I agree that it's impractical in the forseeable future (battery technology will need significant improvements - no one will tolerate a gas-powered version because it'll sound like a flying chainsaw; sense-and-avoid will have to improve very significantly to avoid collisions with birds, powerlines, people etc, failsafes will have to be better than "just shut the motors down" because that could drop the drone on a person or a moving vehicle and cause injury and/or damage. These are all hard problems that won't likely get fixed in the timeframe Bezos quotes. So I agree it's just a publicity stunt.
However, n-copters (n>=1) are not necessarily good-weather toys only. Even fixed-pitch types like the AR drone can withstand pretty gusty winds outdoors. You can also make a small something-copter VERY good in very gusty winds if it has collective pitch control on its rotors. Collective pitch rotor systems can react extremely fast. I have a number of RC helicopters, and even my tiny T-Rex 250 could be flown in 20 mph gusting winds because it was collective pitch (a fixed pitch heli would be lost with that much wind). I've flown my T-Rex 600 quite happily in gusty winds over 30 mph - I had to give up not because the heli was hard to fly, but the wind was blowing water out the tear ducts of my eyes and making the helicopter too hard to see.
The main problem with a drone (of any number of rotors) would have that was sufficiently powerful to lift a 5 lb package AND deal with potentially strong gusting winds is that it'll also be very powerful too. A blade collision with a person would likely lead to a pretty serious injury; the Amazon drone would need to be at least as powerful as a T-Rex 600, and model helicopters of this power have already resulted in deaths.