In Dembski's book, he uses an f(x) = x function to show how it cannot gain information. But, that isn't a model of evolution because that isn't how evolution works. Evolution increases variation through mutations, recombination, duplication, gene flow, genetic drift, and gene shuffling and then converges via selection. As this is the basis of his book, the entire point collapses.
Erik Tellgren meticulously deconstructed Dembski's law here along with Tom Schneider.
Anyway, complex biochemical pathways have evolved - without intelligent intervention. (B.G. Hall (1982). "Evolution of a regulated operon in the laboratory", Genetics (journal), 101(3-4):335-44.)
You can read more here.
Wheels do not need intelligent intervention to spin, but they generally need intelligent intervention to be built. Organized matter performing specific functions may point to intelligence. That does not mean intelligence resides within the object as it performs a function or remains static. In cases where complex biological pathways evolve, intelligence must be involved in order for the results to be intelligible. Furthermore, the observed material must have properties to make it intelligible, properties which could very well be a product of intelligent design. If one wishes to assert the contrary, he can do so only on the basis of philosophy, not science. At bottom there is nothing inherently supernatural about intelligent design. It happens all the time, and its results are often, but not always, tangible.