The "event" that Geoff Hinton concludes "X" should, rationally, update your posterior probability of X, unless you put zero weight on that evidence.

→ View original post on X — @erikbryn