The one big difference with the SpaceX organization is the willingness to "blow stuff up" while they are developing, which is enabled by the amount of sensor data they get on each test. That allows them to have a higher revision cadence that decreases development time. Compared to old space tech, where a rocket failure (even during early development) spells the end of the program (such as the X-10 -- it was looking good until a failed landing leg caused an impressive looking loss of vehicle, and then no more program).