I honestly don't know what the job market is like for a CS PhD, so I can't rebut this statement. On the other hand, we interview a lot of PhD.s, and by and large they have little clue about how to provide value. They want to sketch out ideas in Matlab, and throw it over the wall to a bunch of programmers (which in the academic world are postdocs and grad students) to deal with the "implementation details". Many can't write a simple for loop.

That's sweeping, overstated, and unfair of me, but the job skills are quite different. Applied, directed research aimed at producing tangible and sell-able results is very hard, and by-and-large not what they do, at least based on who applies and interviews (which is of course a terribly biased and perhaps non-representative selection).