Re: poly: Computers induce fast growth?

From: Hal Finney <>
Date: Thu Sep 10 1998 - 21:36:07 PDT

Robin Hanson, <>, writes:
> These are all reasonable criticisms, but don't you find this framework at
> least a useful resource when thinking about what speculations are how
> plausible? Given how little else we have to go on, I find it comforting
> to at least of some sort of empirical context to work with.

It is interesting that it is consistent with other lines of argument.
It can be seen as adding confidence in the Bayesian sense to the
predictions of Drexler and Moravec, etc. We have some a priori belief
in the probability of major changes in technology over the next century.
Your analysis shows that such changes would be broadly consistent with
historical patterns.

It's also interesting to speculate about what happens with the change
after the next one. On June 10, Carl Feynmann had predicted:

: There are still a few people in stage 2, and billions in each of the
: subsequent stages. Each stage takes a long time to squeeze out the
: previous stages. So we can look around us for occupants of the next stage.
: Most of us won't have to look very far, because we are them: we are
: information workers in a bit-centered economy. If the proverb that an
: internet year equals seven ordinary years is true, we can say that the new
: doubling time is two years, and the DT factor is 7. If there are 6,000,000
: people in the new economy, and 6,000,000,000 in the old one, it will take
: roughly 20 years for the new economy to become dominant.
: Stage 7 will become dominant sometime between 2024 and 2038, and will have
: a doubling time between a week and six months.

I'm not sure of Carl's starting point for counting 20 years until the
next transition, but I think it was around 2000, suggesting transition
near 2020 to two year doubling, similar to Robin's results. He then
estimates from 4 to 18 years until the next transition, which would be
2 to 9 doublings, consistent with the data. The speedup factor he gives
for that transition is 4 to 100.

Received on Fri Sep 11 04:20:23 1998

This archive was generated by hypermail 2.1.8 : Tue Mar 07 2006 - 14:45:30 PST