> that view is now in question
Does anyone else find it fascinating that for computers to be considered "safe" they must be as close to 100% as you can get, when the equivalent human drivers have fatal accidents every single day?
Does anyone else find it fascinating that for computers to be considered "safe" they must be as close to 100% as you can get, when the equivalent human drivers have fatal accidents every single day?
Statistically, autopilot is orders of magnitude safer than humans.
If there's a story here, it's about how the US doesn't require underrun protection on trucks, thus making them death traps, but that wouldn't be anywhere near as good for circulation as a "rise of the machines, they want to kill you" story.
I wouldn't be surprised if autopilot is safer, but I think it is overstating things to say this is a statistical fact.
Source? I hear that it is imperfect software.
> Karl Brauer, an analyst with the auto research firm Kelley Blue Book, said the accident served as a signal that the technology might not be as advanced and ready for the market as some proponents have suggested.
> "This is a bit of a wake-up call," Mr. Brauer said. "People who were maybe too aggressive in taking the position that we're almost there, this technology is going to be in the market very soon, maybe need to reassess that."
I think that's perfectly reasonable, e.g. it may take 5 or 10 years to reach some kind of approved minimum safety level, especially since such safety approval standards/processes are still being developed. Just because Tesla's cars already have an "auto-pilot mode" doesn't mean the technology's ready for widespread deployment.
Regarding "as close to 100% as you can get", that's not stated in the article. Instead, it says:
> The federal traffic safety agency is nearing the release of a new set of guidelines and regulations regarding the testing of self-driving vehicles on public roads. They are expected to be released in July.
> At a recent technology conference in Novi, Mich., the agency's leader, Mark Rosekind, said self-driving cars should at least be twice as safe as human drivers to result in a significant reduction in roadway deaths. "We need to start with two times better," Mr. Rosekind said. "We need to set a higher bar if we expect safety to actually be a benefit here."
Again, I think that's reasonable. While a fully-automated road system would probably be safer than human drivers, we're not going to get an overnight switch; the harder problem is the transition, with manually- and autonomously-driven cars coexisting.
It's not just a case of swapping out more and more cars, since self-driving cars will behave and interact in different ways (as an obvious example, looking at the occupants won't give you any information about intentions/gestures/etc.); hence it's worth waiting until a threshold higher than break-even is reached before beginning the transition.
I can't even imagine the resistance to self-driving cars once they're actually viable. Who even enjoys driving on city roads anymore? Or on highways? I drove on American freeways and it was mind-numbingly dull. Just acres of concrete all along the road with nothing interesting to see or do.
I drive motorcycle (and enjoy that), I doubt autopilot will come to those anytime soon..
In the future though, it will always be your fault, because you are one person trying to prove an accident was not your fault against a big-money company. Good luck.
That's why I always wear videoregistrator on my helmet. It paid off for itself many times already.
I would expect that it's quite easy to autopilot a tractor on the road. They already use autopilot in the field.
[1] https://www.precisionfarmingdealer.com/articles/2211-retrofi...
In Lithuania it's even worse, around 60% (2014 data).
[1] http://americanhistory.si.edu/collections/search/object/nmah...
Who even enjoys writing code any more? It's just line after line of characters on a screen, with nothing interesting to see or do.
Yet another american claim that "I don't like something, therefore no one else must like it either", with the usual view that because something is terrible in America, it must be terrible everywhere.
Driving has now become a choice between the dull, automated driving I did in America, or the mad, chaotic rush I see on Indian roads.
Neither of these are enjoyable or even particularly safe
Keep in mind, you're just below the world average (17) at that value. So while Im sure Indian roads are as bad as their reputation, there are worse.
I live in Thailand, on the same list (when sorted by deaths by 100K inhabitants), we're #2 (right behind Libya), with 36.2; the deaths per year for here was 24K, based on a population of just 65 million.
My point is, I know what it's like to drive on dangerous roads, and I still dispute your claim that driving is either boring or dangerous.
Even driving across the Eyre highway from Adelaide to Perth (~2700km), part of which is the '90 mile straight' - a section of road with 0 bends for 145km - is something I can look back and be glad I've done it.
If you personally don't enjoy driving, or feel unsafe driving, you personally can take actions based on that, such as taking a bus or taxi, or car pooling etc.
But don't somehow think that because your homeland has a reputation for bad roads, and America has apparently boring roads, that no one else on the planet enjoys driving.
So if you see a Tesla car crashing on days with white sky, you want to know whether this can happens to you as well.
Computers have to be 100% safe because self-driving cars take control away from the user. And where does the blame lie in an accident? With the owner of the vehicle, the 'operator' at the time of the accident, the 3rd party developers who were contracted by Tesla/Nissan/BMW, or the car company themselves? Getting customer service and a dealership to admit responsibility for parts failure as it is today is near impossible.
It's fine for people to be be pro self-driving cars but it is wrong to dismiss any and every criticism of an emerging technology on that basis alone.