Hacker News Reader

Prohibiting the sale of individually identifiable information is a very weak protection. It's alarming how quickly multiple dimensions of "anonymized" data can be used to zero in on any target person, and it won't take long for amoral opportunists to start selling de-anonymized data to the highest bidder. See also k-anonymity.

View on HN

Combine the data with publicly available information about politicians and other public figures, feed it to a machine learning algorithm and I bet you could get very high accuracy results very, very quickly.

Look no further than the news in Congress last week. Devin Nunes' claim is that he saw conversations where names were masked, but it was obvious who they were about.

When doing this procedurally, it's amazing how few data points are needed to pinpoint a specific person.

Reply on news.ycombinator.com