This raises an interesting dilemma. Here in NSW a person reaching 80 is required to undergo a driving test to ensure they are capable. If they own a self-driving vehicle would that enable them to pass the driving test?

Trials of 'autonomous vehicles' on European roads, particularly in Germany, Scandinavian countries and elsewhere, have demonstrated the effectiveness of this mode of transport going forward .  This includes a detailed model for the related infrastructure requirements for its rollout across cities and towns in these regions. 

One of the major findings from these studies/trials is significant traffic throughput improvements have clearly demonstrated infrastructure changes for the effective use of this technology (autonomous vehicles, IoT, etc) will be 'minimal' (the studies' word, not mine).  In this regard, they often refer to the following scenarios:

a. Currently a driver, using their eyes, can generally see what's happening 2 or 3 cars ahead in medium to high volume traffic conditions.  Using on-road tags/monitoring devices/etc., your autonomous vehicle can 'see' 50 vehicles ahead (or more depending on the power of the inroad devices).  FYI, such devices are all over current roads sensing traffic movement to control signaling, speed signage, traffic warnings, etc., etc.  And today, costs of such devices are dollars, rather than hundreds of dollars. For instance, locally, there are 000's of such devices all along the M4 motorway in Sydney, not to mention current local traffic roads across cities in Australia.

b.  If your autonomous vehicle can 'see' 50 vehicles ahead of you, it's (not yours) reaction time will be faster than you when required.  Mathematical equations have been formulated for such an environment which demonstrate autonomous vehicles can therefore travel closer together (vehicle separation), thus delivering a greater throughput on busy roads.  Applying such equations to traffic volumes during the morning peak hour on the Sydney Harbour Bridge [SHB] (i.e., pre-Covid volumes) roughly indicates that the same volume of vehicles between 7:30am and 9:30am (approx. weekday peak) could be accommodated on just under 6 lanes, instead of the current 8 lanes.  Thus you could:

- increase the volume of vehicular traffic, or

- repurpose 2 lanes for other vehicles (e.g., multi-passenger vehicle like buses or light rail in BOTH directions, goods only vehicle lanes,  bicycle lanes [current bicycle lane is becoming crowded and remains a little difficult to access]).

It raises the question, why are we building tunnels to accommodate vehicles crossing Sydney Harbour when there is still significant existing infrastructure that can be READILY repurposed for TODAY's environment, which would be billions of dollars cheaper?

The BIG question is that Tesla themselves have stipulated that for safety the "driver" must have their hands on the steering wheel and be alert to what is happening around them. If the car is driving itself it is just a matter of time before the "driver's" attention will be somewhere else and will be unable to react in time for an emergency. It happens now while they are, supposedly, wide awake. 

Given the accidents with aircraft with two highly trained pilots who are wide awake and using "automatic pilots" who get into trouble trying to sort out what is happening with the computer controlling the flight of the aircraft it really raises questions on the actual self driving cars.

I just hope Microsoft have nothing to do with the software installed in these vehicles!

Never! I don't even trust other drivers' ability. Why would I trust a robot?

No, I don't use commuter trains. Last train I travelled in was a steam one while on holiday in the Scottish Highlands.

NO WAY!