Re: limits to exponential growth (RE: Other signposts towardsthe Singularity)

From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Mon Jul 23 2001 - 14:44:41 MDT


James Higgins wrote:
>
> At 03:24 PM 7/23/2001 -0400, you wrote:
> >Supposing that Moore's Law suddenly breaks down in 2005 and transistors
> >abruptly stop shrinking, I'm still not nervous, because I know that
> >computers will suddenly start having more and more symmetric
> >multiprocessing CPUs, rather than bigger and bigger CPUs, and everyone
> >will talk about how it was "obvious" that Moore's Law would someday switch
> >over to parallel power per dollar instead of serial speed per dollar. Or
> >we could see serial speed continuing to increase at the current rate even
> >as the number of CPUs also begin to double every couple of years, which
> >would also break down Moore's Law, albeit from the opposite direction. Or
> >FPGAs (field-programmable gate arrays) might become common and break down
> >our yardsticks.
>
> Moore's Law has nothing to do with CPUs! If transistor counts stop
> increasing but systems start shipping with multiprocessing CPUs (not in the
> same chip) Moore's Law is broken. Likewise, if transistor counts continue
> to increase as predicted but systems also ship with multiple CPUs there is
> no effect on Moore's Law.

Yes, I know. That was exactly my point - that Moore's Law might break
down completely, and yet computers would go on smoothly increasing in
cost/performance with scarely a blip. If this happened, everyone would
immediately begin speaking of Moore's Law as if it strictly controlled
cost/performance and had never had anything to do with transistor sizes.
If transistor counts kept increasing but systems also started shipping
with multiple CPUs, Moore's Law might technically remain intact but would,
from the perspective of a computer buyer, suddenly appear not as a steady
acceleration but as a speed limit that had been broken (!).

Moore's Law is now really a social observation rather than a technical
one, and is used that way almost universally; it generalizes so readily to
so many observed phenomena in the computing industry that one speaks of
"Moore's Law for performance per dollar" or "Moore's Law for disk
drives". As I don't work for Intel, I have now accepted this and have no
trouble with, e.g., Ray Kurzweil's graph showing the operation of "Moore's
Law since 1900" for clerks, mechanical calculators, vacuum tubes, and
transistors.

Moore only *thought* his Law was about transistors. Imagine if Isaac
Newton had first observed that apples fall at a constant acceleration, and
formulated Newton's Law as "The velocity of a freely falling apple changes
by a constant amount in a given unit of time." Would we still be speaking
of Newton's Law as if it applied only to apples?

-- -- -- -- --
Eliezer S. Yudkowsky http://intelligence.org/
Research Fellow, Singularity Institute for Artificial Intelligence



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT