Say they have a check valve in their titanium system that fails. They could spend a lot of research dollars on the fundamental science of why that valve failed or just replace it with a burst disk. From an operational engineering point, they may feel like they solved the problem. But from a scientific standpoint we haven’t learned much other than “check valves are bad”.
It’s akin to the old joke about a mathematician, physicist, and engineer trying to find the volume of a cow. The mathematician tries to formulate a triple integral for the cow’s shape. The physicist drops the cow in water and measures the displacement. The engineer says, “let’s suppose the cow is a sphere...”
Point being, often in engineering they’re shooting for “good enough to get the job done” not to increase fundamental understanding
Regarding the cost, the aerospace industry needs some scrappy young start ups. However, quality pedigree is a large chunk of the cost of flight and SpaceX has shown some gaps in the past. I think a big question is whether they can keep that scrappy startup up culture while also building processes robust enough to meet space flight quality standards