A Littlewood student, Stanley Skewes, estimated that only when counting the first few numbers no smaller than 10^10^10^34 can one witness the underestimation of the number of primes by Gaussian integral logarithm. This is an incredibly large number, probably the largest number ever considered in a mathematical proof up to that time.
