I have trouble taking economic models seriously. I suppose ever since I learned that increasing the minimum wage doesn't increase unemployment (Card and Krueger, Dube, Lester, and Reich), I've basically come to the conclusion that for most models, the exceptions to the rule (externalities, indivisibility of goods, etc.) are more common than things that actually follow the rule. The one exception is that I still do believe that the bias in policy should be towards free trade, but I admit that neither trade models nor trade empirics are my specialty. And instead I heavily lean towards empirical regression analysis. Of course this requires a whole new set of assumptions (linearity, normal distributions), and maybe I'm just not good enough at math to know that these assumptions are just as stupid as assuming magical perfect free markets. Some statisticians (David Freedman, he of "randomization does not justify regression" fame) would probably think I'm foolishly naive for how much faith I put in (well-identified) OLS, even though I think I'm healthily skeptical about things like omitted variable bias, p-hacking, specification searching, publication bias, etc. I hope I'm not deluding myself.
I'm thinking about all this because of the classes I'm teaching. I have very little interest in the standard undergrad models, and I think I just taught a development economics course without drawing a supply and demand curve a single time. I just skipped all that and went straight to regression, the "real" way to get an answer. It might be crazy, but I think maybe not. Several in my cohort in grad school had not been econ majors as undergrads, and I don't recall having seen many supply and demand curves in grad school, so they might be perfectly wonderful economists and not know what a Cournot competition graph looks like.
It also came up in a discussion with an older colleague whom I very much like and respect. He said something offhand to the effect of "we'll have to see if all this quasi-experimental literature is still with us in another 15 years..." which can do nothing but belittle to the point of nothingness everything I know and value. So reasonable people can have different opinions on the matter, apparently. Like many things, the truth is likely somewhere in the middle. (But left of center. Definitely left of center.)