Mean-variance optimization is about to begin its 60th year and by all accounts it has aged extremely well. Originally proposed by Markowitz [1952], it was not until the mid-1970s that institutional investors seriously embraced this technology to structure portfolios. The impetus was twofold. Congress enacted ERISA, which for the first time imposed fiduciary liability on the stewards of pension assets. And from the beginning of 1972 through the end of 1974, the U.S. stock market lost more than 35% in real terms. Institutional investors were desperate to find better ways to manage risk and to avoid the new legal consequences of failing to do so. Since then, it has survived and prospered because it works. As with many innovations, however, practitioners of the old technology resisted change and defended their resistance with a variety of excuses, which persist even today. I would like to comment on three criticisms of mean-variance optimization.