Originally published on the blog on 16 June, 2010.
Yesterday during the Oil Industry hearings on Capitol Hill, I saw one congresswoman make a statement that hit a nerve for me. She held up two photos, one was of the oil containment booms used during the Santa Barbara oil spill of the 1960s. Another was of the booms being used today in the Gulf of Mexico. Where, she asked, was there any kind of technical advance in oil containment and why had nothing been done in 40 years to address the problem?
It's a good question, and one for which the answer is most certainly complex. It's quite clear that there are lots of technologies out there that most of the oil companies haven't even bothered to explore that they probably should have, and equally clear that lots of improvisation is going on along the coastline despite the oil business' reluctance to try any such thing themselves.
But it's also a naive question, one that I believe is rooted in what Red Herring once referred to as "the Tyranny of Moore's Law." It's naive because the presumption is that there's always some better technology waiting in the wings, always some new quantum leap yet to come that will solve all our current problems and make a greater future possible, and that something 40 years old in general design must -- almost by definition -- be obsolete.
Moore's Law is the name for the observation by Intel (NasdaqGS: INTC) co-founder Gordon Moore that the number of transistors that could be manufactured on a microchip in an economically viable manner would roughly double every 18 months. It has pretty much held true since the observation was first made in 1965 and was the driving force behind most of the technological advances of the past half century. Companies that bet "against" Moore's Law at virtually every point in the past 45 years have been proven wrong, often catastrophically. Even those that survived were often forced out of the electronics business.
But Moore's law has had a major negative effect. It's lulled us into the incorrect belief that everything can and should advance at that kind of exponential pace, that given enough R&D, every business should be able to achieve the same kinds of cost reductions and product improvements that Intel and others in the semiconductor field have managed. In fact, it should be quite clear that the semiconductor industry is the exception. It's the only business in which advancements follow such a clear trajectory or such an exponential one. Everywhere else, things proceed along much less predictable and certainly non-exponential paths of advancement. Yet over the past 45 years, the belief in there always being some disruptively better technical solution for every problem has become so ingrained that it's usually not critically questioned. When Red Herring wrote about the problem in 2003, they spoke of it as a key cause of the Dotcom boom and bust: millions of would-be entrepreneurs applying Moore's Law rates of growth to businesses for which no such expectation was reasonable. Today, we're applying those same kinds of expectations to everything, including such labor-intensive tasks as oil recovery, health care and screening of airline passengers. Usually when we try to apply the technical lessons of Moore's law to other fields, we fail, often catastrophically.
Yet we hear this ingrained expectation everywhere, including on Capitol Hill yesterday. The presumption is that a device invented 40 years ago should necessarily have been replaced with something significantly different and better. Something exponentially better. But the science and technology of separating crude oil from salt water does not follow the trajectory of Moore's Law. It could be argued that there's no real good technology to replace floating booms for the purpose for which they were intended. They need to float, they need to be flexible, they need to be lightweight enough that huge quantities can be easily transported quickly to anyplace they're needed. Are they perfect? Certainly not. Should they be supplemented with other types of technologies? Certainly. Could those new solutions benefit from new computer technology too? Probably. (In fact, one of the major improvements in boom usage has been the employment of satellite imagery and other computer-enabled technologies to determine where and how to best deploy them.) Has the oil industry failed to put sufficient efforts behind those other supplemental technologies? Quite likely.
But are we likely to find a "quantum leap" type technology to replace booms? Something as lightweight, flexible and easily deployable while being much, much more effective? Personally I doubt it. Again, oil recovery doesn't follow the trajectory of Moore's Law, it follows some very old, very well understood laws of physics. As Scotty often told Captain Kirk, we can't change those. At best, we can use the physics and chemistry in new devices, at worst, we can't do much else.
Most technologies don't follow Moore's Law. Cars are largely unchanged in the past 50 years or longer. Put my grandfather the bus driver (who passed in the 1930s) behind the wheel of a modern vehicle and most of it would seem pretty familiar to him, right down to maintenance items like changing the oil, filling it up with gas or putting air in the tires. There have been slow incremental improvements, but fewer than you might think: most of the big advances in vehicles are the result of better automation and electronics -- they too were driven by Moore's Law. Look around at virtually everything in our world and you'll see stuff that mostly hasn't changed much, or where the changes were driven by better electronics. Technologies that don't utilize electronics or aren't highly-dependent on computerization haven't changed a whole lot.
My bike is similar in operation to the one I had in grade school. Better materials technology, better design and better machining have made it lighter, stronger and given it a lot more gears. However, it seems that basic mechanics was pretty-well developed 35 years ago. Moore's Law-driven technologies have certainly contributed to the improvements (mostly through computer aided design and manufacturing), but there was no quantum leap. I could not ride cross country a whole lot more easily today than I could when I did my first long tour at age 16. The most efficient cooling device in my home -- a modern ceiling fan -- would be easily recognizable by my grandmother: put her 1940s model side-by-side with my 2010 model and you'd need an engineer to point out the ways in which the current one is better.
This is part of the driving force behind the current economic malaise around the world. The problems of the past 30 years and were mostly solved with computerization and application of better electronics, which have also driven new growth in new industries. The exponential improvements in computer technology has resulted in second and third order improvements in everything else. As available computing power has continued to grow beyond most people's ability to use it, we are running headlong into problems that are not so easily solved. More advanced electronics can make gasoline-based cars more efficient, more luxurious, even more fun, but they won't help us replace the internal combustion engine with something independent of fossil fuels, at best they'll be a second-order enabler of whatever new drive technology we need to invent. Likewise, we're not going to solve our healthcare cost crisis by hoping that advanced technology will result in a Moore's Law style exponential decline in costs. There is no evidence to suggest that such a thing is possible, let alone likely, yet I still hear those who suggest that high costs for care can be cured by just unleashing Silicon Valley technologies.
It's time for us to look beyond Moore's Law and the kind of exponential growth it has driven in a limited number of industries. Moore's Law is not the norm, it's the exception. 40 years ago flexible booms were designed within a set of limitations that drove the best design possible. Other than some materials technology improvements, they mostly look the same today and they're likely to look much the same in another 40 years because the design limitation are unlikely to change much. 40 years ago healthcare was people-intensive and thus unlikely to benefit much from economies of scale in the way electronics do, and the same is likely to be true 40 years from now. Properly screening airline passengers is always going to be people-intensive, as it is in Tel-Aviv, regardless of how many "revolutionary" technologies GE (NYSE: GE) and Rapiscan (NasdaqGM: OSIS) convince us to spend money on. In fact, most things in 40 years are not likely to look all that different. It's time we stopped pretending computer technology solves all problems, or that everything can be improved at Moore's Law's exponential rate.
-mg
