I'd be surprised if someone from http://csl.stanford.edu wrote that article. My point is that it matters a lot what area you did your PhD in and not just that you finished a PhD in CS. If your area was Machine Learning and you can't get a data scientist job then you obviously messed up. If your area was networks/systems and you can't get a programming job then you obviously messed up! Those cases are very clear (and rarely ever happen). But you can't take a specific case of working in an area outside of core systems or ML and then try to generalize it as the "PhD -- industry gap". CS is an exception, and not an example, of the PhD--industry gap. Try talking to grad students outside of CS and you'll see how envious they are of CS PhDs finding industry jobs.