Let's pause for a moment for a meta-level reflection. You're engaging in metacontrarianism, with the relevant uneducated/contrarian/metacontrarian triad being:

Dubai is a country / No, Dubai is part of the UAE / Dubai has a lot of power and autonomy within the UAE.

The trouble with metacontrarianism is that metacontrarians often seem to forget that even if they're right -- that is, even if the third level of the triad is true -- the first level is still wrong. In some sense, you have to pass through the second level in order to legitimately claim the mantle of the third. (Here, "pass through the second level" means not "go through a stage of being at the second level" so much as "understand why, and in particular that, the second level is an improvement over the first".)

I submit to you that if Alice thinks Dubai is a country because she's never heard of the UAE, and Bob thinks that Dubai is the UAE's version of Istanbul, Bob's model of the political geography of the Arabian peninsula is still better than Alice's, even if Carol, who thinks that Dubai is so different from the rest of the UAE that it "might as well" be a country in its own right, has a better model than Bob.

Now, to return to the object level, I don't actually see why Carol's model is better than Bob's. I don't know that much about the internal politics of Turkey, but I assume that Istanbul, being a major city, is culturally and demographically different from most of the rest of the country, wields a lot of influence in the country's politics, and has governmental policies that most other parts of the country don't have. For that matter, the same is true of New York City, whether regarded as a part of New York State or of the United States. In neither of these cases do I see any need to give up the model that has these cities being politically subordinate to the nation-states (or states) that contain them, and I don't see how the case of Dubai within the UAE is any different (or, anyway, different enough). And, conversely, even if Monaco is heavily influenced in its policies by neighboring France, I don't see that as sufficient reason to remove from my model the notion that Monaco is an independent state, because otherwise we might as well say that Canada is part of the U.S., et cetera.