If humans are too irrational to directly follow utilitarianism (U), then U implies they should come up with easier/less dangerous rules of thumb that will, on average, produce the most utility.

This is a pretty dubious move. Why think there will be easy to follow rules that will maximize aggregate utility? And even if such rules exist, how would we go about discovering them, given that the reason we need them in the first place is due to our inability to fully predict the consequences of our actions and their attached utilities?

Do you just mean that we should pick easy to follow rules that tend to produce more utility than other sets of easy to follow rules (as far as we can figure out), but not necessarily ones that maximize utility relative to all possible patterns of behavior? In that case, I don't see why your utilitarianism isn't collectively self-defeating according to the definition you gave. A world in which everyone acts according to such rules will not be a world that is as close to the utilitarian Paradise as empirically possible. After all, it seems entirely empirically possible for people to accurately recognize particular situations where actions contrary to the rules produce higher utility.