Moore's Law, Progress, and Displacement

The way that "Moore's Law" is usually cited by those in the know is something along the lines of: "the number of transistors that can be fit onto a square inch of silicon doubles every 12 months."

...However, it doesn't quite do justice to the full scope of the picture that Moore painted in his brief, uncannily prescient paper. This is because Moore's paper dealt with more than just shrinking transistor sizes. Moore was ultimately interested in shrinking transistor costs
--Jon "Hannibal" Stokes

I was defeated by Moore's Law before I even knew what it was. I used to play in OthelloTM tournaments. Othello is a board game sold by Pressman Toys, and like checkers or chess, it is a game of skill that requires looking ahead.

In 1987, I came in sixth in the world Othello championships in Milan, Italy, in which over 30 players from a dozen countries participated. I was no threat to finish in the top four, but I was a reasonably decent player.

About that time, some of the first serious Othello-playing computer programs for PC's were developed. One of the best programs was written by a fellow named David Parsons. However, I could beat his program really easily.

A few years later, I bought a new PC, powered by an Intel 486 processor rather than the 286 that I had been using. I took the disk from Parsons' program, loaded it into my new computer...and proceeded to lose every game I played against the machine.

What happened was that with its more powerful processor, my new PC was looking ahead seven to nine moves, compared with just five moves before. With the older processor, a search depth of seven or nine moves would have taken about an hour to return a single move--easily leading to a time default in a tournament game. With the new processor, the computer could make each move in less than a minute.

This increase in power reflected the number of transistors, among other factors. The 286 chip had 110,000 transistors, and the 486 chip had 1,180,000 transistors, or about ten times as many.

Intel founder Gordon Moore had predicted in 1965 that for the next ten years processors would become more powerful. As Stokes points out, his prediction was more subtle and far more complex than is usually described. Readers who want to understand what Moore really meant should consult Stokes' article.

Stokes believes that the simplest approximation of Moore's Law might be,

"The number of transistors per chip that yields the minimum cost per transistor has increased at a rate of roughly a factor of two per year."

Research in computer science has maintained a pace close to that of Moore's Law all the way into this century, and experts seem to expect that this will continue for the next decade or more. What this means in practical terms is that the amount of computing power that you can buy for a given amount of money--say, $2000--doubles roughly every two years.

An economic process that doubles every two years is disruptive. Economists are used to compound growth, but at slow rates. Population might grow at one or two percent per year. Labor productivity might grow at one or two percent per year. Interest in a savings account might grow (after inflation) at a rate of two or three percent per year. Stock market investments might yield a return (after inflation) of four to seven percent per year. Moore's Law describes a process that is growing at 50 percent per year. The cumulative results have been staggering already. What lies ahead is even more astounding.

In the "Rational Exuberance" essay of the chapter on growth, I talked about the potential for economic growth to accelerate as computers continue to improve. In this chapter, I look at the potential effect of increased computer power on various individual sectors of the economy.

I no longer play tournament Othello. I am discouraged by the fact that computers are so much better than I am. In the early 1990's, a computer could beat me by looking ahead nine moves. My current laptop can look ahead twenty moves!

I have been displaced in Othello. The issue of progress leading to displacement is one of the mixed blessings of technology. This chapter tries to predict some of the displacement that will take place in the future.