Maybe it’s obvious, but there are lots of situations where believing an incorrect fact is less harmful than learning the incorrect fact too slowly. “Real” bayesians who are maximising EV won’t have this issue, as they’ll be happy to make decisions on the basis of things they think are unlikely to be true.
A more realistic choice is between being the kind of person who doesn’t respond quickly to changes in the world, or being someone who might respond quickly enough but is sometimes (maybe even often) wrong.
Maybe it’s obvious, but there are lots of situations where believing an incorrect fact is less harmful than learning the incorrect fact too slowly. “Real” bayesians who are maximising EV won’t have this issue, as they’ll be happy to make decisions on the basis of things they think are unlikely to be true.
A more realistic choice is between being the kind of person who doesn’t respond quickly to changes in the world, or being someone who might respond quickly enough but is sometimes (maybe even often) wrong.
Not obvious at all, and useful/helpful to share and remind us all of. 🙂