Google's self-driving car crash is a positive sign
ideas.4brad.comBeing from India, Let me say that it will take huge advancements in AI before self-driving cars could drive on the streets here!
These advances would need to include :-
-Being able to decide what can be considered a road and what cannot.
-Knowledge of Human psychology which additionally needs to cover areas of suicide , homicide, etc.
-Knowledge of Animal psychology
-Understanding of different cultural norms both for humans and animals (I have found that the cows in Bangalore are much less likely to yield way than the cows in Kerala)
Comment of the year!
> A lot of people I talk to imagine that the tech problems have all been solved and all that’s left is getting legal and public acceptance. There is great progress being made, but nobody should expect these cars to be perfect today. That’s why they run with safety drivers, and did even before the law demanded it. This time the safety driver also decided the bus would yield and so let the car try its merge. But expect more of this as time goes forward.
I think most people who say the issues are social and legal don't think the technical problems are solved, they just think that we are inexorably progressing towards solving them at a more rapid rate than the social and legal problems.
1) i thought this meant stock crash, but it refers to the bus / automated SUV crash from 2/29/2016
2) Has anyone made an animated depiction of the crash, I'm curious if I have in my head the right idea about what actually happened
Not animated, but there's a diagram here: https://news.ycombinator.com/item?id=11197681
Me too... I suggest to change the title, as it seems misleading.
> Has anyone made an animated depiction of the crash
Was this a test-vehicle of Google? In that case, it probably had lots of cameras on board.
Yes, it was an autonomous vehicle (with a "test driver" in the car)…there are more details on http://www.engadget.com/2016/02/29/google-self-driving-car-a...
What I find interesting is this comment: Our test driver, who had been watching the bus in the mirror, also expected the bus to slow or stop.
And Google is not going to release that footage, at least not anytime soon.
The article mentions that the robocar can theoretically see other drivers intent: <i>It’s worth noting that sometimes humans solve this problem by making eye contact, to know if the other car has seen you. Turns out that robots can do that as well, because the human eye flashes brightly in the red and infrared when looking directly at you — the “red eye” effect of small flash cameras.</i> But what about the other way? Perhaps the bus driver couldn't see Goog's flashing eyes, and assumed therefore it wouldn't attempt to merge back in.
I'm wondering why there were enough sandbags around a storm drain to warrant the vehicle moving back into the left lane and not any indication before hand. The autonomous vehicle would know when a lane is blocked because of cones, right?
My understanding is that the car didn't change lanes. It was in the right lane the whole time,just hugging the curb. A bus (in the same lane) tried to pass, and hit the Google car when it got more centered in the right lane.
I'm glad that nobody was hurt. It takes some of the emotion about getting to the fixes.
And it's still hard to argue with the fact that out of several million miles driven, this is the first serious accident that was caused by the autonomous car (and the failure scenario looks to have been appreciably complex). You won't find a human with that clean of a record.
You won't find a human with that clean of a record.
Some number of us humans are better than that. In addition to my father who's in his early '80s who's lived almost all his life in Missouri, where driving instead of flying to other places is generally ideal, try this search: https://www.google.com/search?q=truck+driver+million+mile+cl...
Humans only drive about 500K miles in a lifetime, though there are those who drive much more. 2 million miles would be a challenge in any human lifetime, but there would be some professional drivers which attain that.
I'm very bearish on the future of human drivers. Human drivers sleep and wheel and drink and drive much more frequently than computers crash. But the computers will have glitches in edge cases that humans are better equipped to serve. It's part of the trade-off. It we can cut even half of the 30,000+ auto fatalities a year, it's worth it.