I could swear that SH discussed black holes "evaporating" via quantum tunneling several years ago. Am I right, or just hopped up on the beeber?
That's correct - since the exact position of a small-enough particle such as an electron is not really "exact," but more of a bell-curve shaped smear, it can happen that one part of that position "smear" falls outside of the event horizon of the black hole, allowing the electron to escape.
The smaller the black hole, the higher the rate of evaporation.
You're right, but the question is whether the information contained in the black hole could in principle be "read" from the outgoing particles. He used to say no, now he apparently says yes.
For example, suppose I bring together a bunch of neutrons to form a black hole in one place, and an equivalent mass of Lambda hyperons to form a black hole in another. (Lambda hyperons are just like neutrons, except a down quark is replaced by a strange quark.) Would I find that the radiation from one BH has a net strangeness, whereas the other does not? The conventional belief was that I wouldn't.