I tend to agree with the very last comment, ‘But it’s never going to get off the ground’ Well, certainly not in the foreseeable future.
I really do think they underestimated the complexities involved in a machine AI learning all the risks, hazards and anomalies that we as humans take for granted and act on daily to avoid accidents. Just think of all that would be needed to be visually analysed, recognised and avoided to enable that 27kg piece of kit to get from a distribution point to the front lawn of Mrs Pembridge at 16 Crescent Gardens, Worksop. And think of the consequences when, not if, that 27kg drone and its payload has a wobbler and crashes onto the M1 North just outside of Birmingham.
It’s the same with autonomous cars. Can’t see it in my lifetime.
Some wobbler that, the M1 goes no where near Brum ;o)
Strong wind from 1000ft 
It always struck me as such a stupid and pointless idea.
… brutal 
In my defence I’m a Southerner and very rarely venture north of Watford 
Meanwhile, in the USA:
Trials have also been made in the UK with shortish BVLOS flights over the Solent using fixed wing UAS
If tilt wing or hybrid aircraft were used - something like this but built for payload rather than endurance:
https://www.foxtechfpv.com/foxtech-ayk-250-vtol.html then deliveries of urgent samples and drugs, perhaps even human organs between hospitals and labs could become possible although in this country we are always at the mercy of the weather.
They are planning a route over Morecambe Bay between Barrow and Lancaster