Gen. Keith Alexander, director of the National Security Agency and the chief architect of the contemporary surveillance state, announced this week that he will soon retire. His departure was long expected, and the announcement surprised no one. Still, some are using this opportunity speculate on how President Obama could reform our surveillance apparatus after Alexander’s departure. Instead, I’d like to take just a few paragraphs to look back on what he achieved.

I don’t think Alexander is an evil genius. I admire him for his technical expertise, which is beyond anything I could ever hope to acquire, and his commitment to public service. Yet simply having good intentions is never enough. Alexander has consistently ignored the potential consequences of his actions. He has, apparently through mere willpower and charm, built a technological and legal apparatus with the ability to collect, analyze, and store detailed information on the relationships among hundreds of millions of people, if not billions. It’s exactly what Hannah Arendt wrote was the ultimate goal of the Soviet secret police. It’s the closest thing to omniscience ever built by human hands. Meanwhile, Alexander has turned cryptography into a truly offensive weapon for the first time. It doesn’t really matter if he did it all for the right reasons. His legacy will be determined by how these tools are used in the future.

For many, I think what has been most surprising about U.S. surveillance under Alexander has been its emphasis on the connections between people rather than individual information. The past month was a particularly clear illustration of this. Alexander recently revealed that his agency had tested a program to gather data on the locations of cell phones, but that the program had been terminated because it provided, in his words, no “operational value.” Yet last week, we learned that the agency is collecting hundreds of thousands of address books and contact lists every day, even though it doesn’t really have the capacity for so much data and much of the information is basically noise. This program, apparently, does provide operational value.

It’s possible, of course, that Alexander is simply lying by omission about whether the agency has records containing location data, as Sen. Ron Wyden’s statements have suggested (D-Ore.). Yet it is also easy to see why the value of location data might be limited compared to that of a global and intricate matrix of relationships between people. Even without knowing how frequently a person communicates with someone in their address book, a sufficiently large amount of this data would allow the agency to determine who the most important people are in any given social group by calculating how many different connections each person is capable of making between people who aren’t otherwise associated with each other. Kieran Healy offers a very entertaining demonstration. Locations are only useful if the agents already know who they are trying to find.

Note that both locations and address books could be described as metadata, depending on your definition. Locations could be classified as metadata under Section 215, but apparently are not, and the address books have not been classified under that statute or any other because they are not collected on U.S. soil. In any case, they are very different kinds of metadata. One is about specific people, the other is about the connections between them.

Obviously, a surveillance program whose primary purpose is the collection of the second kind of metadata creates serious problems. Does the state’s legitimate interest in learning more about one person justify gathering information about people associated with him? What about the people connected to them? These are the questions which Alexander did not anticipate or couldn’t be bothered to ask.

Our ideas can save democracy... But we need your help! Donate Now!

Max Ehrenfreund is a former Monthly intern and a reporter at The Washington Post. Find him on Twitter: @MaxEhrenfreund