I'm surprised the author didn't cite Bryan Caplan's work at all, like his most recent The Case Against Education. He expounds the signaling model of education which helps explains exactly what the author here laments:
Judging by how little of many college curricula applies skills required in today's jobs, its a wonder why employers care at all. Unless you consider that credentials, such as undergraduate degree, signal that you're a reasonably intelligent, conformist, and conscientious future employee. These traits are probably largely determined before college.
This is explained by the sheepskin effect: The reward of college is not the education, but the signaling value of the credential. Anyone can get a world-class education at their local library, YouTube University, or, as Bryan Caplan says, you could simply walk into their classrooms not even enrolled in school (the professors would be thrilled by the sight of learning for the sake of learning). Hardly anyone does this. And, importantly, employers aren't fighting over applicants with MOOCs because the signaling value is less clear than a college degree. Caplan cites evidence that finishing 3/4 of college bestows something like 10% of the earning potential of finishing 4/4 of college and getting the degree. That doesn't make sense if you believe the conventional human capital model of education.
Maybe that's because education is so highly expected in the United States--we require and subsidize it more than most other countries--so the cost to not meeting that expectation is higher. But that source of that disparity is the high expectations of educational credentials, not necessarily the fault of the unschooled.
I agree with the author's sentiment that something is seriously amiss. We're screwing a shitload of people out of a good chance at high-paying jobs because they had the misfortune of being born to a family to poor to be able to prepare and pay for their kids to go to college, not to mention all the people who simply aren't cut out for years of sitting still in a classroom. I'm curious what he would think of Caplan's book which addresses and helps explain these facts. (Disclaimer: I'm very skeptical of Caplan's prescription, which is to cut all subsidies entirely so that we return to a lower level of expectation in regards to credentialism. But his book is good.)