6.3.6

Kinematic Self-Replicating Machines

Robert A. Freitas Jr., Ralph C. Merkle, Kinematic Self-Replicating Machines, Landes Bioscience, Georgetown, TX, 2004.

6.3.6 Potential Design Errors Make the Analysis Inherently Worthless

A sixth criticism is that an error in some part of a preliminary design could make the entire design unworkable, hence the effort is inherently worthless. Almost certainly, some aspect of any preliminary theoretical design will eventually prove to be unworkable when the design is subjected to greater theoretical or experimental scrutiny. Hence the argument is advanced that any attempt at a design is worthless, because one or more hidden flaws in the specific design might make the entire enterprise infeasible.

This argument ignores the fact that there are usually many possible solutions to a given technical challenge, and more than one way to provide a needed capability. If one approach doesn’t work, then another means may be found. The initial design effort then will have been valuable, both because much of it was later found to be feasible and because the precise specification of subsystems allowed and encouraged closer examination of those subsystems, the discovery of the errors within, and the avoidance of that error by other means not originally contemplated. The design will also be useful if it helps to illuminate broad volumes of the design space in which diverse useful possibilities may reside, not all of which suffer from the same identified flaw.

Drexler [208] offers an example of a proposed nanomechanical system that consists of five essential subsystems, each serving a particular function, including: (1) a motor, (2) a power supply, (3) a vacuum pump, (4) a pressure sensor, and (5) a gas-tight wall. If we assume that for each subsystem there are 10 equally plausible engineering possibilities that are not mutually exclusive, and if each possibility has an independent 50% probability of working, then the probability that all ten options will fail, leaving no workable choice for a particular subsystem, is only (0.5)¹⁰ = 0.001, and so the probability that a successful combination of all essential subsystems exists is (1 – (0.5)¹⁰)⁵ = 0.995. Hence a near certainty may emerge from a combination of possibilities, each of which, individually, is as likely to fail as to succeed.

This argument can be made even stronger when the design itself systematically adopts approaches that can be solved by any of a wide variety of approaches. Present assembler designs are based on the feasibility of subsystems, each of which can be implemented by a wide variety of different methods. For example, the use of positional assembly requires only that some molecular-scale positional device be feasible – if the specific design proposed has some hidden flaw, it can readily be replaced by some other design. As the reader can well imagine, there are a vast number of possible robotic arms, and a vast number of possible molecular positional devices. Only if all possible robotic arms proved infeasible would the system-level proposal fail from this cause.

Similarly, there is a vast space of possible chemical reactions that could be used to build the system. While researchers to date have focused on specific reaction pathways (the better to analyze them in detail), there are many ways to synthesize stiff hydrocarbons and even more ways to synthesize alternative stiff materials which would be suitable for the manufacture of the mechanical design proposals described here and elsewhere. Failure of a specific reaction to work as expected would merely result in the adoption of an alternative synthetic pathway, and even if all methods of synthesizing stiff hydrocarbons were eliminated (which seems a remarkably unlikely prospect, given the wide range of routes by which such structures can and have been synthesized) we would still have available the rest of the periodic table and the resulting combinatorial explosion of possible structures that it enables, from which to choose.

Last updated on 1 August 2005