alt Hacker News> We want to train highly skilled people
After how many postdocs is one "highly skilled"?
Postdocs are foremost a failure of the academic job market (supply of eager researchers much larger than available academic positions) that over time has become some sort of ingrained filtering mechanism.
There are different kinds of postdoc positions.
Fellowships are generally the good ones. You get to explore research directions under the guidance of a more experienced researcher. Because you have independent funding, your mentor is not your boss, and you are relatively free to choose what to work on and how.
On the other end of the spectrum, some postdocs are just people a PI has hired to work on a well-defined project. Either because the institute does not have more appropriate job titles or because postdocs are cheaper than professional researchers. These are fundamentally ordinary jobs, and like any job, they can be good or bad.