This has kinda been tried and hasn't worked. The incentives to the journal system (i.e. promotion and tenure) are too strong.

"Fast, cheap, and high-quality" you can have 2 out of 3 is pretty much sums up the problem here.