Actually, I think “This Turing machine halts” is the more obviously troublesome one. It gives no computable expectation of when we might observe a halt. (Any computable probability distribution we would assign to halting time would put too much mass at small times as compared to the true distribution of halting times.)
Actually, I think “This Turing machine halts” is the more obviously troublesome one. It gives no computable expectation of when we might observe a halt. (Any computable probability distribution we would assign to halting time would put too much mass at small times as compared to the true distribution of halting times.)