Kinematic Self-Replicating Machines
© 2004 Robert A. Freitas Jr. and Ralph C. Merkle. All Rights Reserved.
Robert A. Freitas Jr., Ralph C. Merkle, Kinematic Self-Replicating Machines, Landes Bioscience, Georgetown, TX, 2004.
6.2.1 Design Precedes Construction
First and foremost, it seems unlikely that we can build an assembler in the absence of a design. It is difficult to know how long it might take to develop an assembler (including not only design but construction of early prototypes), but estimates are generally measured in decades. While the authors believe that one to two decades is plausible, and while most estimates range from one to four decades, a minority of knowledgeable observers claim that it will take longer.
Whatever the development time, it should be possible to shorten it significantly by the widespread availability in the technical community of one or more feasible designs. These designs would permit a more careful analysis of the capabilities required for construction, and would therefore focus attention on those experimental approaches that could most effectively speed development. A design believed to be correct, but as yet unbuildable using present-day manufacturing techniques, can enable further design efforts in which the minimal design is backchained toward currently accessible manufacturing techniques such as scanning probe microscopy . As Merkle  noted in a related context: “The direct manufacture of this ‘simple’ system with current technology seems unlikely to be feasible. As a consequence, it will probably be necessary to develop other systems that are easier to synthesize using today’s methods, but which are sufficiently powerful that we can use them to synthesize more sophisticated systems. This might be likened to the ascent of a tall mountain in stages, with base camps established at intermediate elevations. The precise nature of these intermediate stages depends on the design of the final stage. The present proposal might be likened to a final base camp, close enough to the peak that it’s clear that a final assault from this final base camp would reach the peak, but far enough removed that it’s significantly easier to reach the base camp than the peak.” It is useful, of course, to recall that just as in mountain climbing, one should not confuse mere glimpses of the summit with proximity to that goal.
As Drexler  explained in 1992: “In a forward chaining search (as the term is used in computer science), one pursues a goal by taking steps that may lead toward it, sometimes exploring blind alleys. If, however, all possible destinations are considered equally good, then there is no real goal, hence (by this loose standard) forward chaining never fails. In science, it is common to pursue experimental programs opportunistically, choosing next steps based on immediate prospects and a sense of what is interesting, important, and fundable. This process resembles forward chaining with abundant, unranked goals, and it routinely produces incremental advances in knowledge and capabilities.
“In backward chaining, one first describes a goal, then searches for intermediate situations one step removed from the goal, then for situations one step removed from those, and so forth, planning backward toward situations that are immediately accessible – that is, toward potential first steps on an implementation pathway. If there are many potential first steps, then backward chaining can be particularly attractive. In technology, it is common to select a goal based on its near-term feasibility and economic attractiveness and then plan backward to select the necessary parts and procedures. The enormous range of modern technological capabilities often provides many possibilities, hence engineers are more often concerned with cost and performance than with feasibility alone.
“Molecular manufacturing is a technological goal (representing many scientific challenges), but it cannot be achieved in a few steps or a few years. Accordingly, one cannot expect to succeed by combining just existing parts and procedures. Backward chaining is still appropriate, but the links in the chain are intermediate technologies, not mere parts and assembly procedures.”
It is likely that the existence of one well-developed proposal, with its specific parameters and results, would inspire efforts to produce competitive proposals (whether related or diverse) that are improvements in some way. This intellectual competition would contribute to further useful explorations of the design space.
Last updated on 1 August 2005