The non-AI runup in technology

From: Dani Eder (
Date: Thu Mar 31 2005 - 09:45:01 MST

Durable goods are defined by economists as those
products with an intended useful life of more than
3 years. Included within this category are things
like vehicles, furniture, and factory equipment.
In particular it includes most of the equipment used
to make more equipment. At a sufficiently high
level of automation, factories could effectively
become self-reproducing without human labor. This
can be achieved without human-level AI. It just
requires each factory to be automated in the products
it produces, and the set of factories as a whole to
produce all of each other's needed inputs.

To investigate how much progress we are making
towards this end, I have looked at economic data
gathered by the US department of Commerce on the
durable goods manufacturing sector of the economy.

Over the period from 1992 to 2003, investment in
structures and equipment in this sector increased
63%, and output increased 57%. Thus roughly output
has kept pace with investment. On the other hand,
labor hours decreased by 11.5%, yielding a 78%
increase in labor productivity over that 11 year
period. In other words, it has taken 5% less labor
each year to produce a given amount of products.

The plant and equipment in place in 1992 had a range
of ages from brand new to nearly obsolete. Over the
1992-2003 period 5/8 of the in-place investment
depreciated away - either due to wearing out or
becoming obsolescent. This 63% 'old stuff' was
replaced by 'new stuff', and an additional 63% of
'new stuff' was added to expand production. Thus
of the new stuff installed over the 11 years, half
was for replacement and half for growth.

>From this data I calculated that 'new stuff' installed
in the 1992-2003 period was twice as productive as
the average 'old stuff' in place in 1992. That
assumes that the inherent productiveness of the
workers themselves, as separate from the tools they
have to work with, has not changed. That implies
new stuff has been improving at about an 5.3% per
year. The next level of analysis would be to look
at shorter time intervals to see if there is an
upward trend in the rate of change.

A 5%/year rate of improvement may not qualify as a
singularity level change, but it is comparable to
rates seen in the industrial revolution of the
18-19th century, the mechanization of agriculture
in the 20th, and the rapid development being seen
in some Asian countries at present. Rapid
technology change can lead to disruptive social and
economic change. I think it is prudent to keep
an eye on which way the wind is blowing so as not
to be caught unawares.


Do you Yahoo!?
Yahoo! Small Business - Try our new resources site!

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:51 MDT