What about academic operating system research that was done years ago? Exokernel, SPIN, all aim to solve the "os is the problem" issue. Why don't we see more in that direction?