I think the GP refers to Zig's approach[1], which is just “do mark functions as sync or async in there signature, just switch between two modes at compile-time using a global flag”, which I don't think fits well into Rust's values of being explicit about what is happening.
Basically using "await" just means you are declaring opportunity for concurrency at the call site, but it is still valid to await a synchronous function.
And it is valid to call a function synchronously, even if it internally supports concurrency via await.
I’m super interested that you think that’s something that can exist, how would it work, what would it look like in your mind ?