Hmm. I get what you are saying, but I'm not down with your analysis. There will be a certain amount of energy transferred by the impact. The injurious portion of that impact is during the impulse strike of the two cars. Crumple zones, for instance, work by spreading the impulse experienced by the cabin across a larger time. If some of that same energy goes into accelerating the car and then being dissipated later by braking all that is doing is spreading the total energy dissipation across a larger time window. It should lower the impulse transferred to the cabin by using some of the energy to accelerate the rest of the frame as well.yes, you get slightly more acceleration due to the whole car accelerating, but less total impulse due to the partitioning of the energy transfer.

Not that I think there's any significant difference, since in any serious rear end collision the total acceleration will vastly overwhelm the strength of your legs to press the brake pedal.

Well, the reported accidents at least. Show me the numbers on how many people get into little fender benders that never get reported so we can compare some real data.

No way. It's supposed to do that for every car within 300 feet all the time on the off chance one of them is about to cause a problem. Think of the children! Google pinkie-swears they won't do anything bad with the new datastream this represents. Also, there's a Starbuck's we think you'd like on your route to work...

"So, you are saying that you specifically programmed it to behave in a way that increased the danger to my client?"

These kinds of things are inevitable. There were some things the car might have tried to do to evade the oncoming collision, but they all have some additional risks, plus they just move the potential accident closer to the intersection. Sometimes the lowest risk thing you can do is get squarely hit from behind.

Much like the original article.

And what happens if we tell them to behave randomly? A particle filter, for instance, uses randomness to generate a set of states for evaluation. Sensor fusion takes large numbers of highly error prone sensor readings and merges them into state estimates. Both methods introduce uncertainty into state estimation and, therefore, present non-deterministic foundations for reasoning. Even if the reasoning processes are strictly deterministic, you can still get non-deterministic behavior, and that's without even introducing any explicit behavioral randomness.

But, let's be honest: No one has ever provided a definition (that I've ever seen) for sentience which precludes deterministic response. Are you proposing irrationality as a fundamental identifier of intelligence? I'm not sure I'd call it a feature, but maybe it's an inevitable consequence.

I don't think there is an extremely large piece. I think there are hundreds of thousands of little pieces. Also, we create things all the time without understanding them. I mean most people don't have any idea how mitosis works and yet we don't have much problem reproducing. Anyway, the point is that I don't think we need a full understanding of human intelligence to create some kind of intelligent agent.

I own a car, but I've always hated driving. I value the freedom of having the ability to go anywhere, whenever I want. I also value my free time. I would trade the joy of attending to my 45 minute commute in the morning for the next episode/chapter of whatever in an instant. The desire to drive isn't a desire I share. I share the desire to get somewhere, which an autonomous vehicle will, pretty soon, achieve.

Actually, self driving cars will always have to deal with unpredictable behavior from neighboring vehicles. This will never change. They can react faster to other's wrong behaviors, evaluate and plan faster than human drivers, have greater knowledge of road conditions and environment, plus they don't get bored or inattentive. While they aren't up to the task yet, they are coming fast, because they have value for every driver who isn't just driving to drive.

I agree pervasiveness will be more than 5 years, but you'll be able to buy any car with an autonomous driver option in less than 20 years. In 5 years the first autonomous work vehicles will be on the road (specifically long haul trucking). Shortly after that someone will implement the autonomous traffic cop (red light camera on wheels), and traffic cops will be on their way out.

Don't really wanna make it tough
I just wanna tell you that I had enough
Might sound crazy,
But it ain't no lie,
Bye, bye, bye

Well, he clearly owns the copyright on the photographs, so if anyone wants to contest that they are SOL. The privacy concern is legitimate if and only if the pictures were taken in an area where there was an expectation of privacy. A sporting event with people in the stands cheering certainly doesn't seem like a private event...

I don't know...This seems like some kind of alternate reality setup for the Venom creation story. *I'm* not wearing anything woven from this stuff is all I'm saying.

I always suspected that automakers were amateurs. Real engineers use CMake.

Why is this tagged Cod Equality? I mean, I'm all for fish rights, but that's just sil...heh.

All I'm wondering is whether there has ever been a single quote in the codename before? Virtually guaranteed to break someone's build system...

Out of curiosity, how many crew members do you think died in the first 100 or so trans-Atlantic sailing boats? I'd wager it was a hell of a lot more than 1 in 60.

I wish more people understood this. There are plenty of ways during a support contact to show me you are an exception to the norm, but telling me you are the exception just makes you less exceptional.

AKA: Show me don't tell me.