Carrier has arrived: https://www.youtube.com/watch?v=7B5DPovWY7o
How: drone with a harpoon, or a particularly pointy model rocket. Perhaps you'd attach a tether, so you can bring the ship to a specific spot (and reel it in gently without having to compromize its buoyancy so much it crashes hard).
How far do you think you could drag a drone-filled zeppelin before you were detected? How long would it take you to gently reel in a floating warehouse while Amazon alerted authorities and private security about your multimillion-dollar heist?
But now we have computer control stabilizing it in flight, making it vastly easier and safer to fly. Of course Physics hasn't changed, but advances in technology matter.
The only way that might ever change is if we develop a form of nanotubes strong enough to allow use of a vacuum rather than gas.
Besides, even at just 7%, if the Hindenburg had been "filled" with vacuum instead of hydrogen, that would have been an additional 37,000lbs of payload capacity, which would have almost tripled it's lift capability (21,000lbs as flown).
Additionally, helium, which is probably the only gas the public will accept, is almost double the density of hydrogen, and we have a limited supply of it.
I've put some thought into this from time to time and I see a few alternatives. One is to use a tow line with some sort of auto-belay, and shield the rotors so that the line can't get snagged under usual circumstances. Then a drone can be reeled in if the need arises. Another option is to have drones able to adhere to each other (magnets? velcro? carabiners?) so that functional lightly-laden drones can perform an in-air rescue of a malfunctioning drone before it reaches the ground. This assumes that a malfunctioning drone would still have some mechanism to slow its descent, which isn't unreasonable.
Laden drones which can drop down to their delivery point from above, and then return empty back to the air dock seem like they could operate much more quietly and efficiently than drones which need enough power to maintain altitude as they travel horizontally to their target.
Of course the drone would need the capability to throttle up and climb fully laden, possibly even minus one rotor, and the energy reserves to spare.
But the theory would be launching off the airship, and traveling horizontally (at altitude) above the target, then a controlled descent straight down to the drop point.
Presumably you would have two or three ships so one would be in position, while another was moving to/from the ground warehouse, while the 3rd was being loaded for the next batch.
The whole thing only makes sense if the energy density of the drone batteries doesn’t allow them to simply launch from a remote warehouse and proceed directly to their destination at altitude and return all on their own power.