Alex,
Of course, without actually knowing the precise methods used in constructing original code and testing methods, it is difficult to be able to examine the claims in specific detail. However, here are my observations.
There is plenty of information here: http://www.genetic-programming.org/
The first point to be made is that you cannot get away from the fact that program code produced via intelligent means (i.e. human mind) has to be produced first as a vehicle for kickstarting the software 'evolution' process.
Like I said before, you seem to be confusing cosmological arguments with evolutionary principles. In fact, to go even further, so say that active intelligence is required in every case leads to an infinite regress because to explain the existence of an intelligent god you would need an even more intelligent and complex meta-god, and etc.
The first rules of the universe were set up - nobody denies that. Whether it was done by an intelligent being or by accident is not the issue here. The issue is whether evolutionary principles work without additional tinkering once it's started. If you believe it was started by god or some intelligence, fine! That has nothing to do with evolution. Evolutionary principles work without the constant tinkering and redesigning of an intelligence and the amazing programs developed by evolutionary principles prove that nicely.
Every point you brought up has to do with the rules of the universe or the model universe before the evolutionary process was started. Therefore everything you brought up as an issue is invalid. If you can somehow prove that intelligent tinkering with the programs was necessary after the evolutionary process began, then you might have a point. So far you have only misinterpreted or misunderstood the topic.
As you have probably guessed, computer simulations and evolving code, in my view are not legitimate candidates for testing the credibility of randomly generated life, because they are all products of intelligent design. I say this as one who has been involved in both low-level and high-level computer programming - its what I trained for at university
You have not adequately demonstrated any inkling of intelligent design in these programs - you have only demonstrated that intelligence is needed to set up the artificial environment for the simulation. Nowhere in the development of the program is there an intelligence tinkering with the process. Random mutation and selection are all that are in play once the initial rules are set up.
Natural Selection = selection by natural means. The very word "selection" explicitly points to something that has been chosen, which of necessity requires the presence of something able to make a judgement.
Right here you show that you have no basic understanding of evolution or natural selection. To think that some intelligence is necessary to select for fitness is absurd to the extreme! In the real world animals who are more fit for their environment simply reproduce more - it is not an intelligent choice of the environment. In genetic programming, obviously we can't create programs that die of natural causes, so rules are set up to select programs that perform a certain function the best. This simulates fitness in the program's world. This is because genetic-programming is not just a wild process like the real word - it's goal oriented. Interestingly pure evolution simulations, such as Tierra, do just let the programs run wild in the virtual machine, reproduce, and find their own parameters of fitness.
Does this all prove that there is no god that started the Universe? No. All this proves is that an ongoing intelligent tinkering is not necessary for complex, 'intelligently-designed-looking' things to evolve once the initial environment is set up.
rem