This, but to some degree, unironically. If studies aren’t reproducible (or deemed worthy of reproduction) then there’s definitely a disconnect between the folks handing out research assignments and the folks engineering applicable solutions to scientific problems.
That goes two ways. You could be a guy who successfully formulates a mathematical model to support the existence of Neutrinos and face a funding board that has no interest in building a LHC. That’s arguably a problem of malinvestment within the scientific community. Or you could be a guy who successfully formulates a mathematical model for a new kind of mouse trap that’s 10% less efficient than traditional mouse traps. That’s more of a university research assignment problem. Or you could have a researcher who claims he’s the only one who can do a particular thing, because he’s got the magic touch. If the research is unfalsifiable by design, that’s an entirely new kind of problem.