Commenter Pastorius commenting on my article on Moore's Law suggested I have a look at this article by Ray Kurzweil on the accelerating rate of the rate of change.
...a serious assessment of the history of technology shows that technological change is exponential. In exponential growth, we find that a key measurement such as computational power is multiplied by a constant factor for each unit of time (e.g., doubling every year) rather than just being added to incrementally. Exponential growth is a feature of any evolutionary process, of which technology is a primary example.
One can examine the data in different ways, on different time scales, and for a wide variety of technologies ranging from electronic to biological, and the acceleration of progress and growth applies. Indeed, we find not just simple exponential growth, but "double" exponential growth, meaning that the rate of exponential growth is itself growing exponentially. These observations do not rely merely on an assumption of the continuation of Moore's law (i.e., the exponential shrinking of transistor sizes on an integrated circuit), but is based on a rich model of diverse technological processes. What it clearly shows is that technology, particularly the pace of technological change, advances (at least) exponentially, not linearly, and has been doing so since the advent of technology, indeed since the advent of evolution on Earth.
Yep. Change is happening much faster than when I was a kid (50s). We have so much more capable technology than we had then and so many more capable technologies.
What does this mean in terms of solving humanity's technical problems? Say energy for instance. It says we should put off implementing a solution for as long as possible because better solutions are just around the corner. Al Gore's idea that we need to rush to fix our reliance on fossil fuels chop, chop, double quick, is flat out wrong. It would be much better for us to rely on current technologies for as long as possible (at least until the alternatives become economically competitive) because the solutions we will have available in five or ten years will be so much better than the ones available today. The important thing is to avoid, thorough any kind of government program, getting locked in or promoting any given technology.
In the home computer market that is pretty much what we do. If we assume a doubling of capability every two years replacing your home computer every four to six years makes a lot of sense. In four years your "old" machine will have 25% of the power of what is currently on the market. In six years the "old" machine will have 12.5% of the capacity of the latest and greatest. Its economic value at that time (six years after purchase) will be around zero. Which is why if you go out on garbage day looking for a computer you will generally find machines about five to eight years old. Business is a little different. They can't afford to get too far behind in technology. Which is why they get new computers on a three year schedule. They can't afford to give up more than a factor of three to their competition.
Here is another example (based on hypotheticals). Suppose we do a big push on solar and then some kind of nuclear fusion "magic" comes along reducing the price of electricity by a factor of 10X? Under those circumstances the less we spend on alternative energy at today's prices the better off we will be. Given the accelerating rate of change we are bound to come up with some new kind of "magic" of one sort or the other.
It all depends on the learning curve of a particular technology. I'm not sure what the learning curve for solar is but I can guarantee that what we can do in five years will be much better and more cost effective than what we can do today. Kurzweil estimates that the rate of technological change in 2100 will be around 20,000 times the rate of today. That is pretty damn fast. How fast? Changes that we would see in five years at todays accelerating rate will happen in a day around 2100. (I haven't done the math so that number is just an estimated representation to give a feel for what is happening).
What does this tell us in general? Forcing change is wasteful. Real environmentalists (who inherently are conservative) will resist change until the market provides economic solutions, because forcing change increases waste. We saw this in the solar boom in the Carter era. It didn't work out. Because of government subsidies there was a huge amount of waste. Or as that most wise of sages once said:
TrackBack URL for this entry: http://classicalvalues.com/cgi-bin/pings.cgi/6935
Comments
In simplest terms, we do now what we know how to do well: drill more, build refineries, build a few (20) nuclear plants. Those will carry us 20-30 years. In the interim, remain poised to jump on nanosolar and waste biofuel solutions as they come online. Adaptability will be more valuable than long-term solutions. In a side note, I believe adaptability will be a more important survival skill than intelligence as time goes on (admitting that they are related).
Even if the improvement were only linear, we can solve things better now than we could 20 years ago. If Kurzweil is right and we should view this logarithmically, then our point of comparison should be 1928, not 1988. Yeah, we can do a lot more than we could in 1928, eh?
There's only one problem with the Kurzweil theory: trees don't grow to the sky.
The reality is, most progress (indeed, most phenomena) follows an S-curve: a period of exponential growth, followed by an inflection point, followed by a period of logarithmic growth.
Even CPUs--the epitome of Moore's law--illustrate this: back when I was first buying PCs in the 1980s and 1990s, you could increase the CPU speed of your machine by factors of 3X (e.g., when I went from a 33-MHz 486 to a 133-MHz Pentium) or even 4X (when I went from an 8-MHz 286 to a 33-MHz 486).
But for the last decade, we've seen much smaller and more incremental changes, with no CPU that I am aware of going much beyond 3GHz and most of the current high performance chips playing in the 2.4-3.0GHz space. Given that we broke the 1-GHz barrier back in 2000, that's a pretty slow rate of increase compared to the previous two decades.
David Hecht · July 23, 2008 10:38 AM
David,
There is one thing you are leaving out: substitution. Replacing trees with airplanes and then rockets.
Or take CPUs. They are not getting faster by much. We just get more CPUs per chip. I can buy some chips that have 24 CPUs on them. About one CPU per two active pins.
A 3Ghz signal at the speed of light travels 4 inches, less in copper or silicon. It's a fundamental limit.
Computer servers generally are kept longer than desktop computers. They usually have a 5+ year lifecycle. Microsoft releases it's server OS at 5 year intervals (2003, 2008). The desktop 3 yr cycle is for professionals (e.g., graphic artists, developers) and power users. The grunts are still running Windows 2000.
Gideon7 · July 23, 2008 02:20 PM
The late Frank Herbert wrote some stories about a society so rapid and efficient that it became a true horror. To slow things and make them less efficient, they created the Bureau of Sabotage.
The Anti-Idiotarian Rottweiler
Don?t bang your head!
In simplest terms, we do now what we know how to do well: drill more, build refineries, build a few (20) nuclear plants. Those will carry us 20-30 years. In the interim, remain poised to jump on nanosolar and waste biofuel solutions as they come online. Adaptability will be more valuable than long-term solutions. In a side note, I believe adaptability will be a more important survival skill than intelligence as time goes on (admitting that they are related).
Even if the improvement were only linear, we can solve things better now than we could 20 years ago. If Kurzweil is right and we should view this logarithmically, then our point of comparison should be 1928, not 1988. Yeah, we can do a lot more than we could in 1928, eh?