> At least to me, it seems pretty well-scoped as a workaround from the requirements that the compiler has around needing to use generic type parameters in type definitions

The reason those requirements exist is (primarily) to do with unsafe code. Specifically it’s about deciding the variance of the type (which doesn’t matter for a truely unused type parameter).