Firstly, do you really believe that it is possible to make any use of Knuth's work "without having ever taken one college class"? You contradict yourself in that you simultaneously indict the "monkeys" who are under qualified to be programmers, yet you then denigrate the very education system which endeavors to produce that same talent. It is, of course, arguable that coding skills are obtained through real world experience, not in college. However, the fundamental theoretical skills necessary to excel in such are obtained in college. Unless you are some kind of natural-born genius, you need to be taught how to think logically, iteratively, and symbolically. That is the reason you were forced to do so much integration et. al. You use those skills all the time; the fact that you don't recognize it is testament to the success of your higher education, not the failure of it.
As to your other retorts: phenomena such as CDEs, component factories, and other productivity gains are real. You dismiss them as binary "works or don't work". This misses the point. Software is a science as much as a trade, and it is evolving, sometimes in revolutionary fits and starts, but mostly slowly and quietly. Of course there are flaws with EJBs, etc. But the fact that they exist at all, and are improving, and even being evolved out of relevance as better solutions are crafted tells the story. The real world, which you so brazenly invoke, lies somewhere between 0 and 1.