As an example, Englehardt cited how more than 30m Americans have diabetes. "While we’re very unlikely to re-identify any of those users based solely on the knowledge that they have diabetes, I suspect we can agree that nearly every individual in this group would not want this information used by advertisers," he said.
It's even worse than that. A person is unlikely to be in just one of theses groups.
Maybe there's a group for how affluent one is? e.g. Earns more than 100K.
And a more specific, but still generalised location, e.g. State, Rhode Island.
And several interest groups, Gaming, Adventure Sports, Sports Cars, Motorbikes, Romance Novels, SciFi/Fantasy Movies, Sports (maybe even specific sports, Football, Grid Iron, Tennis, Cricket, Baseball, Hurley, Rounders, Cross-country skiing, Water ski-ing, Snooker, Ice Hockey, etc.).
So, how many people would match the set of:
Diabetic, Rhode Island, Affluent, Romance Novels, Rounders, Cricket ? Not many I'm willing to bet.
And the more groups there are, the matching becomes even narrower, say you add age brackets and gender and gender identity to those sets. Along with information like browser agent strings, pretty soon you are going to be able to track specific individuals anyway.