Fortunately the same principles of physics and mathematics apply to both me and Google (and the #3 or #4 most powerful/smart organization, NSA). I trust science a lot more than I trust either every member of every organization to obey the law, or the law to remain both public and supportive of my interests.
Absolutely true in 1970, but probably much less true now (they probably ARE experts in traffic analysis, though, since ~no one but intel agencies cares about that yet). Academia and commerce tend to surpass the government in every OTHER area where they participate, it's just that no one cared pre-70s. I don't think they're using their huge budget efficiently, just like every other government/military program. The bulk of their budget and effort is going to collection, mainly satellites, physical installations, staffing for those, etc.
Judging by what I've seen of USG security, their COMSEC side (defense) is actually comparatively weaker than best practices in most of the high-end civilian world. Their only saving grace is a focus on physical separation and hardware implementations; otherwise they'd be totally pwned, and they seem to be moving "into the cloud" which will lose their their existing advantages.
And, if they had a magic way to factor or whatever, it would be so highly classified (like Enigma break in WW2)that they could never use it on anything which wasn't totally justifiable through other plausible means, and of such incredible value (like "killing OBL" or higher), that I can act as if they can't factor.
I'm pretty sure the NSA has an active program to recruit high-level strictly theoretical mathematicians specializing in algebra and algebraic analysis (typically at the undergrad level).
Agree that the US's COMSEC is degrading, rapidly. As is the USN's nuke reactor program. Sigh.
They do, although the vast majority of NSA staff are more like computer techs/middle managers/etc. Sadly.
It's also a lot harder, IMO, to convince a top theoretical mathematician to go to NSA vs. remain in public, since there aren't exactly amazing resources NSA can offer (better chalk?), compared to someone who needs access to multi-million or billion dollar resources. In exchange for better chalk (and access to other classified work, sure), you can't publish, so you'll never win a Fields, be respected by your mom, etc. (A guy who really is into nuclear physics could be tempted by DOE, by comparison.)
I think NSA would have a much easier time recruiting "true believers" in US national defense than anything else. i.e. "yes, you're an expert in theoretical math, but rather than doing purely abstract stuff, you can help protect your country from the evil terrorists!". An easier case when the threat was existential nuclear annihilation from a USSR your relatives may have fled in the past, than now where we're going up against mud hut dwellers who can't afford shoes and take over planes with boxcutters.
NSA also does the bullshit "EOE" recruiting, particularly in the high school and undergrad level. Sorry, but most high level math people are not particularly "diverse".
Plus, go to the math department at top universities. Find the US citizen. Find the US citizen who isn't particularly counterculture or does recreational drugs or whatever. Convince him to work for GS/GG wages (maybe $100k, tops?) vs. a tech startup or a hedge fund. With paralyzing bureaucracy, uncertain funding in the next 10-20 years, and stifling work rules.
I would not want to be the NSA's recruiting department.
