13 November 2010

Singularity Economics

Robin Hanson's article on Singularity Economics is interesting, but, I think, flawed.  Robin lists various examples of capital, but, in the end there have historically been only two forms of capital: humans and land (or sea).  Capital can be produced that multiplies the effects of a human.  Domesticated animals or tractors improved the ability of humans to farm.  Arrows improved the ability of humans to hunt.  Irrigation improved the ability of humans to carry water.

Land is a different form of capital as it either contains resources that humans can extract, or it provides a location where resources can be grown.

There are also two kinds of investments.  One form of investment transfers capital from one person to another.  The other form of investment increases the total pool of capital available to society.  Making movies is an example of the first kind of investment.  Deep water oil drilling, tanning hides for warm clothes, and boats are examples of the second kind of investment.

Investments that increase the pool of available capital also increase the number of kinds of investments that are available, and increase the number of investments that can be undertaken. In a hunter-gatherer society, each human spent most of their time fulfilling their basic needs: gathering food, making clothes, building shelters.  There were few humans, and each human had little time to experiment with improving how to fulfill their daily needs.  With the development of agriculture, each human engaged in farming could produce excess food.  The number of humans greatly increased, the average time each human could spend on research increased, and the pool of available projects expanded dramatically as well.

During the industrial age, the number of humans greatly increased.  Increased lifespans led to an increase in productivity, partially offset by a need for more training before a person becomes productive.  Increases in capital were not well distributed.  Much of the population continued to live in an agricultural society reusing inventions thousands of years old instead of increasing the productivity of humanity.  Over the past generation, we've done a better job of harnessing the potential human capital.  Over half the world population has moved out of an agricultural society and into a manufacturing or post-manufacturing society.

There are various trends that will affect our pool of investment capital.  Slowing population growth will negatively impact the growth of investment capital.  This is offset by improved education, increased lifespans, cheaper manufacturing, and computers.

Many of the social dynamics in the United States reflect the move toward improved education.  Today, highly educated portions of our country graduate 30% of the population from a 4-year college.  Manufacturing jobs have moved out of the country where lower-cost labor and lenient environmental laws make manufacturing cheaper.  In response, the population wants better education in order to obtain good jobs.  Multi-national corporations hire educated people from around the world increasing the need to obtain a good education.  And society is struggling to innovate in education to increase high-school graduation rates while also increasing the knowledge that graduates leave high-school with.

Increased lifespans modestly improve the ratio of time when a human can make contributions that increase the pool of available capital.

The pool of available capital is not just humans, but humans multiplied by technology.  For each human, the tools the humans needs to multiply their muscles or brains must be built.  

To date, computers have not been strong multipliers of human capital.  So far computers have primarily been useful for communication, helping a group of humans communicate faster, and thus reducing redundancy and speeding up the thought processes of the group.  But current computers are far less capable than humans.  Over the next generation, this starts to change.  Individual computers start to have the hardware equivalent of a human brain, and the number of computers per person increases.  

Over the course of the next generation, we should see the available pool of human capital expand much more rapidly than ever before.  Today, I'd guess 10% of humans have the tools they need to be fully productive, and they are productive for about half of their lives.  In a generation, about 90% of humans may be fully productive for about 2/3rds of their lives.  So improvements in education, continued manufacturing, and modest population increases should provide a 10x to 15x expansion of human capital.

Meanwhile, computers should effectively double or triple the available brain power.

Although human capital should greatly increase, the available land capital may not increase as much.  Also, we don't have a model as to the relative importance of human capital versus land capital.  My expectation is that human capital increases the available land capital