These tag types are not for general programming, they're for use as
constructor arguments.
It is more than just that, as we have the full set of heterogeneous comparisons between optional<T> and nullopt_t, as well as smart_ptr<T, ...> and nullptr_t. It is inconsistent that they themselves don't have a full set of homogeneous comparisons, especially given how simple it is to define them.
Don't pass them
into your hypothetical generic code and expect them to LSP to some
other type (which type? There's an open
set of such types.).
I agree that LSP has nothing to do with this.
I also agree that such a proposal would have to go to EWG as well as LEWG. The homogenous equality comparisons for nullptr_t are defined at the language level and not at the library level, and these would need to be as well.
--