The final confidence of the "diagnosis" is about 50% (you can use the Bayes formula to get that, see here http://www.visualab.org/index.php/cyberchondria-microsoft-ba...). Yes, 50% is better than nothing, but the NY Times article does exactly what serious newspapers should not do –let people think that "the Internet" is a great place to diagnose yourself. It is not.