In most common, trivial real life situations, you have an advantage if you keep spewing forth nonsense and adjust for mistakes as you go along and keep pushing forward semi-blindly - compared to being forever hesitant in search for perfect truth.

The "confident idiot" basically applies a Monte Carlo search over the range of possibilities, and may eventually fall onto a good-enough solution.

I suspect that's the explanation for the D-K effect. If that's the case, then the "idiot" actually uses a pretty sophisticated overall strategy that works well in a fluid world (so not so much an idiot actually).