The problem is that excessive competition in a system originally designed to ensure research money was spent well, is now actually causing the money to be spent badly. Asking for concrete suggestions for improvement is perfectly fair. Here are two things that could be done that would help.
First, ideally, flood the system with money, fund science the way it should be funded, as though it actually mattered as much as entertainment or killing people. Again, providing ten times as much money would produce more and better science per dollar not just in total.
Second, failing that, filter PhD applicants by random lottery; discard a randomly selected 80% or however many it takes, before moving forward as currently.
First, ideally, flood the system with money, fund science the way it should be funded, as though it actually mattered as much as entertainment or killing people. Again, providing ten times as much money would produce more and better science per dollar not just in total.
Second, failing that, filter PhD applicants by random lottery; discard a randomly selected 80% or however many it takes, before moving forward as currently.