Well if autonomous cars are supposed to have better reaction times than humans, it should be the manufacturers that are liable. ...then again, if cars are ever truly autonomous they should obviously take responsibility for themselves 🌈
Interesting comment from my brother on this news.
"Talking of software: an autonomous vehicle recorded the first fatality a few days ago. That's the first fatality in about 3 million miles of driving with all the test vehicles. That sounds pretty good until you realise the fatality rate for drivers overall in the U.S. is about 1.25 per 100 million miles. In order to prove driverless cars can meet this standard, it seems to me they need several hundred million miles of testing. For each software update."
The point he is making is that statistically speaking, real live drivers actually manage to have an accident free record that is much better than nearly all software.
Consider: yesterday ANZ had network wide computer issues, crippling all banking activities, guessing this was following a software update. Can you imagine the issues that could arise when a fleetwide software update is pushed to all autonomous cars? (You would expect bank software to be among the most carefully prepared software used anywhere!)
John
JDNSW
1986 110 County 3.9 diesel
1970 2a 109 2.25 petrol
Had been wondering what the stats really look like for the poor old much maligned human driver.... hadn't searched for that data thinking that there wouldn't yet be enough data for the other side... the ever so "much better, faster" computer driver side. Even now with one fatality in 3 million miles - it is too small a sample.
A stray thought passed through my mind tho - some time ago (maybe when ABS was still in its infancy) it used to be said that race trained drivers at their peak can brake in shorter distance than ABS equipped vehicles. I wonder how long ago that situation flipped?
Neil
(Really shouldn't be a...) Grumpy old fart!
MY2013 2.2l TDCi Dual Cab Ute
Nulla tenaci invia est via
I thought this point needed highlighting.
The pro autonomous vehicle mob always look at the absolute figure of road deaths and, as autonomous vehicles have killed less people in absolute terms, falsely conclude they are safer.
When the statistics are weighted, it shows autonomous vehicles are killers.
Hi,
1.25 per 100 million miles ? I'm not doubting it, just extrapolating that to how many miles must be driven per day.
Scary!
I guess the ratio must change if we go from a world of engineered highways to the chaos of third world cities though.
Perhaps the potential for most benefit will be in Asia, not U.S.A.
That would be a bit ironic.
Cheers
Cuts both ways... cannot say autonomous is better, cannot say its worse.... the sample size on the autonomous mob's side, one death in 3 million miles, is too small for now. (i.e. lets wait for the 100th death, look at the autonomous miles travelled, then see if autonomous vehicles are better at killing people than human drivers)
Neil
(Really shouldn't be a...) Grumpy old fart!
MY2013 2.2l TDCi Dual Cab Ute
Nulla tenaci invia est via
From the article:
AVs also have a hard time recognizing bicycles.I think we have a long way to go before autonomous vehicles are to be let loose on our streets. A very long way.If lighting was an actual issue, then darker-skinned people might want to be extra wary, given that the best technologies today have a hard time detecting their faces.
...and what about related technologies? How many times have robotic japanese love dolls gone round the block without injury to humans?
![]()
Neil
(Really shouldn't be a...) Grumpy old fart!
MY2013 2.2l TDCi Dual Cab Ute
Nulla tenaci invia est via
| Search AULRO.com ONLY! |
Search All the Web! |
|---|
|
|
|
Bookmarks