From: Amara D. Angelica (amara@kurzweilai.net)
Date: Sun Jun 17 2007 - 22:13:43 MDT
Leonard Skinner said:
> When the technological singularity comes, cars will have an infinite
> number of tailpipes and airbags and razors will have an infinite
> number of blades. Ok, so that's a little silly, but still - past
> performance is no guarantee of future results.
>
> As for razors and tailpipes, perhaps it may be for microprocessors.
Leonard is apparently referring to a humor piece in the March 16th 2006
issue of The Economist, "More blades good." Here's a response from Ray
Kurzweil that may clarify this issue. - Amara D. Angelica, editor,
KurzweilAI.net
"Exponentials continue if there is (1) a benefit or reason for it
continuing, (2) the resources for it to continue, and (3) a mechanism for it
to continue. Rabbits in Australia expand exponentially until they run out
of resources (foliage). Razor blades expand as long as there is a market
benefit that provides the mechanism. Information technology will continue
to expand as long as these three factors enable its expansion. I analyze
the resources for continued expansion of computation in chapter 3 of The
Singularity is Near.
"There ARE limits but they are not very limiting. Based on what we know
about the physics of computation, the amount of matter and energy required
to compute are not zero, but vanishingly small. The ultimate limits of
computation would permit one gram of matter to be trillions of trillions of
times more powerful than the computation required to simulate all several
hundred regions of the human brain, based on the most conservative estimates
of the amount of computation required. One cubic inch of nanotube
circuitry, a type of circuitry that aleady works in experiments, would
ultimately be 100 million times more powerful than the human brain. So the
resources ARE a limitation but will not kick in until vast levels are
achieved. There is a clear benefit from continued expansion as more
powerful, more intelligent, more capable information technology always
eclipses the prior generation. And the mechanism is that more powerful
information technology allows the design of the next generation.
"We will get to the point where the latest generation of information
technology will design its own next generation. We also find that as one
particular paradigm in information technology runs of steam, it creates
research pressure for the next. In the 1950's, they were shrinking vacuum
tubes to keep the exponential growth of the price-performance of computing
going, and that approach ran out of steam when they could no longer shrink
the size of the vacuum tubes and keep the vacuum. This gave rise to the
fourth paradigm: transistors. Moore's law, which pertains to the shrinking
of features on a (flat) integrated circuit, was not the first, but the fifth
paradigm to bring exponential growth to computing. We have had smooth
doubly exponential growth of the price-performance of computing for over a
century, going back to the data processing equipment used in the 1890
census, the first American census to be automated.
"When we run out of steam with Moore's law, we will go to the sixth
paradigm, three-dimensional molecular computing, which is already beginning
to work in laboratories. We won't need this sixth paradigm until about
2020.
"None of these factors pertain to razor blades or tail pipes."
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:57 MDT