4 ms·

Speaking from experience, I applied to the NSA a few years ago and its surreal how unrealistic the expectations were. a major problem with NSA employment is th

Speaking from experience, I applied to the NSA a few years ago and its surreal how unrealistic the expectations were. a major problem with NSA employment is the mandatory 5 year contractor like status period with half-pay. the NSA expects seasoned professionals from the private sector to join its ranks with a 33k salary? get real. The agency also prides itself on an almost obscene reliance on Polygraphs, a technology no more accurate or scientifically sound than a babushka shuffling chicken bones and reading tea leaves. I also wasnt exactly blown away by the interview questions. my technical screening for a Linux position came from former members of the military, who grilled me on 2 unix questions before asking if they could switch to windows, confessing they 'didnt know much linux.'

jessriedel 9y ago
> The agency also prides itself on an almost obscene reliance on Polygraphs, a technology no more accurate or scientifically sound than a babushka shuffling chicken bones and reading tea leaves. I can't find out why I see smart people repeating this all the time. It's not true. Polygraphs are very flawed, and they are especially vulnerable to preparatory countermeasures, but they have non-trivial predictive power. When I ask people to justify their dismissal of polygraphs, they generally point to the fact that they are not allowed in court, but this is a terrible reason. Hearsay isn't allowed in court either but that doesn't mean it isn't useful.
- dclowd9901 9y ago
  If it doesn't stand up to "beyond a reasonable doubt" how can you have such faith in it? Hearsay is a fine example, as I wouldn't want my innocence predicated upon someone's flawed memory or whim.
  - jessriedel 9y ago
    Because "beyond a reasonable doubt" is a much higher level of certainty than we have for making many decisions in life? The NSA isn't jailing people based on a lie-detector, they are just using it as a necessary but not sufficient criteria for granting security access. It's OK to deny such access based on weak or circumstantial evidence.
- dilyevsky 9y ago
  "can't find", huh? well did you even try? https://en.wikipedia.org/wiki/Polygraph#Validity https://en.wikipedia.org/wiki/Polygraph#Validity has plenty of actual studies referenced.
  - jessriedel 9y ago
    Yes of course. Did you? The balance of evidence is strongly toward polygraphs having non-trivial predictive power. People frequently conflate unreliability with being fully random.
    - dilyevsky 9y ago
      Don't know what "fully random" is but 60% sounds like pretty close to a coin toss.
      - jessriedel 9y ago
        If you require applicants to pass five independent tests that have 60% accuracy (e.g., interview, polygraph, family/friend interviews, recommendations, criminal record), good applicants are more than 7 times more likely to pass than bad applicants. Indeed, most of our important decisions are made by combining many weak pieces of evidence. Even if some other tests are more accurate, additional less-accurate tests can easily be worth running if their costs are modest. Again, this is not an argument that the information provided by the polygraph actually passes a cost-benefit analysis in any particular case, it's just a criticism of the "all or nothing" mindset regarding accuracy that many folks have on this topic.
dilyevsky 9y ago
> the NSA expects seasoned professionals from the private sector to join its ranks with a 33k salary? they don't. like any other military/police agency in the world they want them young and dumb.

jessriedel 9y ago

> The agency also prides itself on an almost obscene reliance on Polygraphs, a technology no more accurate or scientifically sound than a babushka shuffling chicken bones and reading tea leaves. I can't find out why I see smart people repeating this all the time. It's not true. Polygraphs are very flawed, and they are especially vulnerable to preparatory countermeasures, but they have non-trivial predictive power. When I ask people to justify their dismissal of polygraphs, they generally point to the fact that they are not allowed in court, but this is a terrible reason. Hearsay isn't allowed in court either but that doesn't mean it isn't useful.

dclowd9901 9y ago
If it doesn't stand up to "beyond a reasonable doubt" how can you have such faith in it? Hearsay is a fine example, as I wouldn't want my innocence predicated upon someone's flawed memory or whim.
- jessriedel 9y ago
  Because "beyond a reasonable doubt" is a much higher level of certainty than we have for making many decisions in life? The NSA isn't jailing people based on a lie-detector, they are just using it as a necessary but not sufficient criteria for granting security access. It's OK to deny such access based on weak or circumstantial evidence.
dilyevsky 9y ago
"can't find", huh? well did you even try? https://en.wikipedia.org/wiki/Polygraph#Validity https://en.wikipedia.org/wiki/Polygraph#Validity has plenty of actual studies referenced.
- jessriedel 9y ago
  Yes of course. Did you? The balance of evidence is strongly toward polygraphs having non-trivial predictive power. People frequently conflate unreliability with being fully random.
  - dilyevsky 9y ago
    Don't know what "fully random" is but 60% sounds like pretty close to a coin toss.
    - jessriedel 9y ago
      If you require applicants to pass five independent tests that have 60% accuracy (e.g., interview, polygraph, family/friend interviews, recommendations, criminal record), good applicants are more than 7 times more likely to pass than bad applicants. Indeed, most of our important decisions are made by combining many weak pieces of evidence. Even if some other tests are more accurate, additional less-accurate tests can easily be worth running if their costs are modest. Again, this is not an argument that the information provided by the polygraph actually passes a cost-benefit analysis in any particular case, it's just a criticism of the "all or nothing" mindset regarding accuracy that many folks have on this topic.

dilyevsky 9y ago

> the NSA expects seasoned professionals from the private sector to join its ranks with a 33k salary? they don't. like any other military/police agency in the world they want them young and dumb.