Author seems to assume the latter condition is definitive, i.e. that real machines are not, and then derive extrapolations from that unproven assumption.