I struggle to understand how what you talked about is relevant here. The problem parent talked about is how this laziness could be implemented without losing what makes f-string awesome right now. It’s not viable, which backs my reasoning why at least one alternative format method is required.
That said, I also struggle to understand how you’d claim parent clearly not using Python enough, when your description of str.format shows a lack of understanding to it yourself. One of the advantages of str.format over %-formatting is exactly that you do not need to modify the arguments passed to str.format when removing components from the source string:
>>> '{0} {1} {2}'.format('a', 'b', 'c')
'a b c'
>>> '{0} {2}'.format('a', 'b', 'c')
'a c'
Or preferably (using keyword arguments instead of positional):
>>> '{x} {y} {z}'.format(x='a', y='b', z='c')
'a b c'
>>> '{x} {z}'.format(x='a', y='b', z='c')
'a c'
But again, this doesn’t matter to parent’s argument. Nobody is arguing with you that f-string is better than alternatives for what it can do; we are trying to tell you that there are things it
can’t do, and you did not get it.