The problem with PARC and AT&T labs wasn't that having a bunch of really smart people doing interesting stuff failed to produce brilliant, world changing ideas, it was that the companies in question weren't always good at capitalising on them (arguably for the better, in the case of Unix).
AT&T Labs wasn't too bad at capitalizing on their inventions, though they didn't do equally well on all of them. For a pretty long time the lab was definitely in the black, though. Their most profitable invention, which came about halfway through the Labs' existence, was probably the transistor, which made AT&T a ton of money. I don't remember where I read it, but some old Bell Labs people calculated that it made AT&T so much money that that single invention paid for something like 40 years of the Labs' operating costs.
The problem with PARC and AT&T labs wasn't that having a bunch of really smart people doing interesting stuff failed to produce brilliant, world changing ideas, it was that the companies in question weren't always good at capitalising on them (arguably for the better, in the case of Unix).