In a truly random process you don't see gradual progress fit to serve a teleological end. I don't know if that is Behe's argument, but it is definately mine.
For example, if it were truly random, then we'd expect nubs (early formation of wings) to start growing "randomly" from all areas of the animal's body.
Did they?
Apaprently not. If they did, then they could only be removed by natural selection. So if nubs also grew from the corners of their eye lids, then we can explan why we don't see evidence of this now because that would have made them blind, reducing their chances of survival. But would a nub on their backside hinder their chances of survival? No. So where did these nubs grow? If they were truly random mutations, then we wouldn't expect an orderly appearance serving a specific function.
Instead, they gradually protruded from exactly where they needed to in order for their future generations to attain flight.
EDIT: I decided to look into thie half-wng argument and what follows s a response from an evolutionist:
One of the most common criticisms of evolution is the question "what good is half a wing?" Anti-evolutionists point out that organs must be functional at every stage of evolution, otherwise they wouldn't be selected, and therefore half-formed wings, eyes, and lungs are impossible.
This is, in fact, a valuable insight that tends to get lost in many simple discussions of evolution, so it's small wonder that creationists are puzzled. The reality is there are no half-evolved organisms. Every organism is fully adapted to its environment all the time, assuming the environment is stable. When the environment changes, then it's a whole new ball game and natural selection does its thing. In fact, you can gain valuable insights into how evolution must have happened by requiring that at every step, organisms must be fully functional. You can easily show that some popular stereotypes, for example, lungfish crawling out of drying ponds onto the land, just can't work. Instead, early amphibians had fully developed limbs, developed for locomotion, sensing in murky environments, maybe even clasping during mating, while they were still completely aquatic.
One problem with the half-a-wing criticism is that it ignores exaptation, the adaptation of a trait originally developed for one function to some other function. But apart from exaptation, the half-a-wing criticism is completely false.
Actually, half-formed eyes and wings can be very useful. Any light-detecting ability, however rudimentary, will enable an organism to seek shelter, find food, and avoid predators. Similarly, half-formed wings aren’t as useless as often imagined. The idea that eyes and wings can only function if fully formed is completely false. Indeed, it’s a lot easier to see how partial versions of these organs could function than it is for many other organs. Creationists assume that problems in evolution are insoluble without making even the slightest attempt to see if solutions exist.
That isn't my argument at all. My argument isn't that wings couldn't have evolved because they have to constantly serve a function throughout their development. My argument is that evolution assumes "random" mutations, so why do modifications via "random" mutatons appear teleological?
Why do teeth grow in a shark's mouth and not on his backside? Will teeth on his backside lower his chances of survival? If not, then natural selection says they will remain there, albeit useles.