The description of differential privacy in the article is completely wrong. I am really shocked it was described so poorly.