating price tag coupled to its exponentially improving performance. With this commoditization, Carr argued, IT had made the transition from a particular asset of the individual company to something ‘shared’ by companies, a commodity generally available to all. In the process it had become a standard aspect of infrastructure, a prerequisite for most businesses; it was thus clearly meaningless to appeal to IT spending as a primary basis for ‘competitive advantage’.

Much of this story of commoditization could be told at the level of computer hardware, in isolation from other factors: here, there had been a rapid cheapening of goods related to the technical progress exemplified in Moore’s law, and to the standardization in component manufacture represented by companies like Dell. Already by 2000 the cost of data processing had declined by more than 99.9 per cent since the 1960s, while storage was a tiny fraction of its 1950s price. But for Carr, software also had particular characteristics which help to drive the commoditization of IT in general. Since typical production costs were very high and distribution costs very low, software had extraordinary economies of scale, often making the pooling of resources between firms preferable to the development of particularistic in-house technologies. This supplied an economic rationale for the centralization of IT provision by third parties, who could make the most of these economies of scale by serving many clients. But it also provided an economic basis for the programmer’s communitarian ethic, embodied in professional user groups such as IBM’s long-running SHARE. The resulting standardizations of hardware and software meant that IT typically overshot the needs of its users, since technologies developed for the most demanding users tended to get generalized. This in turn put a deflationary pressure on prices, since it was rational for users to opt for cheaper, older or free technologies that were adequate to their needs, rather than wildly exceeding them. And since software was not subject to wear and tear, once it had saturated a market, new profits could only be gleaned by pushing users through an ‘upgrade cycle’, which they often resisted.

Carr viewed IT as infrastructural in the same sense as the railway, telegraph, telephone, electrical grid and highway systems. For Carr, the consolidation of this infrastructural status was a realization of IT’s tendency to be cheapened, standardized and made generally available, issuing ultimately in its conversion into a grid-based utility — the apotheosis of commoditized IT. Increasingly, IT goods — software services, data storage and even computing power itself — would not be purchased as the fixed capital of individual companies, but would be based in vast centralized data centres and delivered as services over the Internet by a handful of very large providers. On this trajectory, IT was following a path previously taken by electricity provision — a historical analogy that Carr would spell out in his next book, The Big Switch.