61–70 of 288 posts
I'll believe this when there's independent evidence. Also, I want to see evidence to see data for all pilot AI driving products: traffic collisions, injuries per collision, arrive sooner or later than humans, and pedestrian injuries rate.
Fox claims it's a better henhouse guard than humans. But of course. Their insurance company isn't an unbiased source. I'll believe this when there's independent evidence. Also, I want to see evidence to see data for all pilot AI driving products: traffic collisions, injuries per collision, arrive sooner or later than humans, and pedestrian injuries rate.
Thing is the structure of the data for traffic accidents is, that we have a lot of data were we don't care much and very little about the accidents we care about a lot. We care more about fatalities than severe injuries, more about severe injuries than about light injuries and more about bodily harm than about fender benders. However, the number of events is precisely the opposite, which makes it easy to just swamp the signal we care about with events we don't care about as much.
> For example, Waymo’s driverless cars have traveled roughly 4 million miles, but traffic-related fatalities are routinely measured per 100 million miles. Across America's roadways last year, there were 1.37 fatalities for every 100 million miles traveled, totaling 42,795 lives lost in vehicle traffic crashes
That is, the expected number of fatalities at this point of data taking assuming Waymo's self driving cars are comparable to humans is 0. If we assume that rate is ten times the human rate, then the expected number of fatalities observed is still 0. If we assume that they kill people at a 100 times the rate of human drivers, then there is still a decent chance, something slightly worse than 50/50, that we didn't observe a fatality yet.
> “It’s interesting that we were able to show statistical significance with so little data,” Victor said. “That has to do with two things: one is that we have very robust baselines from Swiss Re, and the other is that we have incredible safety performance.”
Now this claim is probably true, in the sense of conflating all accidents with accidents we care about. And actually I think that there is a type of accident, low speed collisions when one tries to squeeze past a badly parked car, that is probably a very large part of all insurance claims where computers just should be really good. Computers have a engineered very precise world model and distance sensors all around and don't rely on human intuition guided by human impatience to just hope that squeezing past that will just fit.
Aren't they operating under a lot of constraints? The comparison should be with human drivers operating under similar conditions and on similar roads, similar cars(eg. new ones), weather conditions etc. Plus they have a fleet of support folks and standby remote and dispatchable operators which cannot be cheap, we get no insight into the financials except that they're losing several billion dollars a year. If at the e…
It’s a comparison between real-world transportation systems, not a sporting event. The systems are different, and that includes operating under different safety rules. It’s true that they haven’t reached sustainability yet.
...which have very different comparable characteristics.
Earlier quoted context omitted.
I believe for the insurance purposes, this difference is irrelevant. The point is under the current condition, Waymo driverless cars are safer than human controlled cars, thus the question of liability and permit to operate should be cleared. The question whether humans are truly inferior to robots in term of controlling cars is rather a question of ego than a practical one.
If you had insurance for a human whose vehicle could only operate between 9pm to 6am, would it still be more expensive than Waymo's?
Aren't they operating under a lot of constraints? The comparison should be with human drivers operating under similar conditions and on similar roads, similar cars(eg. new ones), weather conditions etc. Plus they have a fleet of support folks and standby remote and dispatchable operators which cannot be cheap, we get no insight into the financials except that they're losing several billion dollars a year. If at the e…
That said, actuaries aren’t stupid. They know whether it’s better to have an autonomous vehicle driving in the idealized subset of conditions that Waymo operates in or whether humans are better.
Earlier quoted context omitted.
Why did you type three paragraphs of nothing? What is the 'correct' way to frame this question?
The correct way is "who goes to jail if a driverless car kills a child?"
If self driving cars reduce the number of fatalities, and specially if they reduce them significantly, would you really be against them just because there won't be an easy target to put in jail if an accident happens?
Earlier quoted context omitted.
Great to hear someone is still pursuing the Uber model long after Uber themselves gave up on it!
I thought the problem with Uber’s self-driving initiative was that they didn’t know what they were doing, except for the main guy who stole the IP from Google.