The same effects have also made long postdocs less common, and it's even possible to jump directly to a tenure-track faculty job from a PhD with no postdoc in between. Alternately, people seem to be doing a kind of "industry postdoc". You get an industry job, but one that lets you publish (DeepMind, Facebook Research, some startups, etc.), and after a year or two you either find you like it and stay, or you use that as a springboard to apply for faculty jobs, treating it as having been basically a high-paid postdoc.
Still plenty of stuff to complain about, but the job market looks far different than in the natural sciences or humanities.
Yup. Economics [0] and comp sci are basically the only doctoral programs worth going into at this point (at least if you're an American citizen). You can debate which one is _more_ worth it, but the fact is that if you're doing anything else, you're in a for a world of pain.
[0] http://noahpinionblog.blogspot.com/2013/05/if-you-get-phd-ge...
[0] https://twitter.com/Noahpinion/status/1015988356184866817
This isn't new. I had plenty of professors that worked at e.g. Bell Labs before returning to academia.
His English sucked, his physics PhD in non linear quantum mechanics useless and his resume-writting skills was non existent.
Yet, he's the only person to read my thesis, the cluster worked only thanks to his efforts, and he administered 20~30 workstations while doing his research. I would not have graduated without him and he is more dear to me than anyone I met at school.