Pay researchers for results, not plans

Quite a good idea so long as the balance between quality & quantity was carefully considered. A focus on quality e.g. Impact Factor & citation numbers after a couple of years might also reign in the predatory publishers and lead to fewer but better papers

The problem with an upfront investment by venture capitalists is agenda capture: the risk of more and more pet projects. Venture capitalists will also want want a return on their investment and so researchers will switch from pleasing a panel of reviewers to pleasing a panel of investors (imagine a dragon's den scenario). Plus... a proposal will still be needed, it will just be called a "pitch". Not sure how this solves the problems this article rightly identifies.

It seems logical that research could be improved—but not simple. We need a more thorough analysis of the problems and obstacles, and due diligence on proposed solutions and their intended and unintended consequences—for different stakeholders. I spent half my career in academia and half in industry—research and innovation. Basic and applied research are quite different beasts, so no easy solution for either. Venture capital and U.S. government (SBIR) funding are both designed to encourage research and reward value creating scientific knowledge, technology and innovation in products, services and other assets.

it's clear that the grant request system in the old fashion is broken and must be found a new solution to this challenge, this proposal it seems a very good idea but it's necessary to enter into more details, how it may work well in all aspect of the science, it's not clear. But this may be a good base to start the change.