“The Silicon Valley guys seem to have the attention span of 3-year-olds.”
The fact the product benefits from faster innovation is a function of more efficiently sharing knowledge and data, not spending less time on individual components.
Designing and testing an engine can be done in an agile style. There's nothing intrinsically "software only" about the methodology.
NB: I ran a startup that made software to manage requirements in big projects. I've spent lots of time thinking about different ways to manage things. Including how not to, but that was after we failed.
The only difference for a SaaS startup is that they're their own customer for product development (in the sense that they build and test in-house rather than someone else paying them to build a product). The cost of failure is still there, it's not measured.
It's weird when I realize I can just push something out the door and not have to wait for the reams of paperwork to be complete.
I read a book recently about the history of Supermarine, and what jumped out was the attrition rate of their test pilots. Ditto for the post-war British jet industry.
There is a trade-off point between safety and fast development, but I'm glad we've moved to the side of safety.
Not only that, they have a bad habit of adopting disproven solutions simply because they 'are mature'.
That is I believe Musk's premise with his rocketry division. It couldn't possibly be this complicated, can it? I hope he turns out to be right. But if he is, gonna be a lot of smug engineers having to take early retirement.
One hull loss per nineteen attempts is not an acceptable failure rate for "the airplane industry", is it? It's difficult for me to put into words the level of respect and admiration I have for the risk-averse, plodding, and methodical aerospace culture that has created modern aviation. (Seriously, have you ever considered the absolute miracle of engineering that is a modern jetliner?) To lump that profession under the same rubric of engineering as the one where a bunch of overcaffeinated "disruptors" write software to get people to click on ads seems... unjust, to put it mildly.
What I meant to convey is that I expect someone to do to commercial aviation what Musk is trying to do to rocketry, and what I expect that to look like is for more effort to be placed into reducing single points of failure by designing complementary systems instead of just putting three of everything on the vehicle and then adding two more layers of stuff to manage it. And also for them to act like progress in software development processes actually applies to them now instead of in thirty years. We got no end of grief for refusing to do pure waterfall, and they were conspicuously silent when we're were able to fix fundamental up front design flaws in 3-8 weeks, and without doing a reenactment of Macbeth starring William Shatner, set in a rehab clinic.
Then again, I expect something to happen to improve the caliber of software developers out there as well, so we'll see how that goes. Often when I expound on the qualities of my profession I am intentionally and consciously imagining that the 30% of my peers that are responsible for 70% of our problems either don't exist, or are off doing something that's not important.
Selling thousands of engines too soon and realizing they all suffered from a critical engineering issue could bankrupt your company. There's a lot on the line!