We've had flying cars almost as long as heavier than air flight - they're not some future tech that has yet to be invented, as much as everyone pretends that they're some magical thing just around the corner. We've had them for over 100 years, and they've sucked as cars, and sucked as airplanes for 100 years. Any materials breakthrough that will "make flying cars possible and too cheap to meter!" can also easily be applied to both cars and airplanes to make them better (see carbon fiber).
Drone delivery is worse than wheeled delivery in just about every option, but "flying is the future," so people have gone down that path, dorking around with it for a few years (which, honestly, sounds like a ton of fun). The results end up being quite reliable, which is "There is no way in hell we can actually make this work in any practical, economical sense."
I think that when you think of sidewalks you're thinking of how they look in your town (assuming you actually have sidewalks) they look quite different in different areas - in the southwest they're wide, expansive and quite level - in new england they're broken frequently by tree roots and frost heaves and, honestly, frequently just don't exist at all in a lot of the cities.
I do agree that people fetishize "Flying X" but saying that drones are worse than wheels in just about every way is pretty ridiculously hyperbolic - they have tradeoffs and it's quite possible that the market winner will be whoever sends out a van with wheeled delivery vehicles for most runs that can call in a drone if it runs into difficulties.
I meant wheeled drones, not flying ones.
Drones may be better for very quick delivery or where road network is very poor.
See https://news.ycombinator.com/item?id=28070084 with amazing https://www.youtube.com/watch?v=jEbRVNxL44c (delivery of medical supplies in Rwanda) and mentions ship-to-ship transport.
Burritos are particularly robust to delivery (along with Pizza, Chinese takeaway, and curries). Most finer meals, and even items such as sandwiches or hamburgers, aren't. I don't see a large market for duck a l'orange avec quadcopter however.
Even food deliveries, though, probably don't pass the sniff test.
For me, seeing some new fancy tech prototype or proposal that looks to be unsuited to general commercial or residential use, I tend to see a submarine military application. Drone distribution in a combat environment could conceivably make sense (though it would risk blowing cover in many cases). Emergency response, research, fieldwork, construction, and disaster response are a few other potential fits. Much less price sensitive, high time sensitivity.
But mass delivery of random consumer stuff in an urban or suburban environment? Not very likely.
I suspect the problem just sounds too cool and is sufficiently engineer-catnippy for anyone to say "no" in time.