Note: I don't have a PhD, this is just based on the article and other things I have read.
It seems to me that this article goes around the two main, most obvious aspects of a PhD;
1) Improve yourself (smarter, better critical thinking, etc)
2) Appearing more valuable to others (Getting grants, getting hired)
So the question is, does it fulfill these goals, and if it doesn't, do prospective applicants realize it?
Certainly there are better uses for your time if you want to appear more valuable to others (work experience), and there might be better ways to improve yourself for the same time and money.
If the average applicant realizes this, but proceeds anyway, then it's not a bubble, as there must be other factors at play when they think about the value of their degree. If they don't, then there is a bubble that will burst when the information becomes more plain.
In my experience, the PhD does makes you smarter in a sort of trial by fire. I haven't met a single (math/CS) PhD student who wasn't intimidated by the magnitude of their research field when they started, or seriously intimidated by the math in a seminal research paper in their area, but came out with a mastery of the subject (which is not to say that there aren't any crap PhDs -- there are). At the very, very least, if you finish your PhD diligently, or even get close to finishing it, you will develop a confidence that you can in fact chip away at the most abstract and difficult problems, and usually come up with a good solution, while building on the efforts of much smarter people. At a very practical CS/algorithmic level, a lot more will seem possible, and you'll reinvent the wheel a lot less.
I've often seen very smart people in CS/ML who didn't have a PhD. However, they sometimes also lack the body of knowledge in advanced algorithms or statistics that comes from 5-6 years of monk-like reading and chipping away at immense intellectual problems that have been tackled by some of the smartest people in the world.
So if you don't mind 5-6 years of near-poverty, intensely difficult problems, a possible lack of interest from your advisor (thankfully not in my case), peer-review rejections that sometimes seem arbitrary and disheartening, and can take 3 months to come around each time, large minds and frequently larger egos, intense orgasmic breakthrough eureka moments that come from spending months or years working on a problem, self-learning some truly amazing things, and drinking discount "St. Remy" brandy instead of Remy Martin, then by all means go in for a PhD -- you'll come out sharper if you make it.
I think this point is often missed. I always compare a PhD to professional athletics for the mind. You need to have the same kind of dedication, constant training, ability to deal with failure, good coaching, spartan lifestyle, etc. I never even realized you could exhaust or even cause injuries to your brains by mere thought before I did a PhD. The final goal is to train your mind to compete at a global level, hopefully doing something useful for humanity in the process. You'll come out being able to pierce through complex systems, where others are merely applying tricks.
The specialization of CS related PHd's were truly more hurtful then helpful (on the whole, not as a 100% "rule") when I recruited $120-$150k programmers (mostly finance / eCommerce giants). From the VP's (paraphrased) "They tend to be too narrowly focused on their field of study, and many times have a hard time adapting to our ___[area of need, actual job requirements and day-to-day, etc.]__" This was more on the actual PHd candidate themselves however. Equate this too a Lib Arts Degree, where, it depends entirely on the graduate to get the position they are going for versus relying on the degree for a job. A few different discussion on HN related that a PHd is for RESEARCH, and, if you are expecting a $20,000 bump in pay, that's a lot of opportunity cost with little actualization of pay off.
I had actually never looked (naively) at a PHd focus solely for Research, versus getting a "better" position within an industry.
The specialization of CS related PHd's were truly more hurtful then helpful (on the whole, not as a 100% "rule") when I recruited $120-$150k programmers (mostly finance / eCommerce giants). From the VP's (paraphrased) "They tend to be too narrowly focused on their field of study, and many times have a hard time adapting to our ___
Sounds like they were looking for programmers to do the kind of boring work that CS PhDs would get sick of really quick. Other companies would love to have exactly the same CS PhDs you were rejecting.
It seems to me that this article goes around the two main, most obvious aspects of a PhD;
1) Improve yourself (smarter, better critical thinking, etc) 2) Appearing more valuable to others (Getting grants, getting hired)
So the question is, does it fulfill these goals, and if it doesn't, do prospective applicants realize it?
Certainly there are better uses for your time if you want to appear more valuable to others (work experience), and there might be better ways to improve yourself for the same time and money.
If the average applicant realizes this, but proceeds anyway, then it's not a bubble, as there must be other factors at play when they think about the value of their degree. If they don't, then there is a bubble that will burst when the information becomes more plain.