I don't get quite the same impression. The sense I get is more that such a change would basically need to happen "bottom-up":
> Some of [the scenarios that aren't fixed by this proposal] are just the normal dependency management issues. If you build a library on top of something else that uses one of the changed types (such as intmax_t or something else), then you can’t really upgrade until your dependents do.
> <snip>
> For those of us in large ecosystems who have to write plugins or play nice with other applications and system libraries, we’re generally the last to get the benefits.
In which case the benefit of the proposal (as far as I understand) is that such bottom-up changes can occur without forcibly breaking other consumers.