Suppose that one can simulate as many i.i.d. Bernoulli random variarbles with parameter p as one wishes. Explain how to use these to approximate the mean of the geometric distribution with parameter p.