I had a lightbulb moment when someone said 'the point of iterative approaches is not to find bugs, it's to do something (small) successfully and build confidence+learn'. There's a subtle but important difference between the iterative approach that SpaceX takes and 'debugging through exhaustive retries', and I'm worried NASA would look like the latter (and admittedly, some of the more recent starship launches look that way too).

The ability to pick a small-but-well-defined goal as an interim milestone - and stay focused on it - is a key skill, and too often I've seen waterfall-like companies slowly scope-creep their first MVP until it's a lumbering mess. You almost always need someone with a strong personality to push team to 'get it done', and that level of ownership is really hard to come by in an organization historically built around ass-covering.

I think Commercial Crew is the right model for NASA. Pick the design objectives, provide some level of scaffolding regulation (i.e loss-of-crew calculations), and then contract out to private sector to actually 'get it done'. (Yes Starliner was a failure, but Dragon is definitely a success. A 50% hit rate and success of the program overall is better than Artemis)