My Colleagues’ Reaction
“So my great idea is very simple,” I said to my colleagues. “We just have to put our best design into the program, and let the machine run unattended for two or three days. Then we will have a next-generation lens. It will be like a new species, much superior to all existing lenses. We’ll file for the patent, make the lens, offer it for sale, and right away we’ll capture the whole market. We’ll be the only ones that can supply the new design. Our business managers will set the selling price at the maximum the market can bear. Our company will get rich quick. All of us will get a prize!”
My enthusiasm was not reflected in my colleagues’ faces. They knew that the optical physicists had already tried many times exactly what I had just proposed. The program would make appreciable improvements in the first few cycles. But with each cycle the improvements were smaller than those of the previous cycle. After about ten cycles it was not worthwhile continuing. A hundred or even a thousand cycles more would produce little additional improvement. The design always stagnated at a point far from the optimum. It was impossible to find a path to the optimum design, no matter how small the steps were. The designers had to depend on their own intuition. If it seemed to them that the program moved the design in a certain direction but then stagnated, the designers would usually interrupt the program to try a new initial design, from what looked like the most promising starting point. The designers might even add or remove optical elements (lenses or mirrors) from the design. The program could not make such leaps because it lacked the designers’ intuition. The program did not take large steps either because, if it did so, it would lose the path. The program certainly did not use intuition because nobody knew then (and nobody yet knows) how to program the human creative spark.
Thus the design process was really an interaction between a human designer and a machine. The mathematical ideas programmed in the computer could later refine each leap of human intuition. This combination of human intuition and mechanized refinement has improved lenses. But the process has taken years and has never been “automatic” in any sense. Machines cannot yet replace human thought.
Over the years our lens designs had improved, and the company had its niche in the market. But our lenses were very expensive because we invested a great deal of human intuition in their design. There were really very few applications that justified so much precision and refinement. The company had not become rich. Darwinist evolution failed when it was applied to automatic lens design.
“You said that just to make fun of Darwinism!” they accused me. “But we know that evolution has worked, because we and all the other species exist!”
“The existence of many species is not a proof of evolution, if life is the product of some other creative design mechanism,” I responded. “But if any one of you knows why Darwinism works, go and explain the missing ingredient to the programmer. He will gladly write your idea into the program and we will all be rich!”
They went away angry, arguing among themselves. The next day at lunch one of them complained that no one had taken into account my debating skills. I took that as a clumsy, unintentional compliment because I have never had any training in debate. Some of my colleagues insisted on a rebuttal. They sought for a champion to answer me, but they never found one.
I made the presentation on February 26, 1969. From that day on I have never heard of anyone who successfully applied Darwinism to the automatic design of any product, well enough to produce a new generation of the product and capture and dominate the market.
My enthusiasm was not reflected in my colleagues’ faces. They knew that the optical physicists had already tried many times exactly what I had just proposed. The program would make appreciable improvements in the first few cycles. But with each cycle the improvements were smaller than those of the previous cycle. After about ten cycles it was not worthwhile continuing. A hundred or even a thousand cycles more would produce little additional improvement. The design always stagnated at a point far from the optimum. It was impossible to find a path to the optimum design, no matter how small the steps were. The designers had to depend on their own intuition. If it seemed to them that the program moved the design in a certain direction but then stagnated, the designers would usually interrupt the program to try a new initial design, from what looked like the most promising starting point. The designers might even add or remove optical elements (lenses or mirrors) from the design. The program could not make such leaps because it lacked the designers’ intuition. The program did not take large steps either because, if it did so, it would lose the path. The program certainly did not use intuition because nobody knew then (and nobody yet knows) how to program the human creative spark.
Thus the design process was really an interaction between a human designer and a machine. The mathematical ideas programmed in the computer could later refine each leap of human intuition. This combination of human intuition and mechanized refinement has improved lenses. But the process has taken years and has never been “automatic” in any sense. Machines cannot yet replace human thought.
Over the years our lens designs had improved, and the company had its niche in the market. But our lenses were very expensive because we invested a great deal of human intuition in their design. There were really very few applications that justified so much precision and refinement. The company had not become rich. Darwinist evolution failed when it was applied to automatic lens design.
“You said that just to make fun of Darwinism!” they accused me. “But we know that evolution has worked, because we and all the other species exist!”
“The existence of many species is not a proof of evolution, if life is the product of some other creative design mechanism,” I responded. “But if any one of you knows why Darwinism works, go and explain the missing ingredient to the programmer. He will gladly write your idea into the program and we will all be rich!”
They went away angry, arguing among themselves. The next day at lunch one of them complained that no one had taken into account my debating skills. I took that as a clumsy, unintentional compliment because I have never had any training in debate. Some of my colleagues insisted on a rebuttal. They sought for a champion to answer me, but they never found one.
I made the presentation on February 26, 1969. From that day on I have never heard of anyone who successfully applied Darwinism to the automatic design of any product, well enough to produce a new generation of the product and capture and dominate the market.