Moore’s law, which says that number of transistors you can fit onto an integrated circuit doubles every 18 months, has been the benchmark measurement for technical progress in electronics for decades. The doubling of transistors on a chip translates to a doubling of computing power and so–it was believed– Moore’s law was the reason why people in 2007 could carry a computer in their pocket, the Apple iPhone, that was four hundred times more powerful than the first Apple computer that debuted in 1976 (as measured by Hertz).
But because Moore’s law applies only to electronics, it can’t be used to forecast technological progress in other areas, or even in areas of computing that don’t involve transistors, such as in quantum computing.
Researchers from the Santa Fe Institute now argue that a theory proposed by Theodore Wright in 1936, called Wright’s law, is actually a better reflection of technological progress than is Moore’s law. In their working paper, “Statistical Basis for Predicting Technological Progress,” they detail how they looked at technological progress rates from 62 different technologies including chemical compound manufacture, mechanical engineering, etc., and found key similarities.
“Moore’s law says that costs come down no matter what at an exponential rate. Wright’s law says that costs come down as a function of cumulative production. It could be production is going up because cost is going down,” Santa Fe Institute lecturer Doyne Farmer told Futurist Update.
More importantly, Wright’s law can be applied to a much wider variety of engineering areas, not just transistors. That will give technological forecasters a new way to measure and predict progress and cost for everything from airplane manufacturing (its original use) to the costs of building better photovoltaic panels.
“It means that if investors or the government are willing to stimulate production, then we can bring the cost down faster. In the case of global warming, for instance, I think that a massive stimulus program has the potential to really bring the arrival date for having solar energy beat coal a lot sooner,” said Farmer.
He and his colleagues are expanding their working paper into more expansive study that further details the relationship between costs and the rate of progress. “We’re trying to make nice, probabilistic forecasts for where solar will be with and without stimulus, what’s the distribution of times that will happen with business as usual or a scenario,” says Farmer. They plan to submit their final work to Nature next month.
Source: Doyne Farmer (interview), The Santa Fe Institute. “Statistical Basis for Predicting Technological Progress” (PDF) is available from the Santa Fe Institute.
Source: Futurist Magazine Update (email newsletter) September 2012