As the sibling comment says, the 999 mechanical engineers will not train anybody. So it is still: one person trained replaces one person working. People with P

As the sibling comment says, the 999 mechanical engineers will not train anybody. So it is still: one person trained replaces one person working. People with PhDs can already work on more than just scientific positions, so it is not a huge problem that we are producing too many. As you say, we should be more honest about career changes. I don't think we need hugely many professorships either. But we should open up the hiring pool. If professors do more and more admin work anyway as their career progresses, why do you need 15+ years experience and personal connections to do that job? In industry, you have the other extreme: MBAs with no knowledge about the subject matter get hired as managers. And if most of the labor is done by master and PhD students, why not open up the scientific hiring pool to people from other specializations or from outside academia? I would say my postdoc and research time was maximally 10% thinking about physics, 30% coding, 30% writing papers, 30% data analysis (and 20 bonus % meetings and people skill). You could easily train a good developer or engineer in a year to do this, about the same time I needed to get up to speed with the topic. So yeah, there is a shortage of positions. But to fix that, you need a lot more funding (and honestly I'm not sure it makes sense if society allocates too much more resources to research. A bit, certainly, but I'd allocate more to health care, primary education, etc. first.) But the second problem is the baroque way to get a career in science. To fix that, you'd need to start hiring like a tech company. And this could actually make the competition worse, but you would be selecting for the best, and not for the people who make the most personal concessions for their career.