> but if that is a problem people should expect to go build a new system
Imagine you're walking on the street and see a helicopter crashed on a tree. You don't need to be a helicopter mechanic, or aeronautical engineer, or a pilot, to know that the helicopter shouldn't be on that tree. And you definitely don't need to propose new helicopter control methods to apply common sense and conclude that it's a bad thing that the helicopter crashed on the tree and we should try to not do that again in the future.
Sure, but imagine you go to a university and see a pretty standard mentor-apprentice relationship designed to perpetuate a culture. We don't need to be psychologists to apply common sense and say it is done that way for a good reason and alternative approaches should be trialled in a different system.
It would be good to have a better system, but we can develop it without risking the valuable academic culture. You can see how valuable it is because people keep signing up to it despite the archaic nature of its traditions.
Imagine you're walking on the street and see a helicopter crashed on a tree. You don't need to be a helicopter mechanic, or aeronautical engineer, or a pilot, to know that the helicopter shouldn't be on that tree. And you definitely don't need to propose new helicopter control methods to apply common sense and conclude that it's a bad thing that the helicopter crashed on the tree and we should try to not do that again in the future.