It reminded me of habits and processes I've seen in non CS departments at my local (large) university. It's extremely common for students to work as TAs or as paid researchers on projects to support their own research.
Contrast this with my experience in grad comp sci studies, where almost everyone was attached to a funded research project and you typically tried to find some funded project that was a decent match for your interests. Very rarely (never?) did a student in my department "work" on a funded project and then do their personal research in an area different than that. Grad students often hacked around on the side, but weren't doing deep research, writing papers, or planning dissertations that weren't part of their funded research project.
So maybe, the idea of professors doing unfunded research is simply a continuation of the normal process - the "day job" that lets you do what you really want to do on the side.
I hope that the "sorting students by using poor teachers" is simply speculation on the parts of the authors. I like to think that the role of a teacher in the classroom is to continue to struggle to best teach the students who show up and show the passion and effort to learn.
I actually skimmed over the section in the paper linked from the article and the paper authors say "Researchers may be poorer teachers to low-ability students and thus better screeners" and are thus providing benefit to employers who can avoid having to screen out lower ability employees in their field. Maybe the real idea is to make sure only those students with passion and are willing to make big efforts to learn show up in the first place and the linked paper talks some about this. I just hate to think teachers abandon well intentioned students as part of some screening process. I'm not suggesting that screening type courses aren't OK (you have to appreciate that some students will have an aptitude and some will not for a specific subject) but rather that it would bug me for a student with the aptitude to not be able to count on an honest teaching effort from a professor.
Otherwise, we might as well chuck the college/university system period and learn on our own or by finding a solid mentor/apprentice setup. Come to think of it, maybe that's what we should do anyway. The PG essay "After Credentials" expresses the logical conclusion of that idea pretty well. It's why I think more about my portfolio of work more than my degree collection and sometimes wish I had spent more time on that portfolio instead of working from a very grade-focused viewpoint during my undergrad and grad school time.