Interview with Mark Gubrud
Questions by Sander Olson.
Answers by Mark Gubrud.
Question 1: Tell us
about yourself. What is your background, and current position?
I am a graduate student in
the Physics department at the University of Maryland. I am currently working
on SQUIDs (superconducting Quantum Interference Devices) - quantum computing
projects for the Department of Defense. Quantum computing could potentially
be used to solve certain specialized tasks, such as decryption, simulating
quantum systems, and perhaps others we don't know yet.
Question 2: You have
studied the concept of molecular nanotechnology extensively. How would
you describe the evolution of the concept of molecular nanotechnology?
How has it affected concepts, such as the concept of a technological singularity?
The question is a very interesting
one. The concept of a singularity follows directly from the original concepts
of molecular nanotechnology, put forth originally by Richard Feynman. But
it was Eric Drexler'S Engines of Creation that really began the movement
to what I would call a "technology cult" - this description of technology
in quasi religious terms. When Vernor Vinge
came along and put forth
the concept of a technological singularity, it really only sharpened Drexler's
original works. If you have either AI or nanotechnology, that could lead
to a singularity, because the rate of technological progress would be set
by technology, rather than the speed at which people work.
Question 3: What are
the necessary precursor technologies for nanotechnology. Is Artificial
Intelligence necessary?
To some extent we already
have both AI and Nanotech. If you look at the latest Pentium chips, they
already enable some AI applications, and have elements that are 250 nanometers.
If you look at the International Roadmap for Semiconductors (ITRS), you
see that within 20 years transistors will be 20-50 nanometers. WE will
be able to match the power of the human brain. So we don't need Drexlerian
assemblers to have Artificial Intelligence.
Question 4: Marvin
Minsky recently argued that a 1 megahertz machine could become sentient
with the right programming. Do you agree?
Minsky is more of an expert
than I am, but I don't agree with him. Right now you have machines beating
masters at chess, you have the Google search engine, which I can ask a
question on any topic that you can think of, and it will come back with
more information than I could think of, ever. That system is already intelligent
in that it knows alot about numerous subjects. So there are things machines
can do that I can't do, and things I can do that machines can't do. So
the question is when can a machine do both things? We can now work on many
of those things, such as vision, processing visual information, and speech
recognition and common sense. To reason, to question, it's clear that's
where we're going. I see search engines as a form of superintelligence.
The clock speed of the human brain is quite slow, so you can substitute
a really fast machine for a really big machine.
Question 5: But aren't
these advanced AI techniques, such as search engines and neural nets, really
just idiot savants, capable of only a very limited form of intelligence?
After all, the Deep Blue supercomputer beat the human world chess champion,
but didn't know enough to come in from the rain. How would we program sentience?
I'm not sure what you mean
by "sentience" but programming self-awareness is a trivial problem. You
simply put a robot in a room, and give it tokens for itself and other objects.
If this machine can describe external objects, why couldn't it describe
itself ? I think the problem is to make a machine that can be aware of
anything at all. As a computer's knowledge of the outside world increases,
it's only natural to turn inward and describe itself. I don't think this
is inevitable, because it all depends on how you design it. You don't have
to make clones of human intelligence, I just don't think that doing so
is particularly difficult. If you have a sophisticated system that can
describe the world, and itself, that is all that self awareness is.
Question 6: What about
Doug Lenat and the CYC project? Do you think that this project, which feeds
general information about the world into a computer, will lead to intelligence?
CYC is a piece of the puzzle.
I think what Lenat is doing is valuable, Lenat is a smart guy. The progress
that he has made hasn't lived up to his predictions, but that isn't surprising.
The whole field of AI is littered with unfullfilled promises, but Hans
Moravec deals with this issue, noting that in the 1970s computer researchers
realized that they didn't have the computer horsepower to do things like
computer vision. But Moravec notes that given Moore's law, in a couple
of decades we should be able to do
these things. I don't think
that his time frames are necessarily accurate, but I do think that his
basic technological predictions are accurate.
I think, however, that Moravec
is crazy, when he actually talk about robots being our children,
that they will replace us. He also talks about transfering consciousness,
which is ridiculous.
Question 7: Many people
have argued that the pace of technological change is increasing. Do you
agree?
You do see exponential trends
in certain areas of technology. But if Drexler's nanomachines work, then
you would have quite a bit of change. Kurzweil notes that the overall change
is a cascade of "S" curves, with early exponential growth, followed by
saturation and plateau. Then you have another "S" curve replace it. There
is no general law of accelerating returns, but we are facing in the next
few decades a time of very great technological change, primarily driven
by nanotechnology and microelectronics.
Question 8: What is
your opinion of Drexlerian nanotechnology?
I don't know if it is possible
or not. It may very well turn out to be impossible. I don't know. The people
who claim to know are wrong, we won't know one way or the other until more
research is done. I think that Drexler and Merkle make a good case for
assemblers. However, there has never been a simulation of an assembler
undergoing replication. If such a computational simulation were to take
place, it would prove two things: it would prove that that particular design
would almost certainly not work in reality. But a very slight modification
almost certainly would work.
Question 9: What about
the August 2001 issue of Scientific American?
Much of the article was critical
of Drexlerian nanotechnology. There was very little science in that issue.
Smalley and Whitesides make some good arguments, but there is entirely
too much posturing going on by everyone. Smalley raises a good point: can
all the steps of molecular assembly be performed without requiring 10 "fingers".
Drexler shows that there are certain designs that don't need 10 fingers.
What needs to be shown is that all the steps of replication for one of
his assemblers can occur without needing 10 fingers. I don't know how Smalley
can be certain that these assemblers are impossible.
Question 10: What about
the objections of individuals such as Lyle Burkhead? He argues that nanotechnology
will not be used primarily because it will not be economically beneficial
in the majority of circumstances.
His basic premise is that
Drexlerian nanotechnology won't work because it is so different than our
current manufacturing system. If you think about doing something like growing
a building, pouring concrete is not so hard. On the other hand, if you
develop self-replicating systems, which you may want to do in order to
develop molecular computers and other components, it makes sense to develop
an assembler. Once you have an assembler, why not use it to grow buildings?
It certainly is more convenient.
Question 11: I want
to ask you about Ray Kurzweil. Is it fair to say that you agree that many
of his predictions are accurate, but you have moral issues with his arguments?
I have issues regarding his
timescale. For instance, he says that non invasive brain scanning will
soon be feasible. I think that's
ridiculous, I don't see
that as happening. In order to obtain a molecular scale map of the brain,
you need to cut it up, you'd need to kill the person. What Kurzweil is
proposing is that the human race will be extinct within 100 years, and
he advocates this. In place of people, there will be computer simulations
of people. In the end there may not even be these simulations, just computer
processes as part of some super intelligence. And Kurzweil thinks this
is cool, it's sexy, it's spiritual. He actively proselytizes for that end,
for the extinction of the human race. He describes it in religious terms,
as a sort of ascension to cyberheaven. I think that this is fundamentally
wrong. It's a symptom of a deep disorder.
 |
| About Us | Advertise | Contact
Us By Email: calin [at] nanotech.biz |
|
| Copyright © 2005 Nanotech.Biz |
| Disclaimer: No content, on or affiliated with Nanotech.Biz should be construed as or relied upon as investment advice. While every effort is made to ensure that the information contained on Nanotech.Biz is correct, the operators of Nanotech.Biz make no warranties as to its accuracy. In all respects visitors should seek independent verification and investment advice.
Investing in nanotechnology, investing in nanotech, where do i invest in nanotechnology, how do i invest in nanotechnology, nanotechnology companies, nanotech companies, nanotech products, nanotech stocks, nanotechnology stocks, nanotechnology investment, nanotech investment, nanotechnology investing, nanotech investing, nanotechnology reports and white papers.
|
|