Golem (Genetically Organized Lifelike Electro Mechanics)
| January 17, 2007 | Posted by William Dembski under Evolution, Intelligent Design |
Does the following point up disconfirming evidence against the creative power of unguided evolutionary processes? What has become of this project to vindicate standard evolutionary theory?
The golem@Home project has concluded. After accumulating several Million CPU hours on this project and reviewing many evolved creatures we have concluded that merely more CPU is not sufficient to evolve complexity: The evolutionary process appears to be hitting a complexity barrier that is not traversable using direct mutation-selection processes, due to the exponential nature of the problem. We are now developing new theories about additional mechanisms that are necessary for the synthetic evolution of complex life forms. Some of these new mechanisms are based on ideas of modularity, regularity, hierarchy, symbiosis and co-evolution. These ideas are resulting in a new generation of artificial-life systems. If you are interested in our developments, please follow our recent publications at Brandeis DEMO Lab and Cornell CAD Lab. We thank the more than 30,000 participants for their assistance, and hope it was an enlightening experience.
SOURCE: GO HERE
35 Responses to Golem (Genetically Organized Lifelike Electro Mechanics)
Leave a Reply
You must be logged in to post a comment.
You know, EvoDevo, homeotic genes, acquisition and fusion of genomes, etc. Personally I think discussing RM+NS is like beating a dead horse even though there’s still many Darwinists that support it. I’d rather move onto discussing these supposed “engines of variation”. I have a feeling they’ll get more popular so think of it as a pre-emptive strike.
Oh, and for those who never noticed it I put together an OE article a while back that is relevant to this topic:
http://www.overwhelmingevidenc.....k/ga_chess
Dave,
There are two levels at which GOLEM was too simplified. One is, as you mention, that it used only RM+NS, which is adequate for evolution in general but not for modelling biological evolution, which includes a lot of other mechanisms. The RM+NS model is Darwinism, and it is about a hundred years outdated.
The other level is that the GOLEM authors used a designer approach. Their simulated genes coded directly for features, such as joints and bars, on the simulated organisms. The complexity they could reach was thus limited by the limited number of features used. If you specify a given number of arms and legs you can never get an eyelash. Biological evolution works at a much lower level, and genes are very distant from arms and eyelashes. This allows for a huge number of features available in the phenotype because these features correspond to combinations of genes. If instead of, say, 10 genes coding for 10 traits, GOLEM had used 10 genes that built traits in combinations of 3 they would have 1000 traits available with the same number of genes. The ratio in biological systems is generally much higher than this.
Bfast,
At the fundamental level, evolution is chemistry. You cannot have a simple RM+NS=evolution model because the system is very complex. It’s like space exploration, or weather forecasting, or polymer chemistry, or nuclear power stations. Complex systems require complex models even if the underlying natural laws are the simple chemistry and physics laws. And yes, as it generally happens in chemistry, you need the right ingredients in the right amounts to get the right results.
Also, any model that includes RM+NS does demonstrate evolution. Gene frequencies in the population change over time and the performance in whatever is being selected for increases. However, performance will stop increasing once the limit of the model is approached. And the same probably happens in biological evolution. Organisms can only be so fast, so strong, so quick to reproduce, and so complex. There is no reason to assume that biological evolution has no limits – it just has limits far beyond the limits of a simple model that puts together a few simple pieces.
ludwig
There are two levels at which GOLEM was too simplified. One is, as you mention, that it used only RM+NS, which is adequate for evolution in general but not for modelling biological evolution, which includes a lot of other mechanisms. The RM+NS model is Darwinism, and it is about a hundred years outdated.
Precisely what are these “lot of other mechanisms” and which of them do not depend on random mutation as the ultimate source of variation?
You’re flat wrong about Darwinism. NS is Darwinism. RM+NS is NeoDarwinism or The Modern Synthesis and its beginnings were about 70 years ago and it was fully fleshed out in the gene centric theory less than 50 years ago when Francis Crick came up with the central dogma of molecular biology.
Maybe you should read more and write less, Ludwig, until you get more of the basics down.
GOLEM was a wonderfully illustrative failure. It illustrated the bottlenecks in complexity that are hit when you design without being able to use abstraction. Trial and error techniques with feedback, which is precisely what all natural (unintelligent) evolutionary mechanisms are, with every kind of feedback, cheating, and importing of information mechanism that generations of exceedingly bright engineers and programmers could imagine, have been used to further Computer Aided Design and Computer Aided Manufacturing (CAD/CAM). Perhaps the toughest job undertaken with the biggest reward for improvements is printed circuit board layout. Thousands upon thousands of interconnections in 3 dimensions, a virtually infinite number of possibilities, and no way at all to find the most optimal routing pattern. Even WITH a human guiding the best auto-routers it’s a tough job and the heck of it is that a PCB isn’t anywhere near as complex as bacterial DNA and a ribosome.
What GOLEM managed to design with the simple goal of moving the fastest across a flat plane is laughable. It didn’t even invent the wheel. How pitiful is that?
I think that getting Darwinists to admit that biological reality is not as simple as the old paradigm is quite a victory in itself. Instead they must retreat to a highly complex set of interacting mechanisms working in a chaotic system. Never mind that research into these proposed mechanisms has just barely begun and evidence is currently lacking. Does it hurt that much to admit that their case is currently lacking but they believe additional research will bare fruit?
Yes.