I disagree with the article's premises and conclusions, my perspective is that of a physicist who also knows many lapsed physicists.
The premise seems wrong in that he disagrees with the elite overproduction thesis, but the professions he lists are " lawyers, MBA holders,..." and later to finance. But these are the professions likely to be zero-sum. So to me the article really is "I agree with the elite overproduction thesis, I just don't think it applies to physics".
OK lets talk about that then.
If there is something that's gone wrong with pushing forward the boundaries of our knowledge it isn't a lack of bodies producing the standard unit of scientific knowledge: "the peer-reviewed publication". These have grown exponentially. I don't think I've seen anyone even argue the rate of knowledge increase is growing commensurately. It *might* be the case that new knowledge requires exponentially more people looking for it but that's a pretty different article. Personally, I'm skeptical more NSF funding would help.
I do agree with this part:
Our apparatus of scientific education, especially mathematics education, is deeply flawed. It takes too long to bring young people to the research front, and it produces a false view of what doing science is really like.
But this research is in a academic setting, and there
is definitely, unambiguously, an overproduction of physics PhDs relative to academic positions available. Letting more new students in the door wouldn't help anyone!