It’s no secret that information technology has been a bright spot in an otherwise mundane technology and productivity picture over the last several decades. Computers and the internet — and the industries which most aggressively leveraged them — flourished, while most of the physical industries limped along.
But there is also a persistent worry that Moore’s law — the exponential economic phenomenon driving digital advance — is slowing down. And if the key foundation of information technology slows, then the economy might really be in trouble. The prognosis for Moore’s law is thus a chief factor in overall economic possibilities.
In his new book, The Cloud Revolution, Mark Mills argues that not only is Moore’s law alive and well, but that a number of other technologies are now maturing into powerhouses of their own. Further, the confluence of information and these new physical technologies will lead to another “roaring” era akin to the 1920s when radio, automobiles, airplanes, electrification, and other technologies remade the world. In 1965, Intel founder Gordon Moore, with help from Carver Mead, saw that transistor densities on microchips were doubling every year or two and predicted that they could continue to do so for the next decade, at least.
As transistors got smaller, they would get faster and cheaper. At a furious compound doubling pace, they would launch an information revolution. Few predictions in technology work out, and none so spectacularly as Moore’s law.
For the last 20 years, however, many technologists have argued that Moore’s law is now, or will soon be, dead. In the last five years, those assertions have hardened into common knowledge. It’s true that by the most traditional measure of feature size, Moore’s law can’t continue forever. Existing transistor designs can only shrink so far before they reach an atomic limit. Leading-edge fabs (or semiconductor factories), meanwhile, keep growing in cost (now around $12 billion).
And yet Mills shows that the combination of micro chip advances and macro architectural innovations (the cloud, in which millions of microchips are linked to perform common, commoditized tasks) have combined to keep Moore’s law going strong. What we care about, after all, is not just transistor size but computations per second per dollar. Additional innovations such as 3D chip stacking and wafer-scale chips may help extend Moore’s law for another decade or two, at least. (See me nearby holding the largest chip in the world from Cerebras.)
Beyond chips, Mills, who is a physicist and energy expert, broadens our too-often information-centric technological horizon with a host of new innovations in materials and machines.
One of the most exciting fields is graphene, which is single-layer carbon atoms in a honeycomb lattice. “The properties of such a material,” Mills writes, “are nothing less than magical.” It’s amazingly strong and light and has spectacular electrical properties. Rice University professor Jim Tour has demonstrated dozens of practical applications of Flash Graphene, which is special because it’s so easy and cheap to make into single sheets, unlike previous generations which stuck together. Graphene could transform construction by massively reducing the use of concrete and make nearly every building, vehicle, and machine lighter and stronger. It could also revolutionize biotechnology with nano-cellular applications.
Mills’ book is an antidote to sprawling pessimism. His vision is not automatic, however. It won’t happen if we don’t allow entrepreneurs to grab the antidote and drink it.
No comments:
Post a Comment