This is field-dependent. In some fields, undergrads can already be independent researchers. In other fields, even postdocs are not independent, as doing research requires funding, materials, and labor. The latter is the real problem, as professors in those fields need a steady supply of PhD candidates to do their research.
You finish your college, you're officially an (eg.) engineer now, you can build a bridge, millions of people will use every year, and it's ok... but if you want to try out new compositions and additives of/to cement in a lab environment and measure their properties, you need a mentor, and six different unrelated courses... and preferably a well-known name to co-author your article, to get it published at all.
As an engineer just out of the school, you wont be the one responsible for building a bridge millions of people will use every year. It takes more years to get there.
Furthermore designing a bridge using a century old design, well understood materials, and sticking to all the well known, well documented and legally mandated best practices we have accumulated over the years is an entirely different skill than trying to come up with a novel and better way of designing a bridge and then conclusively proving that it is still at least as safe as the old way of designing a bridge.
But actually I mostly agree with you. PhD should be focused on apprenticeship and nothing else.
Here is the thing: The purpose of a PhD is to demonstrate the ability to conduct scientific research independently. When a student enters a PhD program, they should have all the prerequisites to do that. If not, the university has failed and the study program needs to be reformed. A PhD student needs to have the ability to conduct research and write a thesis, and they get more time for that than they will ever have again as a postdoc.
In my opinion it's really that simple: Do the research and write your PhD thesis, and finish it in time. This "training" idea comes from the US, and the only effect of this in Europe I've seen so far is that unnecessary and off-topic PhD courses with point systems have been invented. At least in my area, these are completely useless and merely delay work on the PhD thesis.
What they should do, however, is to stop the grade inflation and stop handing out a PhD to everyone, no matter how bad the thesis. That's the real problem, I've seen plenty of lazy, extremely low-effort theses, and every time the lazy PhD student passes the defense and gets the PhD. That devalues the PhD and the work of those who take it seriously. Please start to let at least the worst PhD students fail.