From: Robin Hanson (rhanson@gmu.edu)
Date: Fri Jul 25 2003 - 14:52:28 MDT
At 04:46 PM 7/24/2003 +0200, Alfio pointed us to:
>http://marshallbrain.com/robotic-nation.htm
His predictions about future progress in robotic technology are not crazy,
even if they are not assured. He ignores the software problem - even if
robots have more computing power than the brain, it may take a while for
software to be good enough. But it is not crazy to think the software
problem might be solved within a decade or two of when we have enough
hardware.
His social predictions are more of a mixed bag. He may in fact turn out to
be right in his specific predictions, but if so it will partly due to luck;
he doesn't show much indication that he understands the sort of economic
analysis that is most appropriate to considering such situations. Alas, his
is the sort of analysis that shows up time and again among technologies
who understand robotics far better than they understand social science.
For example, he doesn't consider that people who leave some jobs might find
others. Automation has taken away most farming jobs, but that hasn't made
most people unemployed; people now do other things that they do better than
machines. Similarly most economists are not worried because for the last
few centuries machines and human labor have mostly been *complements*. This
means that the more efficient one of them becomes, the more valuable the
other becomes. So as machines take over some jobs, that had *raised* the
value of, and hence the wages given to, the other jobs people still do. So
most people remain employed, and focus on the tasks machines are worse at.
However, a closer examination does I think suggest that eventually machines
will become mostly substitutes for, rather than complements of, human labor.
I present my argument for that position in this technical paper:
http://hanson.gmu.edu/aigrow.pdf The basic idea is that types of tasks are
complements to each other, but that machines can substitute human labor for
any given task. My argument isn't necessarily widely accepted though.
(The strongest case for substitution comes I think in the case of uploads.)
Robin Hanson rhanson@gmu.edu http://hanson.gmu.edu
Assistant Professor of Economics, George Mason University
MSN 1D3, Carow Hall, Fairfax VA 22030-4444
703-993-2326 FAX: 703-993-2323
This archive was generated by hypermail 2.1.5 : Fri Jul 25 2003 - 15:00:33 MDT