Indeed and unfortunately. I've been reading up on "the binding problem" in AI lately and came across a paper that hinged on there being an "object representation" which would magically solve the apparent issues in symbolic AI. In the discussion some 20 pages later, the authors confessed that they, nor anybody else, could define what an object was in the first place. Sometimes the efforts seem focused on "not letting the other team win" rather than actually having something tangible to bring to the table.