If you think of it from a return on investment/expected value point of view, why would you ever take time out of your day to refresh yourself on linear algebra and other numerical methods assuming its not out of personal interest just to apply somewhere that already has so many barriers to entry that don't exist elsewhere, on top of admitting it will not pay as much as other positions that would require these skills.
I'd rather spend my time focusing on applying at other positions that will pay just as much if not more without being required to spend nights and weekends relearning a skill that is just not used very much elsewhere and not requiring security clearances, and why would I waste my time applying there when other jobs that do require these skills compensate more for the effort.
On the other hand if this company did pay closer to Meta salaries (the comparison they used) then it gives all candidates, including US citizens who fit more of their eligibility criteria, that much more incentive to actually relearn these skills and makes the expected return on investment potentially worth it.
I agree with you on the overall reasoning, but I conclude that "this job is not worth applying to even if it took 6 seconds of preparation" rather than "this job is not worth applying to because it would take 6 hours of preparation".
