The Go team's stance on generics has never been "generics = bad", but "generics have advantages and disadvantages, and if we find an implementation which finds a good compromise between the advantages and disadvantages, we will do it". Which they now did. Let's hope it really turns out to be a good compromise...