All true, agreed. IMO SpaceX also took the fifty-cubic-kilometer mound of paperwork surrounding orbital launch, surveyed it carefully with a steely gaze, then pulled out their unit and pissed all over it.
OK, I say that with some levity. From what I've seen, it was far more subtle, and this is the aspect, the "secret sauce", that Anduril might be looking to replicate.
What they really did - one of the things, the thing that's in my core interest - was go through the . . tens of thousands . . of specifications, and isolate the ones that actually mattered, versus the ones that were good but onerously labor intensive[1], versus the ones that weren't only not necessary but which were unfalsifiable[2].
One of SpaceX's great advantages is that they could just deep six such very-expensive-hourly-headcount piles of "work", before they got stuck up to their axles in an entirely speculative systems engineering approach or some other brilliant piece of paperwork.
[1] Like looking for "commonality" when no such thing exists or is even possible. SpaceX doesn't waste time trying to show how "modular" their crap is according to some spec or other - when a variant goes far enough afield, they don't hesitate to cut the cord and roll a new product. False equivalence is a poison pill across all of aerospace, but in MIL-STD especially.
[2] Like the DoD mapping of Mil-STD-881 to MIL-STD-498 to goddamn scrums - how the hell do you tell if that's useful or not? Or if it's an actual thing in the world? You can't. It's horseshit, a six million dollar Powerpoint slide.