I'm not sure you are using the same notion of "colorblind" as most people. All that is claimed is that much application code can be *compile time generic* over the async-ness of the I/O functions it calls. Stepping up the requirements to being runtime generic introduces some difficult concerns about the different error sets of different I/O functions, but async vs. non-async can be handled as in the above gist or by writing like one switch statement.
Months later I read Gavin's post (linked above). I found that post immensely helpful to understand Zig's design around sync/async function colors (thanks Gavin!). Gavin's post helped to illuminate for me that Zig functions do have colors, but that the compiler can infer the color in most usual cases. This is still very exciting!
As I see it, it's not to Zig's detriment whether it "has" function colors or not, I don't really care. I'm really excited about Zig either way. But I personally (coming from a decade+ of JS/Python/Go) found the verbiage I found used to describe Zig's behavior here confusing.
