I think the core problem is our ignorance of psychology - we don't know how humans really work, what makes us tick, what are the 'incentives' that should be put

I think the core problem is our ignorance of psychology - we don't know how humans really work, what makes us tick, what are the 'incentives' that should be put in system design to move toward better science, and whether 'incentives' are even a good conceptualization of human motivation. We will not fix science until we understand what makes scientists behave as they do, and until we figure out how to design systems for humans. Though, Maybe we stumble upon a better system via blind variation in system properties and selective retention, based on some novel metric. Scientific psychology is rather weak in explaining and predicting how humans will behave.