If ((Assume it's stupidity) == (discount/ignore the risk)), then assuming it's stupidity is never the safer assumption, even if it's empirically more likely to be the correct assumption, no?

All boils down to an individual's threat model at the end of the day anyway, though.