Moore's Law and Technological Determinism: Reflections on the History of Technology

Just over a year ago, the arrival in my mailbox of a book I had agreed to review triggered some thoughts about technology I had been meaning to articulate. The book was Ross Bassett's To the Digital Age: Research Labs, Start-up Companies, and the Rise of MOS Technology (Baltimore, 2002).¹ In it, Bassett describes the development of metal-oxide semiconductor (MOS) technology, which enabled semiconductor firms to place more and more transistors on a single silicon chip.² This became the basis for what is now known as Moore's law, after Gordon E. Moore. In April 1965, Moore, then the director of research and development at the semiconductor division of Fairchild Camera and Instrument Corporation, published a paper in which he observed that the number of transistors that could be placed on an integrated circuit had doubled every year since integrated circuits had been invented and predicted that that trend would continue.³ Shortly afterward, Moore left Fairchild to cofound Intel—a company, Bassett notes, that staked its future on MOS technology.

It is important to note at the outset that Moore's law was an empirical observation; it is not analogous to, say, Ohm's law, which relates resistance to current. Moore simply looked at the circuits being produced, plotted their density on a piece of semi-log graph paper, and found a straight line. Furthermore, he made this observation in 1965, when the integrated circuit was only six years old and had barely found its way out of the laboratory. The name "Silicon Valley" did not even exist; it would be coined at the end of that decade. Nonetheless, Moore's prediction that the number of transistors that could be placed on an integrated circuit would continue to double at short, regular intervals has held true ever since, although the interval soon stretched from twelve to eighteen months.⁴

Moore's law has been intensively studied, mainly by those wondering when, if ever, fundamental physical constraints (such as the diameter of a hydrogen atom) will interrupt the straight line that Moore observed. These studies note the lengthening of the interval mentioned already: chip densities now double about every eighteen to twenty months, although no one is sure why.⁵ Analysts have been predicting the failure of Moore's law for years. Interestingly, the moment of its demise seems always to be about ten years from whenever the prediction is made; that is, those writing in 1994 anticipated that it would fail in 2004, while some today put the likely date at about 2015. Obviously one of these predictions will pan out someday, but for now Moore's law is very much in force, as it has been for over forty-five years—a fact from which the lengthening of the doubling interval should not distract us. Over the same period, computer-disk memory capacity and fiber-optic cable bandwidth have also increased at exponential rates. Thus, in 2005 we see memory chips approaching a billion (10⁹ ) bits of storage, Apple iPods with forty-gigabyte (3 X 10¹¹ bits) disks, and local networks capable of transmitting a full-length Hollywood feature film in seconds.

But while industry analysts, engineers, and marketing people have studied Moore's law intensively, historians of science and technology have shown less interest. That is surprising, since it cuts to the heart of an issue that they have debated over the years: technological determinism.

Mel Kranzberg and his colleagues organized the Society for the History of Technology in part to foster a view of technology running counter to the notion that technology is an impersonal force with its own internal logic and a trajectory that human beings must follow. The society's founders spoke of a "contextual" approach to technology, in which the linear narrative of events from invention to application was accompanied by an understanding of the context...