The expectation that wrong ideas won't be too harmful is something that turns out wrong again and again. (I am not claiming anyone in particular had the thought Kurweil's ideas were wrong and nearly harmless.) The outside view reference class of any mistaken notion one thinks isn't very harmful (because one can't think of how it could be) is that one is wrong.

I would be very interested if someone could explain how to prove the harmlessness of an idea.