> the car "predicted that [the bus] would yield to us because we were ahead of it."
> the bus was only going 15 miles per hour, while the SUV was moving at just 2 mph.
Watching the video, it looks like Google's plan was to jump out in front of the bus and cut it off. Even had it not resulted in a crash... are they teaching Google to be a bad driver?
>Watching the video, it looks like Google's plan was to jump out in front of the bus and cut it off
It doesn't look like it even did that. It's just barely moving, and it's debatable whether the bus driver even saw it actually moving. If I was the driver behind, I would assume the the SUV was just signalling to move out, but was going to wait for me to pass. If I was passing in a car then obviously I would take evasive action, but the bus driver likely never even noticed.
It looks like google's car logic is slightly flawed. The car should either make a decisive move, or stay put. Creeping into a lane of faster moving traffic at 2mph is a somewhat ridiculous plan of action.
My understanding is that the bus 100% had the right of way and wasn't making any moves that would indicate he was going to let the car in. I don't see how the car (and it's driver) assumed they could cut in.
I bike every where, and it's pretty scary to think of what other "assumptions" the Google cars may be making. I doubt the bus got more than a scratch, but a biker or motorcyclist in that situation could easily have deflected into other traffic and been killed.
I'm super pro-autonomous vehicles. I believe they will save far more lives than harm. But WTF, its like the rules engine had an incorrect entry. When you're leaving the curb lane (parked) you basically have no rights. Just wait for it to be clear. Maybe Google is starting to teach it more than the rules, but also the "rules of thumb" like if you gun it in front of a car, they'll brake a little.
Okay, you can be super pro autonomous but the problem is not technical.
Who is liable in case of accident (financially and legally)?
The car (no legal existence), the algorithm (no legal existence) a developer doing a library that fails and has been used without is knowledge (nop), the developer integrating without control (nop), the scrum master who pushed a bad choice? the financial director for not caring about the cost of risks? The CTO .... no one is liable. Liability is diluted.
Who will pay when people will die?
And given VW dieselgate, RSA weakening of crypto, the actual world does not seems to support software engineers caring about their users or regulations.
I see no reason why liability doesnt follow just like (afaik) all tech does.
When I use my computer, if it malfunctions and hurts someone, its the manufacturers fault (think DELL w/ bad batteries), if an engineer was deeply negligent they may take on liability, but generally the company. If I use my computer for bad I am liable. So on and so forth to wherever the source of the fault is.
People like my mother are losing an astounding amount of work because of poorly designed interfaces.
My mother that still has an economical activity is losing money with modern technologies. And no one is liable.
It is is the "user's fault". The keyboard chair bug.
If also so many people get fucked by the oldest trick on earth of executing arbitrary software by trusting an unknown source it is clearly an information problem.
Either the industry is unable to make the information clear, or the information presented is not adapted to human beings.
In both cases, given the number of victims and money lost do you really think that this industry considers itself liable? Users are treated guilty by default in the mind of the companies.
The PR of IT is that all the bugs from software is the fault of the users. It is denial.
We (software industry) are the root cause some people lose money and no one can sue us for this.
Because we impress them with our confidence. And they believe us.
There's no way I would have pulled out in front of a bus (or even a car or cyclist) like this; there was barely a gap between the bus and the car in front of the bus, into which the Google car poked its nose. It's probably just an edge case, but I agree with the other post that the parked car has no rights, and no expectation anyone in lane will yield.
8 comments
[ 4.3 ms ] story [ 31.0 ms ] thread> the bus was only going 15 miles per hour, while the SUV was moving at just 2 mph.
Watching the video, it looks like Google's plan was to jump out in front of the bus and cut it off. Even had it not resulted in a crash... are they teaching Google to be a bad driver?
It doesn't look like it even did that. It's just barely moving, and it's debatable whether the bus driver even saw it actually moving. If I was the driver behind, I would assume the the SUV was just signalling to move out, but was going to wait for me to pass. If I was passing in a car then obviously I would take evasive action, but the bus driver likely never even noticed.
It looks like google's car logic is slightly flawed. The car should either make a decisive move, or stay put. Creeping into a lane of faster moving traffic at 2mph is a somewhat ridiculous plan of action.
My understanding is that the bus 100% had the right of way and wasn't making any moves that would indicate he was going to let the car in. I don't see how the car (and it's driver) assumed they could cut in.
I bike every where, and it's pretty scary to think of what other "assumptions" the Google cars may be making. I doubt the bus got more than a scratch, but a biker or motorcyclist in that situation could easily have deflected into other traffic and been killed.
Who is liable in case of accident (financially and legally)?
The car (no legal existence), the algorithm (no legal existence) a developer doing a library that fails and has been used without is knowledge (nop), the developer integrating without control (nop), the scrum master who pushed a bad choice? the financial director for not caring about the cost of risks? The CTO .... no one is liable. Liability is diluted.
Who will pay when people will die?
And given VW dieselgate, RSA weakening of crypto, the actual world does not seems to support software engineers caring about their users or regulations.
When I use my computer, if it malfunctions and hurts someone, its the manufacturers fault (think DELL w/ bad batteries), if an engineer was deeply negligent they may take on liability, but generally the company. If I use my computer for bad I am liable. So on and so forth to wherever the source of the fault is.
My mother that still has an economical activity is losing money with modern technologies. And no one is liable.
It is is the "user's fault". The keyboard chair bug.
If also so many people get fucked by the oldest trick on earth of executing arbitrary software by trusting an unknown source it is clearly an information problem.
Either the industry is unable to make the information clear, or the information presented is not adapted to human beings.
In both cases, given the number of victims and money lost do you really think that this industry considers itself liable? Users are treated guilty by default in the mind of the companies.
The PR of IT is that all the bugs from software is the fault of the users. It is denial.
We (software industry) are the root cause some people lose money and no one can sue us for this.
Because we impress them with our confidence. And they believe us.