Oh, but here's the really unsettling part, dear Patrick: Stephen Wolfram demonstrated evolution from the simplest of instructions, leading to the most unexpected outcomes/behaviors, and "natural" selection had nothing to do with it. Wolfram- (programmer)-selection had everything to do with the establishment of the initial conditions. And what subsequently followed offers an object lesson, so to speak.
Wolfram's cellular automata in their various evolutions (from different instruction sets) seemed to settle into a very small number of general "behaviors" or descriptions. They could rapidly (usually) evolve into a condition of virtually complete redundancy (e.g., either black or white; virtually no information content there); or quickly would begin to display regular and seemingly ad infinitum patterned behavior (low information content there); or they could stun you by surprise with the entirely unexpected and unpredictable evolutionary development they progressivly achieved, in which certain "branches" of the "forthcoming" development died out, while other branches flourished and propagated new forms (extremely rich information content probably involved here).
Please check out the Rule 110 cellular automaton, and then try to tell me that you have not seen Darwin's evolution theory captured in graphical form. :^)
Thanks so much for writing, dear Patrick.
Self organization may play a role in producing change, but selection still shapes the direction of change. It makes no difference what causes or produces the change.