review type checks that should be tautologically true

[sethbrenith]

As @Penguinwizzard mentioned here,

This feels like there's a few situations in which a smart compiler would be able to make it always return true - e.g. with a static build and ltcg, I think it may be legal to optimize the "GetTypeId() == mytypeid" ::Is() checks for various object types in our type system to always return true when working on the same type as that passed in - since there's no avenue other than UB to change the typeid of valid objects of that pointer type to something else. Since we're already doing effectively that, it's not a regression with this change, but in a future change we'll most likely want to change this to being closer to VarIs in terms of signature, with the exception of making the last line AssertOrFailFast(VarIsImpl<T>(obj)).

This issue is a reminder to come back to check on calls to VarIsCorrectType.