The best IT systems have an expected uptime a helluva lot higher than 99% though - the expected downtime is orders of magnitude lower than 1%.
That said, I agree that there will be accidents and part of the issue will be when people die as a result there will be a big backlash against the technology, perhaps disproportionate relative to the scale of the problem compared to other killers.
I've said it before, if human driven cars were a new thing introduced now they will likely be subject to much more stringent controls / legislation based on the number of deaths. New tech that results in injuries and deaths gets clamped down on hard - can you imagine the media storm if any technology, never mind cars, was involved directly or indirectly in over a million deaths a year? - but incumbent technology not so much, it is just accepted matter-of-fact. Similar scenario with alcohol, tobacco etc - if they were new products people would baulk at the health implications.
Overall although I can't wait for safe, reliable, affordable AV I am sceptical about the timelines discussed. Sadly, I think I will have retired before I can have one. I think there are too many variables to program for with current and near-future technology; even with adaptive AI I would expect cars to encounter unexpected scenarios and not know how to respond (which happens with humans too, but again, 'human error' is just accepted as a fact of life by people whereas 'system error' is not).
In some ways I'd be more confident about the effectiveness of autonomous flying vehicles because the terrain is much more predictable, you have some challenges around coping with failures I suppose (a car is already on the ground if if it fails so easier to get out without injury etc, compared to something falling out of the sky) but basically you just have to worry about other air traffic, perhaps compensate for weather conditions more but at least you have no pedestrians to worry about, and more flexibility to route around problems rather than being constrained by roads.
Pretty good summary.
Most of the doubters in regards AV seem to start by saying its not going to happen, but sooner or later put a I don't like how its going to change something I like.
I honestly think we will see some mass limited area systems quite quickly. Completely open, ie travel anywhere systems, as we all agree there are some significant issues in that regard.
It would in fact be a hell of a lot easier if we were starting from scratch, we would have a better road layout, better constructed roads, and build in technology that helps the AV rather than making it harder.
However eventually I am fairly sure we will look back on the days when we functioned as we do today and laugh. When I say we, I suspect most frequenting this forum will be long dead.
The thing that doubles or more the challenge of AV is the having to coincide with humans driving around, normally badly.
This morning on my drive in I saw a car on its side on a bend. It was icy, its always icy there. Driver error. It can't be anything else but failure to drive to the conditions.
This is the sort of error that should be dialled out by AV/AI, not saying there may not be some new ones, but as humans we still haven't adapted to stop making the same errors we always have, at least as AV/AI evolves it should get better every generation.
IMHO more than anything we should push some of the tech now that helps improve the safety and is precursor required tech for AV, so lane following, safe distance tech etc
Same journey this morning, once I joined the more main road i use which is single carriageway. I caught a chain of about 15 cars, guy at the back driving a newish volvo was constatntly braking when the car in front wasn't, it was icy in patches and someone was very cautious at the front, doing 45-55 so clearly some of the drivers were struggling to cope in this scenario, being the sort who think if the limit is 60 thats somehow the target. Hes exactly the sort of driver who probably classifies himself as good, or better than average, yet hes failing in the fundamental of keeping a reasonable distance to the car in front. If he had safe distance tech active he would have been further back, and wouldn't have needed to keep braking (when all the other cars in front were not).
