I would argue this is already essentially the dominant model. There are already more plausible projects than there is money to fund all of them, and the credibility of the person doing the proposing appears to be pretty decisive in terms of whether you actually get funding or not. It also isn't that uncommon for people to be a little dishonest about the funding and play games to shuffle it around to accomplish different ends. It would be nice to eliminate that layer, but I don't think it would be a revolutionary advance in quality or anything.