The real issue with scientific publishing is that there is simply no penalty for publishing shoddy research. I know several academics who made quite a big name for themselves on research that was later partially or fully retracted. No one cared about that; there was no real reputational damage done. To tackle poor science, such "poor" scientific inquiry should be "punished" in some way. Similarly, it is terrible for the advancement of science that only novel or significant results get published -- there should be a way for researchers to benefit from publishing well-designed research which simply did not yield interesting results.
How to do that? I think some of Axelrod's tournament provides an answer. Like in his examples, the individual incentives align to yield a pretty poor outcome to the members of his population (he runs an iterated prisoner's dilemma game). However, correctly setting up the iteration parameter's, slowly a cooperative strategy becomes the evolutionarily stable strategy.
I can see how this could also be the case for academics. There is no law from up above that dictates that "number of papers published" is the ultimate metric of success. There is a culture, and processes, and institutions which have led that to be a leading indicator of academic success. If there were real motivation and impetus to change this, there is no reason to imagine that other metrics (and processes) could emerge that would much more highly value scientific integrity and thoroughness.