Uncritical funding of projects submitted by superstars seems to be one significant problem in funding that overly reinforces the status quo (like other rich get

Uncritical funding of projects submitted by superstars seems to be one significant problem in funding that overly reinforces the status quo (like other rich get richer phenomena). In some fields, only a small fraction of applications (8-15%) are funded and much of it goes to "superstars". I would propose that the bottom 50% of applications be removed by review and the rest be funded at random. In order to reduce spam applications, it would be necessary to adjust the probability of funding by the total amount of money being applied for the previous 24 month by a given PI. To encourage cross-disciplinary research, I would award the top 1% of projects non-randomly but only if the PI has not been previously funded by the granting agency.