I just did the same experiment. The thing claims I'm male, 3 times out of 3. I also read the article, and either the "gender genie" doesn't really follow the principles in the article, or it was not designed to work on blog entries. Perhaps the writing style of the blogging population is different from the research population (I have some ideas about that).
Bah!
Hey, have a good day.