Home > data science, math, news, rant > NSA mathematicians

NSA mathematicians

August 25, 2012

When I was a promising young mathematician in college, I met someone from the NSA who tried to recruit me to work for the spooks in the summer. Actually, “met someone” is misleading- he located me after I had won a prize.

I didn’t know what to think, so I accepted his invitation to visit the institute, which was in La Jolla, in Southern California (I went to UC Berkeley so it wasn’t a big trip).

When I got to the building, since I didn’t have clearance, everybody had to stop working the whole time I was there. It wasn’t enough to clean their whiteboards, one of them explained, they had to wash them down with that whiteboard spray stuff, because if you look at a just-erased whiteboard in a certain way you can decipher what had been written on it.

I met a bunch of people, maybe 6 or 7. They all told me how nice it was to work there, how the weather was beautiful, how the math problems were interesting. It was strangely consistent, but who knows, perhaps also true.

One thing I’d already learned before coming is that there are many layers of work that happen before the math people in La Jolla are given problems to do. First, the actual problem is chosen, then the “math” of the problem is extracted from the problem, and third it’s cleansed so that nobody can tell what the original application is.

Knowing this (and I was never contradicted when I explained that process), I asked each of them the same question: how do you feel about the fact that you don’t know what problem you’re actually solving?

Out of the 6 or 7 people I met, everyone but one person responded along the lines, “I believe everything the United States Government does is good.” The last guy said, “yeah, that bothers me. I am honestly seriously considering leaving.”

Needless to say, I didn’t take the job. I wasn’t yet a major league skeptic, but I was skeptical enough to realize I could not survive in such an environment, with colleagues that oblivious. They also mentioned that I’d have to stop dating my Czech boyfriend and that I’d need to submit information about all my roommates for the past 10 years, which was uber creepy.

Nowadays I hear estimates that 600 mathematicians work at the NSA, and of course many more stream through during the summer when school’s not in session, both at La Jolla and Princeton. Somehow they don’t mind not knowing how their work actually gets used. I’m not sure how that’s possible but it clearly is.

This mindset came back to me, and not in a good way, when I read this opinion piece and watched this video in the New York Times a couple of days ago.

William Binney, a mathematician, was working on Soviet Union spying software that got converted to domestic spying after 9/11. In other words, they used his foreign spying algorithm on a new data source, namely American citizen’s raw data. He objected to that, so strongly that he’s come out against it publicly.

The big surprise is how come they let him know what they were actually up to. My guess is he was high enough up the chain that they thought he’d be okay with it – he’d been there 32 years, and I guess he was considered an insider.

In any case, watch the video: this is a courageous man. The FBI came into his house with guns drawn to intimidate him against his whistleblowing activities and yet he hasn’t been cowed. Indeed, after getting dressed (he was coming out of the shower when they exploded into his house), he explained to them the crimes of George Bush and Dick Cheney on his back porch.

As he explains, “the purpose is to monitor what people are doing”. He explains how people’s social media data and other kinds of data are linked over domains and over time to build profiles of Americans over time: “you have 10 years of their life that you can lay out in a timeline, that involves anybody in the country”.

Describing the dangers of this program, Binney was extremely articulate:

  1. “The danger here is that we could fall into a totalitarian state like East Germany”
  2. “We can’t have secret interpretations of laws and run them in secret and not tell anybody. We can’t make up kill lists and not tell anybody what the criterion is for being on the kill list”
  3. “Just because we call ourselves a democracy doesn’t mean we will stay that way.”

There you have it. The good news is that that guy is no longer helping the NSA do their thing.

But the bad news is, plenty of mathematicians still are. And if you want to find a community more trusting and loyal than mathematicians, I think you’d have to go to a kindergarten somewhere. Not to mention the fact that, as I described above, the problems are intentionally cleaned to look innocuous.

Another example, possibly the most important one of all, of mathematics being manipulated to potentially evil ends. We will have trouble proving actual evil consequences, of course, since there’s no transparency. The only update we will get is via the next whistleblower who can handle guns pointed at him as he leaves his shower.

Categories: data science, math, news, rant
  1. Tom
    August 25, 2012 at 9:24 am

    Engineers and scientists have long been faced with the prospect that their discoveries or inventions might be used for evil. Technology is a double-edged sword, but does that mean decisions should be made to impede discovery or invention?

    Our society has benefited greatly from technology that was developed with a military motivation. Are the people who developed it to blame? Could they have stopped it?

    I believe technology, invention, and discovery should be unimpeded no matter the motivation behind funding or developing it. It’s up to society to learn to use it responsibly. I think it’s fair to criticize its inappropriate use, but I’m not on-board with criticizing the people who develop it or their choice of employer.


  2. Cathy test
    August 25, 2012 at 9:31 am

    Too convenient. It’s not like these guys are developing algorithms that the NSA *may* use. They are developing NSA algorithms that they are not allowed to talk about to anybody, ever. That’s not unimpeded! And how can our society learn to use this stuff responsibly when it’s secret?


  3. mathematrucker
    August 25, 2012 at 10:30 am

    As I mentioned a couple days ago these are really personal choices. I would estimate that these decisions are usually made at least “98% locally” without letting “global” stuff (beyond one’s immediate family’s subsistence considerations) overrule.

    But I also have to admit, it’s really easy for me to get up on my horse like Jimmy Kimmel does and spout off on subjects like this because I’ve been single all my life and never had anybody to support but myself. I have no idea if I’d have avoided the so-called “beltway bandits” (and the NSA) like the plague upon completion of a math M.S. just 150 miles away from DC during the Reagan era if I’d had children to support.


  4. Tom
    August 25, 2012 at 10:31 am

    The NSA never publishes any of their work? They didn’t participate in the analysis of cryptographic standards that are used globally for commerce? They’re only capable of evil? Clearly, that was not the opinion of Mr. Binney over his previous 32 years there. My larger point is that maybe your perspective on them is not an entirely fair assessment of the utility of their organization.


  5. August 25, 2012 at 3:56 pm

    When I was a long-haired PhD aerospace engineering candidate in the early 1970’s, my thesis adviser once expressed essentially Tom’s view on the need to encourage the choice to work on potentially destructive (even massively such) projects. Occasionally emerging from my research basement into campus tear gas, I was aghast, but he patiently explained his (indeed rather visionary) viewpoint, based on intimate knowledge of newly-formed DARPA projects.

    At that time, if my memory serves his narrative, there was only a single $5M project being secretly dedicated to diverting an Earth-fatal meteoroid, including the use of “nuclear devices”. If humankind were to never have developed such gawd-awful devices, he opined, such an option would not be available, and much of Earthling life could be rather severely set back, if not obliterated.

    Of course, one should also be allowed to object, even risk unjust imprisonment for whistleblowing, to projects that are against individual ethos. Whether short-term benefits to one’s particular environment and viewable future or long-term benefits to Earthlings is determinative is up to such ethos.


  6. August 26, 2012 at 10:15 am

    I first heard Cathy O’Neil on the Frontline documentary series, “Money, Power and Wall Street.” That is why I started to follow her blog.
    The point of that series was that the unregulated, secretive actions of the few can and did have a huge negative impact on the lives of individual citizens and the communities in which they live, not to mention the world at large.
    It also outlined the ways in which the perpetrators, in their headlong quest for money and power, took complete and calculated advantage of the ignorance of the public, the impotence of those charged with oversight of their activities and the unwillingness of our elected officials to challenge their actions in any meaningful way.
    These are some of the key conditions under which the NSA is operating today.
    We have entered a brave new world where the massive harvesting of our personal information is enabled by our obsessive use of the very devices through which we willingly reveal ourselves to each other and, at the same time, unwittingly lay bear our actions and predilections to huge business interests and the state.
    If we are left only with dependence on a few honest souls like Mr. Binney to alert us to the huge potential for abuse of this environment, then we may as well shred the Constitution and face up to our citizenship in something quite different from a democracy.


  7. K
    August 26, 2012 at 11:56 am

    “And if you want to find a community more trusting and loyal than mathematicians, I think you’d have to go to a kindergarten somewhere.”

    That made me laugh — totally true! But I think mathematicians who work in finance are also quite suspect, and they can’t plead ignorance. Also I find data collection by businesses much more creepy, as you describe it in this post https://mathbabe.org/2012/08/20/the-death-spiral-of-modeling-e-scores/ . The government coming after me based on some data collection sounds like a remote possibility, while a company coming after me and figuring out just how to rip me off to the maximum extent possible based on the kind of person I am — that’s very easy to picture.


  8. August 26, 2012 at 12:14 pm

    I love people who care. All of you are good writers. Are you caring ? Americans talk to much and kefetch. Please care more.


  9. August 26, 2012 at 12:16 pm

    Brian Young :
    If we are left only with dependence on a few honest souls like Mr. Binney to alert us to the huge potential for abuse of this environment, then we may as well shred the Constitution and face up to our citizenship in something quite different from a democracy.

    Uh, as in “Hunger Games”, the masses passively accepting vicious penance for the crimes of the corrupt few? I think not – we fight on, even suffer torture, prison and death (as did Michelle Obama’s ancestors, constructing the White House) until the promise of a truly representative republic is finally fulfilled for those who follow us.


  10. amy
    August 26, 2012 at 9:29 pm

    William Binney, i am just a little puzzled that he was not put on ice thus far and is allowed to talk about these matter in public. usually, if u have a whistleblower the nsa takes silenlty care of this. that’s wat i am being told. so, either he has pressurezed the nsa with other things or …..
    where did this guy go to school ? any ideas


    • August 27, 2012 at 7:06 am

      talk to much.write to much. keeps you busy and not giving.going to work now.


  11. June 19, 2013 at 12:56 pm

    Tom wrote:

    I think it’s fair to criticize its inappropriate use, but I’m not on-board with criticizing the people who develop it or their choice of employer.

    If math is being used in a bad way it’s important to criticize the people and organizations who do it. That’s the only way to stop it. “Abstract” discussions of the issue won’t help much.

    Of course, these criticisms should be made in a reasonably polite way.


    • Nancy Jamison
      June 19, 2013 at 6:09 pm

      call me sometime.


    • beewhy2012
      June 20, 2013 at 9:18 am

      Glad you made this comment. I’d almost forgotten Mathbabe’s post on this whole sordid question. Snowdon’s revelations show how easily more than abstract concerns have been forgotten due to lack of “solid” evidence. Binney and Thomas Drake both went through hell in what most of us consider a just cause and yet their clear and expert opinion wasn’t enough to start the engines of full and proper oversight. Our media encourage us to live in an ever-forgetful present and that serves the purposes of Big Brother and the Ministry of Truth just fine.


  12. John Mitchell
    July 10, 2013 at 6:40 pm

    The claims by Tom and others that those who work for the NSA have no ethical responsibility to consider the ramifications of their work in regard to the Constitutional rights of U.S. citizens are unsupportable. We all have a universal duty to consider the ethical consequences of how we live out lives. Claims to the contrary are merely excuses used by those who prefer to ignore moral issues for the sake of their personal convenience, affluence, and security.


  13. Bobito
    September 7, 2013 at 8:09 am

    Those who work for the NSA, or building missiles, or building bombs, etc. have blood on their hands. If they can live with it that’s great for them, but I don’t respect their decision. (“they” includes my own grandfather).


  14. October 2, 2013 at 9:16 am

    word! that is horrific, i just keep faith that paranoia is ultimately unproductive, i had this look around bletchly park (a very interesting museum but slightly well… weird), the way they operated in the war was very compartmentalised

    however what strikes me most is that turning had a sort of solution, but an actual turing complete computer would have done the word searching way more efficiently, instead they built, whether this be through time considerations or paranoia… like 400 Turing bombes “computers”

    I think the moral of this is that this sort of paranoia is always less efficient, they will always need many many more people to chase the spotty teenager in his basement somewhere, or indeed worse, the scary terrorist… I have become convinced mankind needs to move forward from this

    a casual estimate of bitcoins current computing “power” is about 4 times the most powerful private parallel computer, also when some guys needed to do protein folding and couldn’t afford one (or likely tolerate the commercial association), they made folding@home


  15. George B.
    October 5, 2013 at 12:10 pm

    To be pedantic…. La Jolla, Princeton, & Bowie are not NSA. They are all branches of the Institute for Defense Analysis, a FFRDC like Mitre & Southwest Research Institute. That means they can pay far better than federal pay schedule. IDA’s sole client is the NSA.


  1. August 25, 2012 at 12:04 pm
  2. August 27, 2012 at 5:17 am
  3. August 27, 2012 at 10:41 am
  4. August 28, 2012 at 4:56 pm
  5. June 7, 2013 at 7:03 am
  6. June 10, 2013 at 6:40 pm
  7. June 10, 2013 at 8:03 pm
  8. July 8, 2013 at 9:41 am
  9. July 16, 2013 at 3:24 pm
Comments are closed.
%d bloggers like this: