Technology evolves, ergo automation evolves, until...

Jeff Davis (
Mon, 02 Nov 1998 20:13:38 -0800

In his Re: Automated Workforce post of 12 July to the Extropians list Danny <>, wrote:

>Theres also an article in an issue of Discovery which talks about self
>replicating robots in the Sahara desert, which create an endless amount of
>solar panels out of the elements in the sand, including chips,
metal railings,
>factories, furnaces etc... It also said that the technology for this is
>currently available and would supply the whole world with power in a few
>years, only taking up the size of the state of Arizona, I'm not sure how much
>it would have costed, prolly cheaper now (article I read was 3 years ago), but
>definately accomplishable if we really wanted to do it (I think they said
>between 1 and 100 billion dollars). If its global every country could
>contribute who wanted a part of it, it would become free to the consumer.
>Plausible free energy, does anyone have more information on this?
>Specifically the Sahara desert project? We need to start this project soon,
>what is stopping us?

I am particularly drawn to that last comment:

		"We need to start this project soon,
			what is stopping us?"

I, too, strongly believe that a powerful opportunity is staring us in the face. Consider the unremarkable correspondence between Klaus Lackner and myself.

NOTE: Klaus Lackner and Christopher Wendt, at Los Alamos Natl. Labs, <>, are responsible for the journal article--Exponential Growth of Self-Replicating Machine Systems--which appeared in the May 1995 issue of Computation and Mathematical Modeling, Elsevier Pub., and which resulted in the article in Discover Magazine about self-replicating, solar-energy-array-building robots--auxons.

I wrote:

>Dear Mr. Lackner,
> I readily confess to an almost obsessive fascination with the idea of machine self-replication.
> Could you email me a copy of your article "Exponential Growth of Self-Replicating Machine Systems", and any additional materials relating to the subject which you would not feel imposed upon to include.
> Perhaps it is my lack of skill in performing such searches, but I find surprisingly little information "out there" on this topic. I've seen Moshe Sipper's Self Replication page, and read the NASA summer study, and requested info from Bob Freitas, its editor. In the end, however, I'm puzzled by the limited amount of work and, even moreso interest, on the topic.
> In the conventional economic model all costs originate as labor costs. Garden variety automation (the industrial revolution), familiar, and for the most part unthreatening, incrementally reduces the amount of labor cost per unit of production. Ho hum. But machine self-repication appears to take automation to a whole new level, reducing the per-unit labor cost of ALL SYSTEM COMPONENTS ACROSS THE ENTIRE SYSTEM, AND ALL SYSTEM OUTPUT to a vanishingly small quantity. Isn't that the implication, or have I missed something? Doesn't that essentially signal the obsolescence of the old economic order?
> So why no gold rush; no international race to perfect machine self-replication? Doesn't he who perfects such a system get to own everything?...rake in all the chips?
> So, you see why I'm puzzled. What do you think? Where did I get it wrong?

He replied:

">Dear Mr. Davis,
>I essentially agree with your assessment of productivity and cost of production. My only disagreement has to do with the competitiveness of a future economy. II am not so sure that the first one to do it will win everything. My expectation would be that the first system is limited in what it can generate in output. The major output of an early system would be basic feedstocks for industry. Energy, raw materials, basic chemicals etc. The reason is that anything complex we consume is not made in numbers that would lend themselves for this production scale. TVs for example such a system could make overnight for the entire world population. I would expect this to happen over time which in turn gives the chance for many people with innovative ideas and capital to enter into this competition.
>Like you I am not sure why there is so little effort in this area. I suppose for most people this goal is still to far out to be considered realistic. Others have suggested that it would not be to the advantage of a company, but I don't believe that to be correct. The trick is to define intermediate goals that don't have this all or nothing flavor.
>Klaus Lackner

I can't shake the irksome feeling that shortly, someone with the vision and indomitable drive of a Bill Gates will seize this opportunity and make it into a reality, and I will be left thinking, "That could have been me."

	A final note: 
	Nanotech is on its way, and I await its arrival with dynamically optimistic enthusiasm.  Meanwhile...
	The self-replicating solar-energy-array-building "auxon" system envisions producing PHOTOVOLTAIC CELLS in unprecedentedly vast quantities ("the size of the state of Arizona", or "10 percent of the Sahara desert")by currently available technologies.  Photovoltaic cells from single crystal silicon(if indeed they are cells of the single crystal and not amophous silicon type) are produced by a procedure with marked similarities to that used to produce memory and microprocessor ic's.  A manufacturing base which could produce thousand square mile quantities of photovoltaic cells, if converted to the production of computing resources, could be expected to crank out some serious -flops worth of computing power.  Mayhaps enough bring about the nativity of the first generation, somewhat clunky, pre-nano superintelligence?  Just a thought. 
			Best, Jeff Davis

	   "Everything's hard till you know how to do it."
					Ray Charles