This is why sincere stupidity is actually worse than insincere stupidity: the sincere tend to insist on their folly.

e.g. in this dialogue form (which I see way too much of on LessWrong):

A: X's action Y was stupid, and X should have known this because of Z.

B: But X's action was entirely justifiable according to V and W!

B's statement is in the place in a discussion where a refutation would go, but doesn't actually address the folly; and seems to claim that sincerity makes stupidity less bad. Whereas in practice, sincere stupidity promises more stupidity in the future.

(A's statement is an assertion about the processes leading X to commit Y, rather than merely the folly of Y; however, A is asserting that bad results that could have been reasonably predicted should have been. The discussion can then go into a long thread about the meaning of "reasonable", possibly with one of A or B subtly dissing the other's Bayesian-fu.)