> No, we'll still get the years of use from the "old tech" while new tech is being developed
(NB)-C < (N-k)B'-C in some cases. (Yes, I realize that NB ignores time value of money. For N>>k, that can be reasonable if C is not too large.)
The numbers matter.
> What if we decided not to build cars until we developed the Prius?
What if we didn't invent false choices? (Hint: I'm not saying always wait.)
You're the one insisting that all-in is always the right solution. That's simply false, as is always wait.
The numbers always matter and they're often context dependent. (Example - pretty crappy solar made sense in some situations while it didn't in other situations. Better solar makes sense in some of the latter, but not all. And so on.)