In software we (or at least I) often see situations where careful forethought put systems and checks in place that survive long after the responsible parties are gone, and it can be difficult to identify or accept that the project failed this year because Tom quit last year. There are too many forces arrayed against looking too closely at those situations, and once the wheels come off there's no longer a quorum to even rationally discuss such a thing, except amongst a few veterans having a beer and trying not to draw too much attention to themselves.
If you idiot-proof something well enough, it can take a long time for the Universe to invent an idiot good enough to pull the whole thing down. We tend to look at Rome as one of the oldest well-documented cases of the power of logistics, so it's not outside of the realm of reason to suggest that a good process might survive even a generation or two before Chesterton's Fence wins out. But whether that's due to a loss of cleverness or just human nature is likely going to be hard to prove.
If you idiot-proof something well enough, it can take a long time for the Universe to invent an idiot good enough to pull the whole thing down. We tend to look at Rome as one of the oldest well-documented cases of the power of logistics, so it's not outside of the realm of reason to suggest that a good process might survive even a generation or two before Chesterton's Fence wins out. But whether that's due to a loss of cleverness or just human nature is likely going to be hard to prove.