Almost never do software companies even attempt to design secure systems. I'm not sure this requires new fundamental research so much as slightly giving a shit.
Especially since the world isn't Dilbert where your boss goes "oh, authz? lol nah, just yolo it" and you go "dangit, alright boss". Instead, security requires eternal vigilance and zero missteps along the thousands a project takes in its lifetime.
I think there's a reason HNers who pitch this idea never give any concrete examples of entailments of their proposal: it doesn't even sound good superficially. e.g. How this actually changes security issues. In fact it just sounds even more convenient to blame engineers.