Physical complexity. The arrangement of molecules in the computer or the brain.
Complexity refers to interconnectness and intereaction of systems. The more connected, the more complex.
The amount of information a network contains is constant if all nodes in that network have a value regardless of whether those values have meaning to a human or other intelligence.
How are you defining information? The amount of information most certainly can and does change, otherwise, no one would ever learn anything.
http://en.wikipedia.org/wiki/Information
http://en.wikipedia.org/wiki/Kolmogorov_complexity
http://en.wikipedia.org/wiki/Information_theory
You keep bringing up "interaction". If it's essential for you to know where the interaction is, it's between the molecules making up a compound and between the atoms making up the molecules, and between the sub-atomic particles making up the atoms.
I am referring to the definition of "complex" which you keep avoiding. Of course it's essential for me to know that. That's precisely why I do, in fact, know it. It's also essential for you to know what complexity and information are, because it's clear you don't. I provided some handy URLs above to get you started.
Those interactions will of course change when data passes through them, but the number of interactions between particles, the complexity, is not increasing. In order for this to be the case, the network would have to develop new nodes.
Yes it is. Let's go back to your CPU example, or even that of a hard drive example. There are more interactions happening at various states of energy and as more and less work is being done. Obviously, powered off is a less complex state. Even less is disassempled. Even less would be the atoms that make up those parts floating millions of years apart in deep space. Same particles but clearly with different numbers of interactions happening and different rates of information, contrary to your assertion. CPUs can be off, idle except for the most bare work to remain powered up or at various percentages of activity. Same thing for a hard drive. The actuator arm can be idle or thrashing, interacting not at all or a lot or somewhere in between, changing the atoms on the platter or not. A CPU can be manipulating signals and changing the values of memory or opening and closing logic gates... or not.
Developing new nodes on a network is irrelevant to your CPU example.
You still seem to be saying complexity and information can't change, that it's simply a measure of how much material is in a given space. The question remains, how are you defining complexity? While you're at it, information also.