Per my understanding, there's less concern over unifying, say, the strong and electroweak interactions. You have to posit some ridiculous coincidences to explain facts like why the charge on protons and electrons are exactly opposite, but nothing whatsoever prevents you from doing so. Then you get separate, compatible theories describing separate domains. That's basically what the Standard Model is: QCD + electroweak, agglomerated as separate interactions, not unified.

However there's better reasons to be concerned with the relationship between general relativity and quantum mechanics. The problem is that while their domains are generally separate, we know they must intersect in cases like black holes or the early universe. And in these scenarios, the theories produce incompatible predictions, leading to issues like the black hole information paradox.

Physicists are very interested in any theory of gravity that makes predictions compatible with quantum mechanics. Some of the contenders unify with the other forces (string theory) while some do not (loop quantum gravity). A unified theory would be huge, but a non-unified "agglomerated" theory that nevertheless makes a single, unified prediction would be almost as exciting.

So the answer to your question is that popular physics writing often uses "unification" as a shorthand for "unified prediction" rather than "a single interaction." That's what this article is doing: the referenced Wheeler-DeWitt equation is NOT a unified theory because it does not explain QCD or electroweak.