I never quite understood

alpha, the probability of the type one error is the probability of rejecting the null assuming that it's trueType II error (rejecting the null, which is that there is at least 64oz)

the type two error is the opposite assuming some value under the alternative hypothesis.