Remove list of implications of dynamic/object identity conversion from 8.7

Branched from #985.

In section 8.7 ("The dynamic type") there's a list of implications that stem from `dynamic` and `object` being identity convertible (e.g. about tuples and generics). These are already covered in 10.2.2, and it would be better to just link there. At the moment, if some new identity conversion is introduced based on some sequence of types (as per generics and tuples), we'd need to add to 10.2.2 *and* 8.7, which seems redundant to me.