Isn't this changing type from being conditionally copyable to unconditionally copyable? And then there is no way to define such type with the suggested syntax?
In the review thread there was a long discussion on the matter, and seems like no better option has arise.