please elaborate on comment for base:Data.Type.Equality.(==)?

[CC'ing Richard, as I'm guessing he's the author of the comment.] I have a question regarding the comment on the type family Data.Type.Equality.(==). "A poly-kinded instance [of ==] is *not* provided, as a recursive definition for algebraic kinds is generally more useful." Can someone elaborate on "generally more useful". Thank you.