Now I`m sure I`m overlooking something here but wasnt Dembeskis argument against ToE and for IC based on the difficulty in obtaining a purely random string of numbers?
Some work has been done on using evolutionary principles on data strings to (with a programmable array) produce a working circuit...
http://www.cise.ufl.edu/~nantonio/artificial_evolution.htm
So doesnt this provide a chance to test this, let Dembeski provide a set of data strings that he feels confident are perfectly random then see if they can be evolved into a form that programs a working circuit....