I'm not entirely sold on the idea, because the article shows no example of defining and using special error codes. It leaves an impression that if anything you want to recover from is IO and HTTP (no HTTP/2, too) then it's all good.
Still it looks interesting and I will take a deeper look into it later