It seems to be making the argument that C++ 26 has new reflection features that outperform Rust code gen due to the unavailability of high quality proc macro features.
I see. I genuinely think the author is just trying to show off a new C++ feature, not argue that it's better than the Rust feature is. Of course, since they are different, there is comparison necessary, and proc macros aren't perfect. But it's really just about explaining why this is a cool feature, not trying to say Rust is bad.
Unless I misread, there is an assertion that Rust macros have to implement a lot of “compiler” logic themselves (whether or not through a separate crate) that is [will be] unnecessary with the new C++ feature. So not “bad”, but more clunky for the annotation implementor.
