> Most of the US military, the best funded in the world, is decades behind the private sector on computing technology
This assertion embeds many assumptions not in evidence. Closed source computer science R&D, not only to which the US defense establishment have access but in many cases funded, is decades ahead of open source and academia in many key areas, not just in obvious areas like cryptography but data structures and algorithms, as well as exotic hardware.
It doesn't explain UFOs, but the idea that the most advanced public computer tech is decades ahead of what elements of the US DoD have doesn't pass the smell test. At worst there is parity and in some areas there is quite a bit of anecdotal evidence of capability that should not be possible if you solely went on public computing literature. (As a well-known example, there is empirical evidence that graph analysis capabilities exist that are vastly more scalable than is explainable by literature.)
Not everyone aspires to be a code monkey at Facebook for the money. There are extremely talented computer scientists working in US government, especially the secretive parts, that could instantly get a job at any FAANG company if they wanted to -- I've met many of them -- but they don't. The notion that the US DoD does not have access to "the best engineers and scientists" is dubious.