Uber’s first self-driving cars will start picking up passengers this month

[pcslookout]

https://techcrunch.com/2016/08/18/u...-will-start-picking-up-passengers-this-month/

Can't wait! Hope this catches on!

 

[MrSquished]

hmmmm. maybe because there is an engineer and co-pilot at the ready to take the wheel I'd try this but no way I am getting inside a totally driverless car anytime soon

 

[pcslookout]

hmmmm. maybe because there is an engineer and co-pilot at the ready to take the wheel I'd try this but no way I am getting inside a totally driverless car anytime soon

Yeah.

 

[smackababy]

As long as they aren't like the terrible, scammy Uber drivers around here, I'm all for it. If you do Uber pool, most drivers will sit and wait for 10+ minutes before coming to get you in hopes you cancel and have to pay them the $2-5 cancellation fee. It's total BS.

 

[pcslookout]

As long as they aren't like the terrible, scammy Uber drivers around here, I'm all for it. If you do Uber pool, most drivers will sit and wait for 10+ minutes before coming to get you in hopes you cancel and have to pay them the $2-5 cancellation fee. It's total BS.

What city is this ?

 

[smackababy]

What city is this ?

Chicago.

 

[JSt0rm]

Can't wait to buy a self driving car once all the kinks are ironed out.

 

[bradly1101]

If all the cars can talk to each other, I think it could work. But what if a deer jumps in front of the car and there's a car next to you and no time to stop? Will the driverless algorithm choose to hit the deer and risk it killing you by coming through the windshield, or will it swerve into the car next to you hoping for a minor accident? What if it was a human and not a deer? What would the car do?

We need ethical algorithms but who decides? Maybe legislation...

 

[smackababy]

If all the cars can talk to each other, I think it could work. But what if a deer jumps in front of the car and there's a car next to you and no time to stop? Will the driverless algorithm choose to hit the deer and risk it killing you by coming through the windshield, or will it swerve into the car next to you hoping for a minor accident? What if it was a human and not a deer? What would the car do?

We need ethical algorithms but who decides? Maybe legislation...

See, here is the fallacy in that argument: a car with sensors can see a deer or a human as a potential "threat" long before humans notice and AND can react long before humans can (if they notice it). This fixation of "oh, the robot will kill a person!" is stupid, especially considering a person in the exact same scenario will do just that.

The vast majority of accidents are likely didn't / couldn't see it, wasn't paying enough attention, or just bad judgement on one or more drivers. Those will be far, far reduced.

 

[Carson Dyle]

Can't wait to buy a self driving car once all the kinks are ironed out.

How far away are we from truly driverless vehicles, though?

It's one thing when a bug causes your OS to crash and you lose the changes you're making to a spreadsheet, quite another when a bug causes the vehicle in which you're a passenger to crash into the side of an emergency vehicle.

And it's totally a different thing to use the technology while you're behind the wheel and watching to keep your car from decapitating you vs. relying on it exclusively.

I have a feeling that even once the technology is "there" (which means good enough to only kill someone occasionally) it's going to have a difficult time being approved for use, and could take many years to get that approval.

 

[Red Squirrel]

I was not aware the tech was actually this far out to the point where you don't actually need a driver. Even Google does not have fully self driving cars yet as far as I know and they were the firsts to work on it.

 

[Carson Dyle]

I was not aware the tech was actually this far out to the point where you don't actually need a driver. Even Google does not have fully self driving cars yet as far as I know and they were the firsts to work on it.

They are, and they aren't.

The cars are fully self-driving, but they need someone to babysit them. Which is why these Uber vehicles will have an "engineer" behind the wheel. In essence, what is being reported is an experiment and no more autonomous than google's vehicles.

the company is preparing to add self-driving cars to its fleet of active drivers in Pittsburgh as soon as this month.

The company will deploy around 100 modified Volvo XC90s outfitted with self-driving equipment. Each vehicle will be staffed by one engineer, who can take the wheel as/when needed, and a co-pilot to observe and take notes. There will also be a “liquid-cooled” computer sitting in the trunk recording trip and map data.

 

[bradly1101]

See, here is the fallacy in that argument: a car with sensors can see a deer or a human as a potential "threat" long before humans notice and AND can react long before humans can (if they notice it). This fixation of "oh, the robot will kill a person!" is stupid, especially considering a person in the exact same scenario will do just that.

The vast majority of accidents are likely didn't / couldn't see it, wasn't paying enough attention, or just bad judgement on one or more drivers. Those will be far, far reduced.

In the city I live in there are pedestrians and bikes that will come off a sidewalk really fast. I witnessed one accident where a guy just ran into traffic. I wondered if he was suicidal.

I just wonder what this interim time will be like until there are fully driverless cars. We've already seen one death (of course relying on beta driverless software seems foolhardy), and other accidents. Driverless cars know what they're doing, humans do not. So until we're all driverless (or at least everything else that could cross the road is geotagged along with all the non-driverless cars), I'm skeptical.

 

[MongGrel]

I've seen enough videos of self driving cars screwing up I still would not trust one as a taxi.

 

[Ns1]

I've seen enough videos of self driving cars screwing up I still would not trust one as a taxi.

Can't be any worse than human drivers.

 

[Ns1]

As long as they aren't like the terrible, scammy Uber drivers around here, I'm all for it. If you do Uber pool, most drivers will sit and wait for 10+ minutes before coming to get you in hopes you cancel and have to pay them the $2-5 cancellation fee. It's total BS.

They're working for seventy cents a mile less costs, so they gotta do what they gotta do.

 

[Carson Dyle]

Can't be any worse than human drivers.

Until the car you're riding in with no driver drives you under a tractor-trailer and leaves your head 100 yards behind the rest of your body.

 

[Ns1]

Until the car you're riding in with no driver drives you under a tractor-trailer and leaves your head 100 yards behind the rest of your body.

seems pretty mild compared to what humans behind the wheel are capable of.

 

[Red Squirrel]

Oh did not realize they'd still have a driver. But yeah as fully autonomous I'd still like to see how well they fare in a typical winter condition, which is the majority of the year in some places. There are no visible street lines, the side of the street is hard to detect using anything like radar, due to the irregular shapes (the snow banks). Though I imagine they could also be GPS assisted but anyone who has played pokemon go knows you probably would not want to trust that for being able to tell where on the road you are.

 

[Imp]

So...um... How much do these things cost and how much could one get for the parts?

I expect numerous Uber riders will need pick-ups by container A1739 at the port.

 

[pcslookout]

Uber is worth $62.5 billion they can afford this.

Supposedly the rides are free for this. Right now.

 

[s0me0nesmind1]

How far away are we from truly driverless vehicles, though?

It's one thing when a bug causes your OS to crash and you lose the changes you're making to a spreadsheet, quite another when a bug causes the vehicle in which you're a passenger to crash into the side of an emergency vehicle.

And it's totally a different thing to use the technology while you're behind the wheel and watching to keep your car from decapitating you vs. relying on it exclusively.

I have a feeling that even once the technology is "there" (which means good enough to only kill someone occasionally) it's going to have a difficult time being approved for use, and could take many years to get that approval.

I will take 50 deaths by automated cars any day over the ~32,000 per year that we have now with non-automated. Statistics don't lie, and if you guys would learn to think for yourself instead of what the media wants you to think - you would realize that you are statistically better off.

 

[kranky]

if there's a human behind the wheel (i.e. computer-assisted driving), I'm fine with that. But not if it's 100% computer controlled with no human.
I used to work by an area with a few hospitals and it was common to have ambulances coming with sirens blaring. If the robocar has a green light at an intersection, how does it know it should stop anyway because an ambulance is coming down the cross street?

What if the robocar has a red light and the ambulance is behind it? When that happens, we just go through the light if cross traffic is clear so the ambulance can get going.

 

[pmv]

See, here is the fallacy in that argument: a car with sensors can see a deer or a human as a potential "threat" long before humans notice and AND can react long before humans can (if they notice it). This fixation of "oh, the robot will kill a person!" is stupid, especially considering a person in the exact same scenario will do just that.

The vast majority of accidents are likely didn't / couldn't see it, wasn't paying enough attention, or just bad judgement on one or more drivers. Those will be far, far reduced.

But the argument is still valid. Its not whether such things will more more-or-less common but self-driving cars will change the legal context of such events - the machine will have to make decisions about whose safety to prioritise. At the moment that happens in a way that isn't explicitly defined before the event. Now it will have to be explicitly decided and encoded somewhere.

Also, as there will be a competitive market for such cars, maybe there will be market pressure on manufacturers to prioritise the safety (and convenience) of the car occupants over that of anyone outside? "Your car keeps stopping for kids running out into the road - its slowing me down! I'm gonna buy that other, more aggressively-programmed, model!" "I'm not risk-averse, I want a car that will overtake aggressively, even if there's something coming the other way!".
Maybe there will even be unofficial 'hacks' available for those who find their car drives too damn cautiously and safely for their liking?

Of course that happens now with human drivers, but it will just all happen in a different, and more explicit, way, so presumably the legal arguments will be different. There will presumably be something like those old computer games where everyone programs their virtual-robot to show varying degrees of aggression and different tactics to fight each other.

 

[pmv]

I will take 50 deaths by automated cars any day over the ~32,000 per year that we have now with non-automated. Statistics don't lie, and if you guys would learn to think for yourself instead of what the media wants you to think - you would realize that you are statistically better off.

Yeah, but if everyone then drives everwhere, because its much easier and you don't even need to know how to do it...
(a) How long before everyone not driving gets banned from the roads/streets in order to make those roads/streets more predictable an environment for the robo-cars?
(b) What's going to happen to the level of obesity and inactivity-related disease?