Several years ago it was shown that dogs are less prone to the sunk cost fallacy than humans (couldn't find a non-gated version; pop science article here.)
The humiliation continues: pigeons are better than humans at the Monty Hall Dilemma. The question is whether humans lack whatever circuit is allowing pigeons to inductively (one assumes) approximate the logic of the Dilemma, or whether there's something in human cognition interfering with otherwise at-least-equal optimizing ability in this task.
A reasonable way of thinking about the first option is that for this specific behavior, for whatever reason, pigeons are more optimized, though for most things they aren't very good generalists.
Next step: finding out whether hamsters one-box or two-box on Newcomb's problem.