Just a couple of observations.
Fred asks how the nylon-digesting enzyme would be reported. Let's assume that we are presented with the nylon bacterium and its carbohydrate eating progenitor devoid of historical knowledge of which food source was available first or which bacterium arose from the other. We could recognise that the two bacteria were very closely related and a gene differed by only one base (addition or deletion?) But which version would be the true version and which one would be the "broken" version? Fred is only making this judgement because we are aware of the historical context. But lose the historical context and add a thousand years of mutation and evolution, how could we distinguish between the "true" code and the "broken" code.
Fred finds support in the comment that "its been observed "more than once"". But were the other occurences the identical mutation or can a variety of mutations produce an adequate protein configuration for metabolising nylon? Obviously if any of the mutations resulted in a severe disadvantage (inability to metabolise carbohydrates) without an offsetting advantage (ability to metabolise yet-to-be-invented nylon) they would be removed by selective pressure.
If as Fred suggests, "certain portions (hot-spots) of the genome have been pre-programmed for hypermutation during environmental stress to toggle bits in an effort to find a possible combination that improves survival in the degraded environment", why does this not occur with all individuals in the bacterial population? Why does the switch only flick (or exist) for a select few? And if such versatility is the produce of intelligent design, why is the alternative state so inefficient? Why not have an efficient carbohydrate metabolising gene and a seperate, efficient nylon metabolising gene to cope with the possibility of a degraded environment?
Fred says "But it is clear that 0 is the favorable state (in general), and the study confirms that this method is a whopping 98% more efficient. So based on this knowledge, we actually lose information when we go from 0 to 1."
Hmmmm. How much "information" did we start with for a nylon metabolising process? How much "information" did we end up with for such a process? Seems like a net increase in "information" on nylon metabolism.