but they generally are not the type to be filtered by an email domain requirement.
They are exactly the type to filter by something as "trivial" - 99% of their target audience is Math nerds with .edu emails.
The other 1% will go the other 99% of the way to acquire the needed materials to satisfy the target condition. Which in this case, is a room-temperature check compared to the challenges.
They do, as do machine learning/AI firms, insurers and other actuarial firms, and, of course, academia. Like every other postgraduate specialization there are subreddits full of threads of people discussing "what am I going to do with this doctorate I'm getting if I don't end up becoming a tenure-track professor?".
PhD's in Math are very rare, and uneconomic. Aside from Wall Street and Langley, no one outside of SV Talent recruits are paying for someone who has spent their prime thinking years considering the viability of certain types of "up-my-sleeve" numbers - no one else has that capital for specialty, almost certainly fruitlessly, since infosec advantage isn't sum-net-zero. Any APT that will pay will have a slight advantage; that opportunity cannot be simply absconded.
That is why NSA skews the average with their hiring practice, let alone indirect contractor influences - although the pure math SME's are held tighter to the chest than even private contractors can boise.
> aren’t logically inconsistent.
Sure, if you remember PhD's in Ecliptic Curve Cryptography or Number Theory or applied but pure 'XYZ' field of promising arcane mathematics are extraordinarily rare, and skew towards a certain demographic as well. The motivated, undeterred, socially-inept few.
And the former is similarly not evidence that they mostly hire people with edu emails.
I can tell you, objectively, statically, that those who have Math degrees have a lower chance at needing help resetting their password to their .edu email. And a much, much higher chance at actually graduating with a grace period and mental clarity to leverage it for a brief window during their opportunity. You can (kinda) check yourself at https://nces.ed.gov/ and https://analytics.usa.gov/,
As the excellence required increases, the numbers get low enough, you can hire ALL the talent. And have enough 'explanatory' budget left over for institutional-preserving things and normal bureaucratic neo-con noise.
You can hire more Math PhD's than anyone, and still mostly not hire math people.
There are very, very few Math PhD's that can, even theoretically, threaten the current risk portfolio of our nation. But if they even did exist, you would not want to signal them out by being the only one they hired.
All signals require noise. Work cannot be performed absent a temperature gradient.
PhDs in math are not in fact rare, and having attended numerous cryptography conferences I can assure you they're a lot more socially normal that computer nerds in general are.
Even if they hired the sum graduating phd class of every math program in the country it wouldn’t change the fact that math phds are not their hiring target.
They have to hire N non-math phds for every M math phd they hire to support their hiring metrics. Like every other large technical bureaucracy in the world.
None of that has anything to do with advanced capabilities and, again like every other technocracy, has to do with management and ops.
The other 1% will go the other 99% of the way to acquire the needed materials to satisfy the target condition. Which in this case, is a room-temperature check compared to the challenges.