Given the power we see in 100+ study meta-analysis in medicine, I think we have to move towards a model where research grants in other fields (psych, sociology, etc.) are jointly awarded to multiple independent groups to jumpstart the meta-analysis for a given new topic from the outset. How to keep the "independent" groups from colluding will be a challenge unfortunately. As will funding for the extra replications. For the reasons you cite, single, stand-alone studies are almost useless these days.
