If this is a question about the meanings of words, that's really just an empirical question about their usage. If it's a question of whether "cars" forms a meaningful cluster in conceptspace, we have lots of different ways of addressing that question that entirely sidestep the Realism/Nominalism debate.

And if it is about how a fundamental aspect of reality -- identity and difference -- works then we don't. The debate about universals is ongoing with people like Roger Penrose and David Armstrong weighing in.

This sort of equivocation does not happen with key terms in a field that has its shit together.

Uh huh. So what exactly are physical laws?

I largely agree with this answer. My view is that reductionist materialism implies that names are just a convenient way of discussing similar things, but there isn't something that inherently makes what we label a "car"; it's just an object made up of atoms that pattern matches what we term a "car." I suppose that likely makes me lean toward nominalism, but I find the overall debate generally confused.

I've taken several philosophy courses, and I'm always astonished by the absence of agreement or justification that either side can posit. I think the biggest problem is that many philosophers make some assumption without sufficient justification and then create enormously complex systems based on those assumptions. But since they don't argue for strenuous justification for the underlying premises (e.g. Platonic idealism), then ridiculous amounts of time ends up being wasted learning about all the systems, rather than figuring out how to test them for truth (or even avoiding analytical meaninglessness).

I'm even less clear on what a "meaningful cluster in conceptspace" means than I am on the traditional philosophical formulations of the problem. What would a "meaningless" cluster in conceptspace look like? Is there a single unique conceptspace, and how is it defined?

Good philosophers must beware of equivocation. Universal is an ambiguous term, so taboo it and distinguish several things it's stood for:

1. predicate - term that can be applied repeatedly. (If 'nominalism' is meant to reduce all universals to predicates, then it's an ill-conceived project, since it seems to be trying to explain commonality in general by reducing it to commonality between words; but if the latter is left unexplained, then commonality itself is left unexplained.)

2. common nature - something intrinsically possessed by all the entities that share a property. An abiding 'essence,' some kind of 'quarkhood' that inheres in all the quarks. Common natures are a posit to explain similarity. They are worldly, thus completely unlike Platonic Forms.

3. common cause - a single cause that has multiple effects. A Form acts as a common cause, but not a common nature, since on Plato's view they causally produce the recurrence of nature's patterns 'from the outside.' (In some ways, they're an anthropocentric precursor to Conway's Game of Life.) Like common natures, common causes can be posited with the intent of explaining why our universe exhibits similarity. If the question 'Why do properties recur at all?' or 'Why are some characteristics of the world the same as each other?' is well-formed, then there is nothing mysterious or ill-conceived about these posits, though they may perhaps by theoretically unnecessary, unenlightening, or ad-hoc.

If you expect it to work like science, it will look like bad science. That is what most of the "philosophy is bad" criticism boils down to.