I don’t think saying PhD is good or bad in general makes sense. It depends on what alternatives you have on your hands. If you are from a third world country and PhD is the only chance to take you a batter place, it makes sense. For the ones who already live in a decent country, it only makes sense if you are very very passionate about being an academician and you should carefully question the arguments making you passionate about that. I think doing PhD is much tougher now than 1983. There is a huge literature you gotta scan and there will be so many other people from both academia and industry working on similar topics so it is very difficult to make an original contribution that is satisfactory enough for the most. Especially in engineering, hot topics having an application area are already studied by the high tech firms. One can work in one of those companies and still can do research, which can be more satisfactory since you can see it applied to the real life. Plus, you will have strict deadlines, much better salary and the time you spend there will be considered as a work experience when you make a new job application. So, in the end of the PhD, there is a risk of finding yourself where you started and asking yourself where you did wrong, especially if you didn’t consider the after-PhD phase during the PhD (wrong topic and wrong supervisor choice, not building connections, not doing an intern in a related company etc.) However, having a supportive supervisor with good connections can hack the game and put you some levels up.
I have the sense that a PhD in computer science is harder to obtain now because the field is more mature, there are many more people working in it, and it just takes more time and effort to get to the point that you're doing original work.
Also, my impression is that, in the USA at least, the economy is fundamentally much more fragile, with the disappearance of the middle class. Someone starting out now does not really have the luxury of doing something purely out of passion. The penalty for needing to change course is much higher now. I think.
I disagree with you on the role of industry in computer science research. In the 80s, there were many companies in which research was happening: Xerox PARC, IBM, Digital, and to a lesser extent at other computer companies. I worked at Computer Corporation of America, a mid-sized software company, in the mid 80s (after UMass, before startups), and they did some amazing, DARPA-funded research in databases. Now it's Microsoft and IBM. Oracle, Apple and others are obviously doing leading-edge work, but I think those efforts are more closely aligned with products. I think there was far more pure research done by software/compute companies 30 years ago.
