I think the fact that the Mind Projection Fallacy is a really strong bias in humans significantly decreases the weight of that possibility. Smart people think it may be true because that sounds like the easiest explanation, for a human, not because they actually thought a lot about it from a strictly rational point-of-view.

That's some kind of general counter-argument against "trust the majority", I think. When you learn that the majority has some kind of bias that supports its belief, you should decrease the strength you assign to the evidence "the majority thinks it's true". P(A|B)/P(A|!B) is small.