I took a huge number of metaphysics courses during grad school and, over time, I changed my mind about pretty much everything, other than my dogged commitment to the law of non-contradiction. Then I stayed stuck, because, of course, I eventually landed on the correct answers. I thought the NYTM article on free will was pretty good, but I also no longer find the question very interesting. There are lots of uninteresting metaphysical questions. Here are a few obviously correct metaphysical conclusions not worth thinking that much about (to me).
Free will: The universe is either deterministic or it isn't. This has nothing to do with free will. We have it. Yes, we often make mistakes in attributing agency to ourselves and others. But often we don't. It is frequently possible to have done other than what we did in fact do. The trick is understanding the relevant sense of “possible,” which has nothing to do with ultimate issues about the nature of causation.
Ontology: Quine is right. To be is to be the value of a bound variable. That is, if something plays a role in our best explanation of some phenomenon, you should believe it exists. Otherwise, not. God, for instance, is the best explanation for nothing. That's why you shouldn't believe in God, or the posits of string theory. (People like Megan who hesitate to call themselves atheists because they cannot “prove” nonexistence are simply confused about ontological commitment. If Megan's p for “God exists” is so low (“vanishingly unlikely”), then God must play no role in her economy of explanation, which is all there is to being an atheist. You don't just get to decide whether or not you are one.)
Universals: There are “repeatable” fundamental “kinds”, which explains why there are relations of causal necessity. Realism about universals confuses the semantic generality of concepts for ontological generality. “Instantiation” and “exemplification” relations add nothing useful to property instances (tropes). There are individuals and that's it. If two are in different locations but it would have made no difference whatsoever to the history of the universe had they been switched, then they are two of a kind. We can have essentialist scientific realism without essences. But it really doesn't matter much: choosing a particularist or universalist ontology is just a Carnap-style choice of vocabulary. It's an open question whether the more elegantly parsimonious vocabulary works out better in the work of explanation. It's probably easier to think like a realist.
Modality: There is exactly one possible world, the actual one. Pace Lewis, more than one possible world is the best explanation for nothing, so that's that. “Possible” means “not inconsistent with the fundamental laws that govern basic kinds.” Modal statements about fundamental kinds (“gold might have had a different atomic weight”) may be grammatical but are not meaningful. Whereof we cannot speak, etc. Bonus: modal epistemology is just epistemology, and epistemology is the psychology and sociology of truth-tracking. Unless there is reason to think that our haphazardly evolved and organized and imaginative abilities for some bizarre reason happen to reliably track truths about the fundamental laws governing ultimate kinds, it's hard to see what thought experiments about transparent iron or a molecular duplicate but non-conscious Zombie me are even supposed to be about, much less explain.
Qualia: Yes! They play a computational function. (This is a joke! I don't know that at all!)
Don't mean to bore you. It's all pretty obvious when you just come out and say it like that, huh?
Also, it has come my attention that some readers of this blog find philosophical jargon forbidding. Sorry! But if a man can't use clubby, exclusive, abstruse jargon on his own blog, where can he? Anyway, if you end up on a game show, and they need the answer to the problem of universals, you're in luck. OK… Back to hard, interesting stuff, like happiness and inequality.