Darwin Credited with Intelligent Design
Exclusive In an unusual mix of metaphors, a pioneer in advanced computing technology gave a lecture at JPL on 11/14 entitled, “Intelligent Design: Using Evolution to Create Complex Systems.” By evolution, it was clear that astronomer Dr. Richard Terrile meant the Darwinian kind. The point of his presentation was that humans can mimic the processes of Darwinian evolution – random mutation and natural selection – to build more powerful problem-solving systems. After all, that’s what biological evolution did, he claimed, when it created eyes, hearts, brains and the most complex, optimized structures in the world.
Dr. Terrile noted how a Macintosh laptop resembles a human brain in terms of mass, volume and power requirements, but falls short of computing power by 1/50,000. Moore’s Law assures that we’re catching up, though. Unfortunately, he said, software design has not kept up with the explosive growth in computational power. The future lies not in staying the course with outmoded methods of software engineering, but in mimicking how nature solved complex problems: by evolution.
The evolutionary process begins with an informational system – a genome.1 From there, the algorithm is simple: mutate, select, repeat. Out of the astronomical random combinations that result, some .02% may be beneficial. Evolution is very adept at sifting out the worthless variants and capitalizing on the tiny fraction of useful ones, he explained. Given the Himalayas, evolution will find Mt. Everest. Not only that, evolution itself evolves. On a second level, evolution can help a system decide the best way to select the optimal variants. Applying evolutionary computing to data analysis, for instance, can discover optimum ways to perform data analysis.
With a touch of humor, Dr. Terrile recognized that his mixed references to “intelligent design” and “Darwinism” might be a little inflammatory, no less his title slide with the image of Michelangelo’s creator. It was apparent throughout the presentation, though, that he really believed that the methods his new Center for Evolutionary Computing is employing exactly mimic the processes of biological evolution. The audience of geeks and gurus interested in this cutting-edge approach to complex systems development, sitting intently, seemed to accept this premise without question.
1Dr. Terrile caught himself assuming the prior existence of an information system. He discussed briefly how the universal genetic code spelled in DNA is optimized against point mutations. The chance of that arising naturally, he conceded was one in a million. Therefore, “there must have been some pre-evolution on a molecular scale,” he speculated. See online book for problems with natural selection acting prior to replication.
If you need evidence that some of the smartest people in the world can believe the dumbest things, look no further than this case. The methods of “evolutionary computing” have nothing to do with Darwinian evolution and everything to do with intelligent design. What was doing the selecting? Intelligence. What was designing the “fitness function”? Intelligence. What was setting the design requirements, the constraints, the algorithm, and the target? Intelligence. Who built the hardware and the software? Who had a purpose and organized the ingredients toward that purpose? Who decided at the end of the run that the outcome was suitable? Human minds. This was the work of intelligent design from start to finish, from input to output, from concept to reality. Darwin had nothing to do with it.
When you add intelligent causes to Darwinism, it ceases to be Darwinism. Charlie envisioned a world of unguided, purposeless, uncaring, random processes with no goal: i.e., evolution, not evaluation. If these experts want to imitate Darwinism, they should take their intelligent hands off the equipment and let come what may. If their computers go extinct, so be it. But to claim their highly-advanced, skilled, goal-oriented efforts are Darwinian is absurd. This is not evolutionary computing, but the latest in iterative and recursive design engineering. Crediting Charlie for this is like crediting Las Vegas for the space program. (That’s Las Vegas, incidentally, with the people removed, and the machines running themselves. Better yet, Las Vegas before man arrived—just a rocky desert receiving copious amounts of energy from the sun and earthquakes—perfect conditions for the emergence of complex systems.)
Dr. Terrile felt it politically expedient to disparage intelligent design several times in his lecture. He inserted remarks to assure the audience he was using the phrase “intelligent design” only in a humorous, provocative way, making it clear he was not one of “them” but is a certified Darwinian. His slides included the icons of biological organisms rising higher in complexity, from bacteria to brain, paralleling the evolution of computer technology (the fallacy of association). Yet the substance of his methodology and technology was all about intelligent design, not evolution. Assuming that the guidance of random variations toward design goals is evolutionary, not ID, is worse than shooting oneself in the foot. It is shooting oneself in the brain.
One of the best critiques of “evolutionary algorithms” is No Free Lunch by William Dembski. A critique of digital evolution by Royal Truman can be found at the ISCID site. You can also find prior entries on digital evolution from 05/08/2003, 07/04/2004, and 08/20/2002.
And now, ladies and gentlemen, you can do your own digital evolution experiment online! Run Perry Marshall’s Random Mutation Generator. As with all computer games, be sure to read the instructions.