From: Stuart Armstrong (dragondreaming@googlemail.com)
Date: Mon Jul 14 2008 - 05:03:43 MDT
The lack of bottlenecks is the extraordinary thing. When making things
bigger, or moving faster, extra strains are put on the structure;
pieces need to be reinforced, energy needs to be stored, and so there
are a host of problems you have to deal with, none directly helping
make the thing bigger or faster, but essential nonetheless.
With Moore's law, all you need to do (simplifying) is to make
transistors smaller. The extraordinary thing is that all the attendant
changes you need to make (cooling, isolating components from each
other, etc...) seem to be minor and cheap; so designers can focus
almost entirely on making the transistors smaller. A single problem,
rather than a multitude.
There are quantitative improvements as well, but these seem to come at
about the rate you'd expect for technological inovations. They
conveniently remove the bottlenecks from the "make this smaller"
process, though. I think the self-fulfilling property aspect of
Moore's law might explain these quantative improvements, but the
general pace of "make it smaller" is dictated by the fact that it is a
single problem.
Lastly, the self-fullfilling aspect may provide an UPPER bound on the
rate of improvement. If software creators have certain expectations,
they will design their software to Moore's law; therefore any dramatic
improvement much faster than this is a commercial waste, especially
since design ideas are quickly copied.
Stuart
2008/7/13 Thomas McCabe <pphysics141@gmail.com>:
> Me, Anna Salamon, and Steve Rayhawk are building a computer model of
> existential risks and what causes them. Currently, we're looking for
> information on what causes Moore's Law-style exponentials in computer
> hardware, so we can predict what will happen given conditions
> different from those of the past forty years. The current favorite
> theories seem to be that 1)., Moore's Law was a self-fulfilling
> prophecy because it was used to set industry goals, and 2)., most
> technological growth is exponential, because the rate at which
> technology improves is proportional to the amount of technology
> already in existence. Any links to other theories, or literature on
> the subject, would be greatly appreciated.
>
> --
> - Tom
> http://www.acceleratingfuture.com/tom
>
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:03 MDT