This is a great talk because it touches on so many questions such as: how much justification can we have for a priori theorems? What does a proof need to be convincing? Is conceptual analysis the appropriate way to responding to ambiguities in ordinary language concepts? Why did people question Turing's proof in 1936 for more than 30 years despite being considered ironclad today?
"The universe itself has an unlimited supply of generosity and surprise built in, and as designers we only need to keep our eyes open to what is here."