This may be an artifact of issue #8128 (closed) (see #8128 (closed)#ticket:8560#comment:77286), but I'm seeing examples involving data kinds where Typeable isn't deducible. This could be an artifact of other problems, BUT since with 7.7 onwards, we have baked in polykinded Typeable, things should always "just work™" right?

the error with current head is as follows (and i'm attaching the code + current finger print too)

EDIT: The request here is really for some generic representation derivable for GADTs. This is a Research Problem, but one that would really help a fair amount of ordinary folk. It won't ever work with Data, as we now know Data, because the types aren't up to the task. But, perhaps some future generic representation will work.

The original description remains unchanged above so that readers can make sense of the comments.