I am going to say this and by-God it is not directed at anyone in particular
. My thoughts nonetheless (not perfect by any means either)
-Conceptualize
-Design (still requires engineering prowess and premise)
-Validate and bounce concept with existing design by plug and play or iteration (Engineering is involved here too)
-Iterate and validate further independently (still require engineering)
-Validate in realtime scenarios (engineering is still relevant)
-Scale, Scale-up, Scale-out, Scale-across (based on the need => industrialize)
When you start with the thought that we are going to design an F35 you have already bound your thought processes, designs and capabilities that are outside your control.
I would rather take a set of fresh engineers and scientist with least amount of industry biases and design a product that may become groundbreaking. Yes it takes time, but guess what if you are taking 20 years to design an existing design... well.