Woman killed by self-driving Uber car

[SumJongGuy]

That's one take.... probably not what it, "boils down to," though.....

Frequency is a HUGE data point, but so are:
Liability (owner of car, software developer, infrastructure owner if tied into a system)
How do systems talk? A lot of this tech will involve group and bulk data

It'll all be interesting.

Well, don't those all count as "problems"? Will the new system cause more or less "problems" than the current human pilot system does??

Liability was addressed above with the proposal to grant blanket immunity to the industry same as the gun industry enjoys.. With the current foxes in the government hen house that is probably a slam dunk..

[fmac]

One thing that’s interesting about this is that there was a human ‘override’ driver present. I wonder if the controls were overtaken at any point.

If the car was still automated, it brings up the big, ethical question of this tech: should the car hit a pedestrian or should it sway and endanger the lives of the passengers?

[iceman]

You should give the bus a dry run one day. Get a haircut, go to the grocery store, and dinner at a restaurant, all on a bus. Throw in a dentist appointment for good measure.

That's probably a better argument for cities than it is for autonomous cars. Walk. It's better for you anyways.

[iceman]

I thought I fit the demographic, but I guess I am not the typical TGR poster.

Right. We drive the buses, we don't take them.

[grskier]

I don't see how they overcome basic driving conditions encountered by millions every day. The biggest of which is the weather. Rain or snow, even a dusting, and these things aren't going anywhere. You're little scenario doesn't play out during pow days, aka the time you would do such a trip.

At least not without completely ripping up every road in the country and replacing it.

Not necessarily the point with most of it.

And just because you don't solve 100% of the issues 100% of the time does it mean that it's a deal breaker.

[Benny Profane]

You should give the bus a dry run one day. Get a haircut, go to the grocery store, and dinner at a restaurant, all on a bus. Throw in a dentist appointment for good measure.

Theres billions of people living in cities in America and Europe and elsewhere that do it every day. Groceries are tough, but, you have them delivered. In a pinch, call a cab. You know, one of those yellow cars that you can just wave down, or call on your phone or an app. The yellow cars have licensed drivers, not the black ones you hail on an app. Again, much cheaper, efficient, and environmentally friendly than a bunch of robot cars zooming around. Proven technology. Try it. It works. But, wheres the profit in all that?

[Timberridge]

Right. We drive the buses, we don't take them.

Uh oh. You'll gonna awaken...

[Timberridge]

Proven technology. Try it. It works.

Like the corner payphone?

[Benny Profane]

Like the corner payphone?

Corner payphones don't kill people. Robot cars kill people.

[Timberridge]

Robot cars kill people.

Yes, cars, planes, asteroids, etc.. Basically anything that can exert an external force strong enough to damage one's vital organs.

[PNWbrit]

Someone should lose their job over this.

I'm guessing it was a problem with the GPS.

[grskier]

Corner payphones don't kill people. Robot cars kill people.

Please give me sources that helped you derive your opinion.

Why don't we look for millions of calls made vs deaths and millions of miles traveled vs deaths.

[AdironRider]

Not necessarily the point with most of it.

And just because you don't solve 100% of the issues 100% of the time does it mean that it's a deal breaker.

This isn't a computer router that you can just restart when it goes on the fritz. Its a two ton piece of machinery that earns zero empathy from the populace. Its an appliance that can kill people in this idealistic scenario, and if people are getting killed by appliances, those appliances wont sell.

And it is most definitely the point. What good is transportation that only works in Socal or AZ?

[Benny Profane]

Please give me sources that helped you derive your opinion.

Why don't we look for millions of calls made vs deaths and millions of miles traveled vs deaths.

Rumor has it that cell phones give you brain cancer. How's that headache?

[grskier]

This isn't a computer router that you can just restart when it goes on the fritz. Its a two ton piece of machinery that earns zero empathy from the populace. Its an appliance that can kill people in this idealistic scenario, and if people are getting killed by appliances, those appliances wont sell.

And it is most definitely the point. What good is transportation that only works in Socal or AZ?

It's already here, helping.


Let's agree on some metrics if you'd like to quantify and measure these things....

1.18 deaths per 100 million VMT is the going rate using current technology (which includes some level of autonomy for a percentage of the cars in use today).

Vehicles have already started to help, and automate vs the conditions you use... anti lock brakes being near the top of the list.

And, for a good chunk of the populace, getting up a snowy road is a non starter anyway, and I'd imagine that we are a few iterations away from solving those issues.

[Timberridge]

Rumor has it that cell phones give you brain cancer. How's that headache?

Instead of bluetooth you could request your next Subaru have a Trimline phone and a 1000 mile cord.

[Benny Profane]

It's already here, helping.


Let's agree on some metrics if you'd like to quantify and measure these things....

1.18 deaths per 100 million VMT is the going rate using current technology (which includes some level of autonomy for a percentage of the cars in use today).

Vehicles have already started to help, and automate vs the conditions you use... anti lock brakes being near the top of the list.

And, for a good chunk of the populace, getting up a snowy road is a non starter anyway, and I'd imagine that we are a few iterations away from solving those issues.

Wait. I always thought that the primary argument for robot transportation was safety. Now that we have a death, it's all relative now? Whoops, well, there's still less than those dumb, drunken humans. What happens when numbers climb? Then, at a certain point, you have to ask yourself, why robot cars?

[Benny Profane]

Instead of bluetooth you could request your next Subaru have a Trimline phone and a 1000 mile cord.

I don't drive with my head resting on my dashboard.

[grskier]

Wait. I always thought that the primary argument for robot transportation was safety. Now that we have a death, it's all relative now? Whoops, well, there's still less than those dumb, drunken humans. What happens when numbers climb? Then, at a certain point, you have to ask yourself, why robot cars?

Again, you have 1 pedestrian death... but you can't just measure these things by gross volume.... that's my point. Put proper metric to use.

[AdironRider]

The going rate for your metric is in virtually perfect weather conditions, in select cities with said conditions.

Saying it works great in LA or Vegas doesn't mean shit when it snows half an inch on your commute to work.

Let alone people have died as a result of truck crossing a rural highway (Tesla) or a pedestrian jaywalking (Uber) in these perfect conditions already. I encounter those situations literally 10x a day.

Don't get me wrong the concept is cool, just like flying cars, but the execution is going to fail miserably. Unless you implant RFID tags or whatever in every roadway in the land this dream is not happening. We can't even fix potholes currently.

[grskier]

The going rate for your metric is in virtually perfect weather conditions, in select cities with said conditions.

Saying it works great in LA or Vegas doesn't mean shit when it snows half an inch on your commute to work.

Let alone people have died as a result of truck crossing a rural highway (Tesla) or a pedestrian jaywalking (Uber) in these perfect conditions already. I encounter those situations literally 10x a day.

Don't get me wrong the concept is cool, just like flying cars, but the execution is going to fail miserably. Unless you implant RFID tags or whatever in every roadway in the land this dream is not happening. We can't even fix potholes currently.

Again, USE THE SAME METRIC.
You are taking a gross and comparing it to your 'every day' which isn't close to apples to apples.

And my going rate isn't what you opine.

[Dantheman]

And, for a good chunk of the populace, getting up a snowy road is a non starter anyway, and I'd imagine that we are a few iterations away from solving those issues.

Yeah, that argument amuses me. Like the average person is such a pro at driving in bad weather. It's not like the fundamentals of snow driving are all that complicated: Slow down, increase following distance, modulate power delivery, brake early. That's basically it, nothing you can't program. Rain and snow definitely mess with LIDAR, but it's only a matter of time before that is solved.

Wait. I always thought that the primary argument for robot transportation was safety. Now that we have a death, it's all relative now? Whoops, well, there's still less than those dumb, drunken humans. What happens when numbers climb? Then, at a certain point, you have to ask yourself, why robot cars?

Christ, Benny, this isn't that hard to understand. There's usually 35,000-40,000 auto deaths each year. Hypothetically, let's assume that full adoption of driverless cars still results in 10,000 deaths per year (probably a high estimate). That's a lot of dead people, but the safety argument is still valid because it's far less dead people than if people were still driving.

[iceman]

1.18 deaths per 100 million VMT is the going rate using current technology...

Betcha it's a lot higher than that for autonomous cars right now.

[Benny Profane]

We can't even fix potholes currently.

Or build pedestrian bridges.

[PNWbrit]

Betcha it's a lot higher than that for autonomous cars right now.

https://www.washingtonpost.com/opini...ce8_story.html