Volume 1, Issue 3 
3rd Quarter, 2006


How We Can Manage Our Way Through the Intertwined Promise and Peril of Accelerating Change

Ray Kurzweil

page 4 of 15

Moore's Law (Image 6) says that information technologies double their power every year. 

Kurzweil Image
Image 6: Moore's Law

It is actually double exponential growth, but right now it iss doubling price performance capacity every year and it is very pervasive. Consider my own personal experience. When I came to MIT, a computer took up space larger than this room, yet was less powerful than your cell phone today. There have been 24 doublings of price performance just in terms of MIPS. This does not even take into consideration all the ways in which computers today are more powerful. 

This is just one example of many. Moore's Law is just the vertical stripe on the right in Image 6, shrinking the size of transistors on an integrated circuit, but there has been exponential growth for a hundred years. These are the 49 "famous" computers, going back to the first data processing equipment, used in the 1890 American census, which was old punch card machines. Around 1940, we see the relay-based computer that broke the German enigma code, and then vacuum tube based computers predicted the election of Eisenhower in 1952. They were shrinking vacuum tubes, making them smaller and smaller to keep the exponential growth going. That hit a wall, but it did not stop the overall progression. When one paradigm comes to an end, it actually creates research pressure to create the next paradigm. Thus, transistors, which had a niche application in radio, were brought over to computers. 

So Moore's Laws is not the first, but the fifth paradigm to provide this exponential growth. We have been talking for some time now that that will come to an end. The first predictions were 2002.  Intel now predicts that by 2022, the key features of transistors will be a few atoms in width and we will not be able to shrink them further. Will that be the end of Moore's Law? Yes. Will that will not be the end of the exponential growth of computing?

We will then go to the next paradigm. Because we have been talking about the end of the Moore’s law paradigm for some time now, there has been increasing research on the sixth paradigm, which is three-dimensional molecular computing. We live in a three-dimensional world. Our brain, although it uses a very slow chemical switching, is organized in three dimensions. We might as well use the third dimension. When I talked in The Age of Spiritual Machines in 1999 that the next wave would be three-dimensional molecular computing, nanotubes were a very powerful, likely candidate. I pointed out that DNA computing also would be interesting. That was very controversial then, but it is now a mainstream view because there has been so much progress in three-dimensional molecular computing. 

Image 7 shows Hans Moravec's chart. Supercomputers are marching right along now up to 10 to the 14th. 

Kurzweil Image
Image 7: Evolution of Computer Power/Cost

Moravec's estimate of the computation capacity needed to functionally emulate the human brain is 10 to the 14th; mine is 10 to the 16th, which is a little more conservative. Many different ways of looking at this process of performance, such as dynamic RAM, are also going through different paradigms. 

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Next Page>