Self-Driving Cars

[Moliere]

Did you pull the trigger?

Nope. I purposely left the checkbook at home. I’m debating between the 1.5 and 2.0 engines. The 2.0 won’t be avavilbe until January. I like having a bigger engine but also want better gas mileage. They did have the color I want though and I doubled the miles on the car I drove. It started with 6 and I put 6 miles on it.


Sent from my iPhone using Tapatalk

[Pelado]

Did you pull the trigger?

Back to the gun analogy?

[SCcoug]

Whoa, that non-self driving car isn't close to being a self-driving car? Mind blown.

[falafel]

Nope. I purposely left the checkbook at home. I’m debating between the 1.5 and 2.0 engines. The 2.0 won’t be avavilbe until January. I like having a bigger engine but also want better gas mileage. They did have the color I want though and I doubled the miles on the car I drove. It started with 6 and I put 6 miles on it.


Sent from my iPhone using Tapatalk

The what?

I suppose what you were really saying is that you were committed to not make a purchase today. At least I hope you didn't really "purposely" leave your checkbook at home, as in you had to take it out of your pocket or something.

[Walter Sobchak]

I am not so sure that the impediments to human judgment that Walter points to will not eventually(and soon) be seen as acceptable trade-offs for autonomous vehicles. Will pedestrians or cyclists be victims of autonomous vehicles? Probably so at some point. But will the rate of such incidents be equal to or greater than the rate of incidents involving dumb ass moves by human drivers? I seriously doubt it.

Heh. Okay. Now consider this... let's imagine that your autonomous vehicle is presented with a situation where, via an unfortunate set of circumstances, it is heading on a collision course toward a crowd of people crossing the road. The car cannot stop in time, but it can avoid killing many people in the crowd by steering into a wall. However, this collision would kill you (the owner) and your child riding in the back seat. What should the software that is running the car do? If you as a prospective autonomous car buyer were aware that the software has been programmed against multiple use cases where it would sacrifice the occupant, would you still buy it?

[beefytee]

I'm not one of those guys who thinks that there will be no human drivers on the road in 5 years, but I can see it taking over most metropolitan areas in the next 15-50 years. It will be a long time before there are no human drivers.

One of the biggest obstacles mentioned is no lines on the road. I don't think this is a huge hurdle. I think one of the easier solution would be for states to offer electronically marked up maps that are based on surveying that is happening anyways. Some processes may need to be updated to make things more precise and more information may need to be gathered. e.g. where certain speed limits change, and other signage etc., but that is a relatively simple problem. States would be able to update information as things change for construction and other events. Cars would use multiple positioning technologies (not just GPS) to stay in the right place. Lines would only be needed for human drivers then.

The biggest issue to deal with will be inclement weather. There's a reason these services are starting in Phoenix and Las Vegas.

The issue about who to kill in bad situation is a non-issue to me. It's not a new issue. Human drivers kill lots of people. Self driving cars will kill much less. I'm fine with the car choosing to kill me as a passenger to save other peoples lives as the scenario is so improbable, I'll take my chances.

I'm not saying things are ready to go now, but they are problems that can be address and mostly mitigated.

As far as self-flying vehicles go:

https://www.engadget.com/2017/11/13/...ing-taxi-test/

I think there is less demand for pilot-less flights, because flying is much safer than driving and pilots are a small cost compared to aircraft costs and fuel. Don't get me wrong. Airlines will look to get rid of them as a cost saving measure, but it won't be super urgent.

[Jeff Lebowski]

It goes far beyond that one example.

People almost always overstate the reach of technology when predicting the future (where is my &&#@*@ jetpack?). I can certainly see how we will get cruise control that is more and more automated, but to the point that we replace drivers and remove steering wheels? nfw.

Swish.

https://www.wsj.com/articles/without...pid-1510488000

Without Humans, Artificial Intelligence Is Still Pretty Stupid
There are likely hundreds of thousands of people, world-wide, whose work is sold as AI, says one expert

Great. Let's have a human in the car factory secretly and remotely drive our self-driving car for us. That should work.

[ERCougar]

Can't remember where I read it, but one author described how the takeover will go down. It will start commercial, with long-distance trucking companies and then uber/Lyft/UPS purchasing huge fleets of vehicles. Within 5 or so years of that, people won't have a choice, as liability will just be too expensive.

Obviously, that assumes the programming hurdles will be overcome, and I'm a little like JL in that I think we tend to underestimate the obstacles of current ideas (while not imagining the true breakthroughs that are coming), but we're not very far at all from switching over to autonomous on freeways. Imagine how much more efficiently a trucker could go if he could just turn the truck onto autopilot while he sleeps in the cab.

[Uncle Ted]

Swish.

https://www.wsj.com/articles/without...pid-1510488000



Great. Let's have a human in the car factory secretly and remotely drive our self-driving car for us. That should work.

I remember in the mid to late 80's when computer science/engineering research proposals that had the words "artificial intelligence" were automatically rejected. Looking at the history it seems like the AI research "seasons" go in cycles.

[chrisrenrut]

I test drove a car today that has lane assist and adaptive cruise control. It was pretty cool. The lane assist worked most of the time but it didn’t work when lane markings were missing or severely faded. The tech today is great but it’s not close to being autonomous.


Sent from my iPhone using Tapatalk

I purchased a Honda this summer with both lane assist and adaptive cruise control. Both were extremely annoying and I've disabled both. The cruise control was especially maddening. When coming up on a vehicle one lane over it would engage by completely letting off the gas, which caused an abrupt deceleration. It was almost scary when doing 80 mph on the way to St George.

I don't know if they do firmware updates for cars. Yhat might at least offer a improvement through software implementation.

[Walter Sobchak]

This is the future.

Tesla ‘on Autopilot’ slams into parked fire truck on California freeway

A Tesla Model S reportedly on “Autopilot” smashed into the back of a fire truck parked at a freeway accident scene Monday morning, authorities said.

Two investigators from the National Transportation Safety Board are expected to arrive in Los Angeles on Wednesday to probe the crash, and will focus on “the driver’s actions and how the vehicle performed,” Bloomberg reported Jan. 23.

The union representing Culver City firefighters whose truck was hit around 8:30 a.m. on Interstate 405 in Culver City tweeted that the Tesla driver said he had been using Tesla’s Autopilot system, which performs automated driving tasks.

[...]

The fire truck had been parked in the left emergency lane and carpool lane, blocking off the scene of a previous accident, with a CHP vehicle behind it and to the side, said Culver City Fire Department battalion chief Ken Powell.

Both emergency vehicles had their lights flashing, Powell added.

The Tesla suffered significant damage, and the fire truck has been taken out of service to have body work done, Powell said.

Had any firefighters been at the rear of the truck rather than in front attending to the earlier crash, there “probably would not have been a very good outcome,” Powell said.

“It was a pretty big hit,” Powell said.

The firefighters union tweet indicated that the Tesla had been traveling at 65 miles per hour before the crash, but it was unclear to what extent the car may have slowed before striking the fire truck.

Like I said above, autonomous vehicles put everyone at risk... other drivers, cyclists, and pedestrians. Thank heavens a firefighter wasn't killed.

[creekster]

This is the future.

Tesla ‘on Autopilot’ slams into parked fire truck on California freeway



Like I said above, autonomous vehicles put everyone at risk... other drivers, cyclists, and pedestrians. Thank heavens a firefighter wasn't killed.

That’s the present, not the future.

[myboynoah]

I'll wait to see the results of the investigation. Lots of unknowns out there.

[Mormon Red Death]

This is the future.

Tesla ‘on Autopilot’ slams into parked fire truck on California freeway



Like I said above, autonomous vehicles put everyone at risk... other drivers, cyclists, and pedestrians. Thank heavens a firefighter wasn't killed.

Good point.. human drivers have never had an accident.

Sent from my SM-G930T using Tapatalk

[smokymountainrain]

seems a lawsuit against Tesla could net somebody a fair amount of money.