Computers are now more than a million times faster than they were when the first hand calculator appeared back in the 1960s. (An engineer working at Texas Instruments, Jack Kilby, had invented the first integrated circuit, or semiconductor, in 1957.) This incredible, exponential increase was predicted via ‘Moore’s Law,’ first formulated in 1965: that is that the number of transistors in a semiconductor doubles approximately every two years.
Another way to state this Law (which is not a natural ‘law’ at all, but an observational prediction) is to say that each generation of transistors will be half the size of the last. This is obviously a finite process, with an end in sight. Well, in our imaginations at least.
The implications of this end are not so small. As we all know, rapidly evolving digital technology has hugely impacted nearly every sector of our economy, and with those changes has come disruptive social change, but also rapid economic growth. The two largest arenas of economic growth in the U.S. in recent years have been Wall Street and Silicon Valley, and Wall Street has prospered on the manipulation of money, via computers, while Silicon Valley (Silicon is the ‘plate’ upon which a semiconductor is usually built.) has prospered upon the growing ubiquity of computers themselves.
Intel has predicted that the end of this exponential innovation will come anywhere between 2013 and 2018. Moore’s Law itself predicts the end at 2020. Gordon Moore himself—he who formulated the Law—said in a 2005 interview that, “In terms of size [of transistors] you can see that we’re approaching the size of atoms, which is a fundamental barrier.” Well, in 2012 a team working at the University of New South Wales announced the development of the first working transistor consisting of a single atom. That sounds a lot like the end of the line.
In November of last year, a group of eminent semiconductor experts met in Washington to discuss the current state of semiconductor innovation, as well as its worrisome future. These men (alas, yes, all men) are worried about the future of semiconductor innovation because it seems that there are a number of basic ideas about how innovation can continue past the coming ‘end,’ but none of these ideas has emerged as more promising than the others, and any one of them is going to be very expensive. We’re talking a kind of paradigm shift, from microelectronics to nanoelectronics, and, as is often the case, the early stages of a fundamentally new technology are much more costly than the later stages, when the new technology has been scaled up.
And of course research dollars are more difficult to secure these days than they have been in the past. Thus the additional worry that the U.S., which has for decades led the world in digital innovation, is going to be eclipsed by countries like China and Korea that are now investing more in R&D than is the U.S. The 2013 budget sequestration cuts have, for instance, directly impacted certain university research budgets, causing programs to be cancelled and researchers to be laid off.
One of the ironies of the situation, for all those of us who consider corporate monopoly to be abhorrent, is evident when a speaker at the conference mentioned working at the Bell Labs back in the day when Ma Bell (AT&T) operated as a monopoly and funds at the Labs were virtually unlimited. Among the technologies originating at the Bell Labs are the transistor, the laser, and the UNIX operating system.
It’s going to be interesting, because the need is not going away. The runaway train that is broadband appetite, for instance, is not slowing down; by 2015 it’s estimated that there will be 16 times the amount of video clamoring to get online than there is today.
It’s worth noting that predictions about Moore’s law lasting only about another decade have been made for the last 30 years. And futurists like Ray Kurzweil and Bruce Sterling believe that exponential innovation will continue on past the end of its current course due in large part to a ‘Law of Accelerating Technical Returns,’ leading ultimately to ‘The Singularity,’ where computers surpass human intelligence.
Someone should tell those anxious computer scientists who convened last November in Washington: not to worry. Computers will solve this problem for us.