At 1:50 AM -0700 6/6/99, jonathan_desp@...
>True wet nano is good for the present for nanotech but not for the future
>nanotechnology will going give us the power to innovate much quickly and
>we will have the possibility to change our body. we will want to change
>our old neuron for better nanocomputer. BUT a problem happen because the
>neuron is not really complex.. but their interaction yes ! and if we want
>to have a better brain with nanocomputer we will must construct the
>interaction too and so.. can we ? with software i don`t think so.. it will
>become hazardous and maybe the creation of this brain will can control us
if we want construct a nanocomputer brain with the same co!
>mplexity of our brain,interaction.. well at the end of the experience we
>will be much smarter than the nanocomputer brain.. and so, as we remember
>at the first moment of this experience the goal of this experience was to
>construct a much smarter brain.. foe put it in the place of our old
>brain... that`s a big problemof nanotechnology.. and this problem will
>appear not just with the brain.. with our cells too.. the interaction of
>our cell is really complex.. well yes..... our cells alone is maybe less
>compless than our future design of nanomachine but their relations,
>correllations,their inter-relations.. them ?
I believe that the issues that you raise point to the importance of the
early development of high level artificial intelligence (AI) engineering
design systems. The relationship between the development of AI and the
development of nanotechnology (NT) is complex. On the one hand, NT will
make possible much more powerful computers on which to implement AI, thus
facilitating the development of AI. But I believe that the role of AI in
the development of NT will be even more important, even if the AI systems
are implemented on current day computer systems and the near term
improvements to those systems that will be made without NT.
An important point to note, as Drexler has argued in _Engines of Creation_,
is that the most useful AI systems will not be ones that appear to be
human, with all that implies with respect to personalities, emotions,
ambitions, etc., but rather the most useful systems will be those that are
idiot savants - systems that are very good at solving specific technical
problems, and solve those problems much better than do humans. I think such
systems will be essential for the application of nanotechnology to cellular
repair, reanimating cryonically preserved humans, and enhancing the human
Consider if someone tomorrow developed a machine capable of atomically
precise positional control of mechanochemistry, and also capable of
self-replication. With such assembler-replicators we could immediately
implement the construction of vastly improved materials, such as building
materials with strength to weight ratios 50 or more times better than
steel, etc. With a modest amount of programming, we could make most
consumer goods and industrial commodities very cheaply and very
cleanly,etc. We could also start to build much more powerful computers.
But the application to nanomedicine, beyond perhaps simple elimination of
viruses, microorganisms, and cancer cells, is likely to involve programming
complicated enough to be beyond our current methods of writing software.
Perhaps some simple regeneration can be accomplished by simple programming
of nanobots to provide cell growth fators, or to use conventional
microsurgery to implant stem cells in the right places. But really
understanding how the trillions of cells in the body interact, using the
100,000 or so genes in each cell, will probably require very sophisticated
analysis and programming that is beyond human capability. That would almost
certainly be the case in understanding the interactions among the 10E15 or
so synapses in the human brain.
We need programs that will make, or help us make, more complex programs,
leading eventually to very powerful AI engineering design systems. Dr. J.
S. Hall has touched upon this issue in a recent book review:
So, to return to the issues you raised: wet nanotechnology produced through
evolution will eventually give way to more powerful nanotechnologies, and
how do we interact with, make use of, and benefit from such technologies
without risking those technologies taking control from us.
I think the way to go is through Open Source (http://www.opensource.org/)
development of distributed systems (over the web) to develop better AI
tools to facilitate the further development of even better AI tools, and
also to facilitate the engineering development of NT. And developing
software systems to control vast networks of nanodevices.
There are lots of things that one can think of to do to try to develop
nanotechnology. Lab work is undeniably essential, but requires a certain
minimum of capital (maybe a few $100K, maybe more). My guess is that a few
dozen very smart and determined programmers, including those without any
formal credentials, could do a great deal on the AI development path with
part time effort and without any substantial capital.
Why Open Source? First, it provides a way for a very loose network of
workers to collaborate without formal organizational arrangements. Second,
AI development that "runs away" leading to a Terminator-like scenario is a
risk that needs to be considered. Having the software along the way be Open
Source provides much greater opportunity to catch errors and dangerous
programming in time to correct it.
Just some thoughts from an ex-molecular biologist who admits to knowing
very little programming, that is, no programming except html.
James B. Lewis, Ph.D. James B. Lewis Enterprises
7527 40th Avenue NE, Seattle, WA 98115-4925
alternate e-mail: nanojlewis@...
World Wide Web: http://www.halcyon.com/nanojbl/