Former NSA whistleblower Bill Binney
has been making the interview rounds in the wake of the NSA surveillance leaks, adding some useful context and history
to some of the allegations. In one interview, with the Daily Caller, almost as an aside, he makes a really good point about how in the desire to collect all these haystacks of information, the NSA is missing out on a lot of needles
. There's a discussion of how law enforcement apparently has a recording of a phone call between one of the Boston Marathon bombers, Tamerlan Tsarnaev and his wife, but that nothing was done with it. Binney points out that, when you have that much data, important stuff gets lost.
They're making themselves dysfunctional by collecting all of this data. They've got so much collection capability but they can't do everything. They're probably getting something on the order of 80 percent of what goes up on the network. So they're going into the telecoms who have recorded all of the material that has gone across the network. And the telecoms keep a record of it for I think about a year. They're asking the telecoms for all the data so they can fill in the gaps. So between the two sources of what they've collected, they get the whole picture.
They can do textual processing at a rate of about 10 gigabits a second. What that means is about a million and a quarter 1,000-character emails a second. They've got something like 10 to 20 sites for this around the United States. So you can really see why they need to build something like Utah to store all of this stuff. But the basic problem is they can't figure out what they have, so they store it all in the hope that down the road they might figure something out and they can go back and figure out what's happening and what people did. It's retroactive analysis.
Yes, it may be useful
in hindsight (not that useful automatically makes it legal), but what would be even more useful is if they stopped focusing so much on collecting data, and went back to doing traditional investigative work that focuses on real targets. Piling on to the haystack doesn't help anyone. It comes from the faulty belief that piling more data on top -- even if it's useless data -- must be a good thing.