Microsoft Word Kurzweil, Ray The Singularity Is Near doc

Going Beyond the Ultimate: Pico- and Femtotechnology and Bending the Speed of Light.
The limits of around 
10
42
cps for a one-kilogram, one-liter cold computer and around 10
50
for a (very) hot one are based on computing with 
atoms. But limits are not always what they seem. New scientific understanding has a way of pushing apparent limits 
aside. As one of many such examples, early in the history of aviation, a consensus analysis of the limits of jet 
propulsion apparently demonstrated that jet aircraft were infeasible.
71
The limits I discussed above represent the limits of nanotechnology based on our current understanding. But what 
about picotechnology, measured in trillionths (10
–12
) of a meter, and femtotechnology, scales of 10
–15
of a meter? At 
these scales, we would require computing with subatomic particles. With such smaller size comes the potential for 
even greater speed and density. 
We do have at least several very early-adopter picoscale technologies. German scientists have created an atomic-
force microscope (AFM) that can resolve features of an atom that are only seventy-seven picometers across
72
An even 
higher-resolution technology has been created by scientists at the University of California at Santa Barbara, who have 
developed an extremely sensitive measurement detector with a physical beam made of gallium-arsenide crystal and a 
sensing system that can measure a flexing of the beam of as little as one picometer. The device is intended to provide a 
test of Heisenberg's uncertainty principle.
73
In the time dimension Cornell University scientists have demonstrated an imaging technology based on X-ray 
scattering that can record movies of the movement of a single electron. Each frame represents only four attoseconds 
(10
–18
seconds, each one a billionth of a billionth of a second.
74
The device can achieve spatial resolution of one 
angstrom (10
–10
meter, which is 100 picometers). 
However, our understanding of matter at these scales, particularly in the femtometer range, is not sufficiently well 
developed to propose computing paradigms. An 
Engines of Creation
(Eric Drexler's seminal 1986 book that provided 
the foundations for nanotechnology) for pico- or femtotechnology has not yet been written. However, each of the 

---

competing theories for the behavior of matter and energy at these scales is based on mathematical models that are 
based on computable transformations. Many of the transformations in physics do provide the basis for universal 
computation (that is, transformations from which we can build general-purpose computers), and it may be that 
behavior in the pico- and femtometer range will do so as well. 
Of course, even if the basic mechanisms of matter in these ranges provide for universal computation in theory, we 
would still have to devise the requisite engineering to create massive numbers of computing elements and learn how to 
control them. These are similar to the challenges on which we are now rapidly making progress in the field of 
nanotechnology. At this time, we have to regard the feasibility of pico- and femtocomputing as speculative. But 
nanocomputing will provide massive levels of intelligence, so if it's at all possible to do, our future intelligence will be 
likely to figure out the necessary processes. The mental experiment we should be making is not whether humans as we 
know them today will be capable of engineering pico- and femtocomputing technologies, but whether the vast 
intelligence of future nanotechnology-based intelligence (which will be trillions of trillions of times more capable than 
contemporary biological human intelligence) will be capable of rendering these designs. Although I believe it is likely 
that our future nanotechnology-based intelligence will be able to engineer computation at scales finer than 
nanotechnology, the projections in this book concerning the Singularity do not rely on this speculation. 
In addition to making computing smaller, we can make it bigger—that is, we can replicate these very small 
devices on a massive scale. With full-scale nanotechnology, computing resources can be made self-replicating and 
thus can rapidly convert mass and energy into an intelligent form. However, we run up against the speed of light, 
because the matter in the universe is spread out over vast distances. 
As we will discuss later, there are at least suggestions that the speed of light may not be immutable. Physicists 
Steve Lamoreaux and Justin Torgerson of the Los Alamos National Laboratory have analyzed data from an old natural 
nuclear reactor that two billion years ago produced a fission reaction lasting several hundred thousand years in what is 
now West Africa.
75
Examining radioactive isotopes left over from the reactor and comparing them to isotopes from 
similar nuclear reactions today, they determined that the physics constant alpha (also called the fine-structure 
constant), which determines the strength of the electromagnetic force, apparently has changed over two billion years. 
This is of great significance to the world of physics, because the speed of light is inversely proportional to alpha, and 
both have been considered unchangeable constants. Alpha appears to have decreased by 4.5 parts out of 10
8
. If 
confirmed, this would imply that the speed of light has increased. 
Of course, these exploratory results will need to be carefully verified. If true, they may hold great importance for 
the future of our civilization. If the speed of light has increased, it has presumably done so not just as a result of the 
passage of time but because certain conditions have changed. If the speed of light has changed due to changing 
circumstances, that cracks open the door just enough for the vast powers of our future intelligence and technology to 
swing the door widely open. This is the type of scientific insight that technologists can exploit. Human engineering 
often takes a natural, frequently subtle, effect, and controls it with a view toward greatly leveraging and magnifying it. 
Even if we find it difficult to significantly increase the speed of light over the long distances of space, doing so 
within the small confines of a computing device would also have important consequences for extending the potential 
for computation. The speed of light is one of the limits that constrain computing devices even today, so the ability to 
boost it would extend further the limits of computation. We will explore several other intriguing approaches to 
possibly increasing, or circumventing, the speed of light in chapter 6. Expanding the speed of light is, of course, 
speculative today, and none of the analyses underlying our expectation of the Singularity rely on this possibility.