“Pay the Prize”
The author writes: …then why don’t engineers use Darwinian evolution to design cars or write software? … I am offering an award to the first person who can discover a process by which nonliving thins can create code.
The answer is: that person already exists, his name is John Koza and his process is called “genetic programming”, which is used to design engines, pictures, music, computer code, etc by itself. You can learn buying the John Koza books here in Amazon. So, the author can pay the prize.
The author is missing the truth when he calls these developments “curiosities “. The truth is that they are used in many industries. By rejecting the already known, the whole argument of the book of collapse as a building of cards.
By trying to take advantage of the ignorance of both engineers and scientists of genetic programming, he only manages to show his own or even worse, his bad intentions, improper in a true scientist.
Engineers DO use Genetic Algorithms. But Mr. Sojo clearly did not read the book he is reviewing (!) because I devote chapter 25 of Evolution 2.0 to this very subject.
(His review is not a “Verified Purchase.” The majority of 1-star reviews never read Evolution 2.0. )
The problem is, none of these genetic algorithms operate strictly according to the rules of old-school Darwinian evolution.
In old-school Darwinian evolution, there are no pre-programmed goals, and mutations are random. To old school Darwinists, “Natural selection is the only game in town” to quote Jerry Coyne.
But the following are true of ALL genetic algorithms:
1) GAs never work unless you precisely define a “fitness function” in advance. The program is ALWAYS working towards a goal that a human has designed.
2) GAs never work unless the mutations are carefully controlled and constrained. You can’t randomly vary just any part of the program. If you’re designing car engines and you’re trying to optimize the diameter of the cylinders, then you have to introduce a variable called “diameter of cylinder” and adjust a collection of related variables together.
3) GAs never work without code that a person has programmed into the GA itself. Thus no existing GA could possibly qualify for the $5 million Evolution 2.0 Prize. The prize insists on chemicals to code with no cheating. All GAs cheat.
4) All GAs have to be babysat by highly-skilled, highly-educated, highly-paid, staff members.
5) GAs are little more than a footnote in the software industry. If you go to any software trade show or conference, you are unlikely to find more than one or two booths by GA companies. GA is no panacea. GAs are difficult to work with. Most GA startups have failed.
Often it is cheaper, faster and more straightforward to just hire a programmer and write the code from scratch than to try to rig up a GA. GAs are used for optimizing when you can define a goal and specific variables, when human engineering can’t predict the various combinations.
The limitations of GAs are superbly outlined on Wikipedia:
Now the most important thing I want to say is that the above line of argumentation is usually made for a creationist or traditional Intelligent Design position. This is usually anti-evolution.
That is NOT the argument I am making.
Because what almost everyone seems to ignore is the fact that we observe cells in real time generating resistance to antibiotics, producing hybrids, symbiotic mergers and new species all by themselves and doing it in real time.
Denis Noble of Oxford documents this very well in his book “Dance to the Tune of Life: Biological Relativity.”
Cells evolve. Organisms evolve. And they choose their own goals. The real punch line is that if we truly understood cells, biology and evolution, our GAs wouldn’t require babysitting by guys and gals with masters degrees in Computer Science.
GAs would evolve by themselves. The way cells do.
When Frontline Genomics interviewed me about the Evolution 2.0 Prize, I said, “Bacteria re-arrange their DNA to fight antibiotics in minutes. Cell AI is 1000X superior to anything from Silicon Valley. We need to discover what makes this possible. If Microsoft knew what bacteria know, their stock price would spike 10X.”
In Evolution 2.0 I describe how cells harness a toolkit consisting of Epigenetics, Transposition, Horizontal Gene Transfer, Hybridization, Symbiogenesis and virus activity to engineer new code all the time. GAs use very similar mechanisms.
If we truly understood evolution, if we weren’t just sweeping the world’s grandest mystery under a big rug called “random mutation and natural selection” or abdicating to “Intelligent Design” we would birth multiple billion-dollar industries.
I predict this WILL eventually happen. Eventually both of these camps will recede. This has to happen because if we don’t fully understand evolution, we’ll never understand cancer, disease, aging or AI. The rewards for people to figure this out are too great for the entrenched old-line positions to stay in place.
My book is subtitled “Breaking the Deadlock Between Darwin and Design” because the Darwinists say Chance did it (not the case) and the Design guys for the most part say evolution is a hoax and God did it.
Darwinists underestimate nature. Creationists underestimate God.
BOTH sides are preventing us from actually understanding the science. And both sides are wrong. The truth lies in the middle and it is far greater than either side comprehends.