Re: No AI for Nano/No Nano for copyloads

From: Robin Hanson (
Date: Sat Jul 15 2000 - 09:02:14 MDT

Robert J. Bradbury wrote:

> I've seen arguments that the's tend to drive themselves out of business
> at an ever increasing rate. ... Don't cheap copies only do the same thing for
> labor market? ... isn't the problem that the copies compete against each other
> to drive the profit out of their specific specialty? Shouldn't
> the cost of labor fall to approximately the level of the amortized cost of the
> hardware plus the energy required to operate it?

Once there is enough competitive entry, then yes. This is the situation for any
innovation. When the innovation is available to all, competition drives the
excess-profit from using that innovation to zero. But if the first successful
entrant is not easy to copy, then that entrant can earn excess profits. The
prospect of such profits, however, can lead to lots of entrant attempts, and on
average there are no profits from attempting to enter. But entry, and innovation,
does occur.

More concretely, if the first upload company has some intellectual property, they
may get returns from that property, and that will raise the market wages above just
the cost of hardware and energy. But as competitors find substitutes for that
intellectual property, the excess wage will fall.

> Since labor is what is required to build increasingly more efficient hardware
> (the self-evolving AI model), doesn't it rapidly spiral to the point where you
> have reached the physical limits? And isn't this just another way of looking at
> the singularity?

I don't know. But I did analyse this stuff a little more formally
at or .ps

This archive was generated by hypermail 2b29 : Mon Oct 02 2000 - 17:34:37 MDT