You say innovations, I say counterproductive bullshit. The idea to write a "secure" operating system on the basis of a) C, the single worst and most actively anti-secure language in the history of computing b) unix with its ambient authority/everything is global and myriad of other design flaws is just fucking crazy and about as helpful as efforts to "fix" cattle slavery by tinkering with the design of whips and chains to make them more humane.

So which of these layers upon layers of crap actually really, fundamentally fixes any of the core underlying problems (like using, in 2023, a memory unsafe language full of UB footguns to implement network facing services and not having the semblance of a proper security model)? Rather than just making them incrementally harder to exploit at the cost of ever increasing complexity and overheads that also get imposed on saner technologies which do not suffer from these problems in the first place?

How much faster (never mind secure, simple and robust) would our core computing infrastructure actually be if not everything was organized around a quixotic quest to partially mitigate some insane design decisions of C and the fact we're using a half-century outdated OS design?