Ack! Do we have to go through this again?
Everything is information and no pattern is more special than any other pattern of equivalent Kolmogorov complexity. Thermodynamics encourages the Kolmogorov complexity of systems to increase. Therefore it should not be surprising that the Kolmogorov complexity of information in quasi-stable systems of automata increases as well. It falls out as a natural consequence of the mathematics. You may have a "problem" with it, but that problem lies solely within your prejudices and pre-/mis-conceptions. It is not a theoretical problem. All the handwaving and assertions to the contrary are in clear violation of the mathematics.
The information is always there. Quasi-stable automata systems that operate within useful parameters are not, though they certainly exist in relative abundance in non-biotic nature.
No, not everything is information. Perhaps according to Kolmogorov, but not according to me. The 'pattern' in Don Quixote may be as equally complex according to Kolmgorov's equations than a Tom Clancy novel, however the Don Quixote is more valuable just as the Iliad is more valuable than the latest crime novel. The theory is assigning value mathematically and you cannot assign values mathematically because there is no mathematical way to measure quality.
Here are some other examples which challenge the statement that all information is as valuable as another provided they are of equal complexity. Take this example:
1. the lottery result tomorrow will be 04, 06, 32, 45, 46, 54, 56.
2, the lottery result yesterday was 04, 06, 32, 45, 46, 54, 56.
Now tell me they are both of equal value?????? Not Kolmogorov value, but real life value. Or take another example:
1. someone yelling fire in a crowded theater
2. someone yelling hot dogs on a street corner>
Are you going to tell me that that information is of equal value? Of course not. Would the code for a smiley face on computer screen be as valuable as any of the numerous scientific formulas out there which may have the same Kolmogorov complexity? Of course not because information has to be measured not by its complexity or mathematical features but by its quality and its purpose.
Oh, and there is one problem with Kolomgorov complexity theory - it is descriptive not prescriptive. You can I am sure calculate the Kolmogorov complexity of every single book ever written, however that knowledge would not enable you to write a single one of them (It is like '42' as the answer for 'what is the meaning of life, the universe and everything'). The problem for cellular automata is how to pre-scribe changes not how to describe them.