GM Plans to Launch Autonomous Car Without Steering Wheel or Pedals Next Year

[Hallucinator]

An ignorant owner of a driveless car who lives next door of a skillful GTA gamer who happen to be a professional hacker .....

.....and fun ensues.

Good luck.

 

[ol1bit]

See it in SFO!

GM has a great start! I think way more advanced that other! Watch this!

 

[travisty]

Meh, cars as a service doesn't sound like a terrible idea i suppose it all has to do with minimum possible cost i guess

This is it in a nutshell. If the cost of ride sharing is a fraction of owning a car few will actually own a car.

To paraphrase Thomas Edison:
"We will make [ride sharing] so cheap that only the rich will [own cars]."

 

[travisty]

Self driving cars are almost completely useless on snow covered roads (and bad weather in general), they don't deal with roadwork well or at all, unfinished and unpainted roads are also a major problem.
there is a long way to go before they are the norm for anyone outside of testbed cities.

You're stuck thinking in a linear paradigm. Self-driving cars (AI in general) is an exponential technology.

It may take another doubling time (~18 - 24 months) but AI will surpass every human driver very soon.

A good video discussing exponential technologies

 

[Twisted Kidney]

I'm in the process of teaching my eldest to drive.

My foot will go through the floor at some point during this process.

I'm not sure if I can handle the no pedals thing psychologically. Just put a pedal in there for me to stomp on for my own stress, It doesn't even have to do anything. Sort of like a baby's play table.

 

[nutzo]

I wonder how they will deal with more practical issues though. Like say, cab picks up someone who has partied a little bit too wildly and vomits in the car. Without a driver, how does it know to return to home base for a clean-up before accepting future passengers?

Or even people eating in the car, spilling their drink or leaving all their trash in it.

If this is an Uber like setup, then when you call a car, and it's a mess inside, they just need a button on the app where you can reject the car and send it to be cleaned.
Then they charge the last occupant a cleaning fee.

How about someone passing out in the car and they don't get out at their destination or don't close the door when they get out?
The car would keep beeping and telling them to please leave the car or close the door as the meter is still running.

I'm sure they wouldn't mind charging you for 8 hours while you sleep all night in the car.

 

[Retronym]

Not in my backyard.




*crickets*

 

[GNUse_the_force]

Or even people eating in the car, spilling their drink or leaving all their trash in it.
If this is an Uber like setup, then when you call a car, and it's a mess inside, they just need a button on the app where you can reject the car and send it to be cleaned.
Then they charge the last occupant a cleaning fee.

Couldn't they just ban food and drink consumption in the car and ban anyone who has a certain limit of alcohol in their system using the onboard Breathalyzer ? Bear in mind all autonomous lease cars will have the same policy so your own private transport could limit certain activities under the car TOS / EULA.

 

[Dalexx]

Autonomous cars only make sense when they are the only kind of cars on the road. As long as there are human-controlled cars on the streets and highways, autonomous cars will never be as safe or as useful as they claim.

Let's face it, an autonomous car loses its appeal when someone manually driving their car plows into it and kills you. Is the autonomous car gonna get out of the way like a human would? I don't think so.

You're assumption is that people aren't causing wrecks that Autonomous cars wouldn't cause. Drunk driving fixed, Texting while driving fixed, General distracted driving fixed. Excessive speeding fixed. Do you believe the other drivers on your road and interstate are going to properly react 100% of the time when someone drives recklessly? I don't and I imagine the stats will bare that out.

No one is arguing that there will be zero accidents with autonomous cars. But you can't really argue about fringe situations as to why they will fail.

This will never happen. Local governments strongly rely on DUI's for revenue and if we have fully autonomous cars with no steering wheel or pedals they can't give DUI's. We will, for the forseeable future, have steering wheels and pedals so that the local governments can still give DUI's under the premise that 'if the car failed, you would have to take control, and you were too drunk to take control so here's your $5000 dui"

That's some grade A crazy right there. I'm betting speeding tickets generate far more revenue than DUIs. Maybe some cities *might* be exploiting this, but it will be a state or Federal mandate

 

[PaulP]

You're stuck thinking in a linear paradigm. Self-driving cars (AI in general) is an exponential technology.

It may take another doubling time (~18 - 24 months) but AI will surpass every human driver very soon.

A good video discussing exponential technologies

I laughed when I read this. You must be young. I have heard way too many of these types of claims in the past that turned out to be dead wrong to believe them now. Let's revisit this topic in another two years and discuss why your prediction failed.

 

[GNUse_the_force]

Autonomous cars. But you can't argue about fringe cases as to why they will fail.

That's the thing. So many people have automatically assumed there will be less death and accidents ( there probably will but still ). We can speculate but we just don't know yet, we haven't been through the whole life cycle yet of millions of connected vehicles driving hundreds of millions if not billions of miles to work out the numbers. Sure we all hate human drivers, everyone thinks there the best and everyone else is terrible.. but what about hacking, bugs, errors or miscalculations ? For instance you might ( almost certainly will ) have a situation where annually the death toll is lower for Auto cars, but every decade there is a single hack that wipes out the same number of people as human error in that 10 year cycle. Not withstanding when you step in a car and drive it yourself there is a degree of taking your life into your own hands.. but with Ai cars your putting your life into the hands of an invisible tech employee behind an army of lawyers.

im not entirely against them btw. So don't take this post as confrontational as some might do, just playing advocate here.

 

[Azrak]

So this is going into orbit too?

 

[Dalexx]

That's the thing. So many people have automatically assumed there will be less death and accidents ( there probably will but still ). We can speculate but we just don't know yet, we haven't been through the whole life cycle yet of millions of connected vehicles driving hundreds of millions if not billions of miles to work out the numbers. Sure we all hate human drivers, everyone thinks there the best and everyone else is terrible.. but what about hacking, bugs, errors or miscalculations ? For instance you might ( almost certainly will ) have a situation where annually the death toll is lower for Auto cars, but every decade there is a single hack that wipes out the same number of people as human error in that 10 year cycle. Not withstanding when you step in a car and drive it yourself there is a degree of taking your life into your own hands.. but with Ai cars your putting your life into the hands of an invisible tech employee behind an army of lawyers.

im not entirely against them btw. So don't take this post as confrontational as some might do, just playing advocate here.

I can respect that stance, there is a healthy chunk of unknown that will need to be addressed. But I think this is on par with there will be a computer in every home\a smart phone in every hand. And as I stated in another thread about this. We're a good 15-20 years away from this being something they can consider making mandatory. Which is why in my original post, I said it's seems radical to say no steering wheel this early.

 

[travisty]

I laughed when I read this. You must be young. I have heard way too many of these types of claims in the past that turned out to be dead wrong to believe them now. Let's revisit this topic in another two years and discuss why your prediction failed.

Oh i'm well aware of the main predictions that have been incorrect. Do some research on who founded Singularity University - Ray Kurzweil and Peter Diamandis - and learn what's happening around us today and you'll have a better understanding of what's different today than the 60's. Hint: What the 60's thought could happen is now happening.



Some predictions are saying 2025 is when car ownership is dead.

 

[jfreund]

"My uncle has a country place no one knows about
He says it used to be a farm before the motor law.
On Sundays I elude the eyes and hop the turbine freight
To fall outside the wire where my white haired uncle waits"

 

[GNUse_the_force]

it's seems radical to say no steering wheel this early.

Your right. IMO, in the first few decades your only going to get mass adoption if you have a wheel & pedals with override vs not having anything.People still have biologically hard-coded fight or flight responses and with no means of control your unable to fight or flight. Every living creature has FoF response, it's not about being a Luddite it is about having a self preservation instinct.
I think that's why it seems so freakishly unnatural until the experience can relax that response and provide a suitable physiological 'get out' if needed mechanism. Given this, it seems a little irresponsible of GM to be touting such a system so early on. It might actually delay public adoption of autonomous vehicles with such a sudden 'we took everything away from your control you were used to, and locked you in a high speed traveling metal box controlled by a computer' approach right out of the gate. A bit like when Oculus warned others not to 'poison the well' with badly realized products and destroy consumers confidence in VR.

New technologies need to adapt to humans not the other way round if they are to be successful.

 

[THRESHIN]

I'm all for driverless cars, and I love driving. It's the fact that overall humans are terrible at this task. Nearly every accident is caused by human error in one way or another. Not checking blind spots during lane change, going too fast for road conditions, outright stupidity like running a stop sign, and so on. I don't want a driverless car for myself, I want all the other idiots on the road to have them! But I'm actually willing to make the sacrifice....


Anyway, I'll believe this one when it comes out. GM has a long and proud history of bragging about new tech that never sees the light of day. I remember seeing a new tech presentation from a GM engineer when I was in college. Some neat ideas, not a single one ever came out. This was back in 2003 I think.

And then GM 's other proud tradition....scary recalls. Sure they all have safety recalls. I find the ones GM has are usually pretty serious. My car has had a couple. Spare tire was underinflated, coolant antifreeze mixture was a little off. Big deal...a coworker bought a cruze and his recall was the driveshafts could snap without warning. Yikes. I'll wait this one out thanks.

 

[Unter Dog]

You're stuck thinking in a linear paradigm. Self-driving cars (AI in general) is an exponential technology.

It may take another doubling time (~18 - 24 months) but AI will surpass every human driver very soon.

A good video discussing exponential technologies

Its not the main processing/AI I'm worried about. Its the quality of the sensors and reliability of the hardware.

I'm in automotive electronics continuous improvement and i'm feeling like they are going to need to take a step toward Aeronautics levels of quality and preventative maintenance. We can draw bell curves on how long we think stuff will last in the field so when assemblies and harnesses get x hours old they get replaced. Thermal fatigue and vibration are the main wear mechanisms assuming you design properly. You don't want to rely on redundancy to save your ass if you don't have a steering wheel. Right now automotive looks for sub 100 ppm for controllers (100 failures per million) failure rates which is going to have to get much lower for me to feel good about something like this. A lot of automotive show lower PPM, but they generally "scrub" their numbers.

 

[nutzo]

This is it in a nutshell. If the cost of ride sharing is a fraction of owning a car few will actually own a car.

To paraphrase Thomas Edison:
"We will make [ride sharing] so cheap that only the rich will [own cars]."

And when they don't want people going anywhere (think Riots) you will be stuck, since they will shut down the ride sharing services.

Even if they manage to create a cheap ride sharing service, I would still want to have my own car. Specifically for emergencies, long distance travel, etc.

This would make it easy to eliminate the need for a 2nd car however.
Also would make it easy when your car is in the shop, getting dropped off at the airport, etc.

 

[GNUse_the_force]

would still want to have my own car. Specifically for emergencies,

Nobody gets to escape the zombie apocalypse they are eaten alive in their fancy autonomous cars

 

[nutzo]

Nobody gets to escape the zombie apocalypse they are eaten alive in their fancy autonomous cars

And when zombie steps in front of your autonomous car, the car will politely stop and wait for the zombie to step aside. When you are completely surrounded by zombies, your car will refuse to move. You will either have to open the door and make a run for it, wait for the zombies to break in, or sit there and starve to death. If you make a run for it and the zombie follow you, your car will then be free to go on it's way

They will have to build the AI to recognize zombies so this doesn't happen.

I think I'm putting to much though into this though into this

 

[face2palm]

Oh i'm well aware of the main predictions that have been incorrect. Do some research on who founded Singularity University - Ray Kurzweil and Peter Diamandis - and learn what's happening around us today and you'll have a better understanding of what's different today than the 60's. Hint: What the 60's thought could happen is now happening.

Some predictions are saying 2025 is when car ownership is dead.

Singularity University?
LOL
Their entire site is just riddled with biz-speak and marketing nonsense.

This is the best part. "Singularity University is not a degree granting institution."

We should totally listen to what these clowns have to say.

 

[dvsman]

Three things:

1) AI car vs AI car accident - who is liable? In some states (including DC) there is no fault. If it's a fender bender (no injuries), everyone fixes their own. If it's a major accident with a ticket being issued, then the fault would be the same as an all human accident. And the point someone already raised about new AI cars being so heavily monitored that the available telemetry data will allow the insurance companies / lawyers / courts to figure out who's at fault much easier.

2) As for the question of "Should it protect the pedestrian or protect the occupant" - the answer is it will try to brake and it will be more effective at it (due to better reaction time) than a human trying to make the same decision (again human reaction time will never beat a machine in this). I see people pose this question all the time but I don't see them asking the same when it's a shitty human driver at the wheel.

3) Would I use it? I wouldn't have a problem with riding in one for my commute if the economics of it worked out. It would be the same level of control I would have if I were sitting in a taxi or uber - which is basically none. And I'd guess a computer / AI driver would be much better than the human taxi drivers we have here in DC

 

[Wrecked Em]

This will be an epic failure and will destroy the careers of the fools who are advocating this nonsense. People on this forum should know better. Buggy software crashing on your computer is annoying but buggy software in a self-driving car will kill innocent people. Those who might try and argue that self driving car software will not be buggy should be banned for pretending to know anything about computers...

We've been writing software to aircraft development assurance levels for decades. Those systems are much more critical than anything on an automobile. If a 787 software "crashes", you potentially kill thousands. If a Tesla crashes, you kill a maybe a dozen. Plus, there is much less radiation at ground level, so the systems will be inherently more robust than in the air.

 

[sadsteve]

If we no longer have a steering wheel, accelerator pedal, brake pedal, or any other drive controls, then what is the point of facing forward or even having a windshield? Wouldn't design go back to the 19th century style carriages with all passengers facing each other?

Well, there's a thing called car sickness that a fair number of people have. Riding in an enclosed box with my wife facing me would end up with me covered with puke. She's also claustrophobic so there needs to be windows.

 

[Deleted member 204526]

Simply put, this is the future.

Human-driven vehicles will be outlawed sooner than most imagine. (the vast majority will not own a vehicle either so it won't be a drastic change)

Here we have a myopic city-dweller who has no concept of hilly and curvy country backroads, many unpaved.

 

[Zarathustra[H]]

We've been writing software to aircraft development assurance levels for decades. Those systems are much more critical than anything on an automobile. If a 787 software "crashes", you potentially kill thousands. If a Tesla crashes, you kill a maybe a dozen. Plus, there is much less radiation at ground level, so the systems will be inherently more robust than in the air.

True, stability of software is a greater concern in aircraft, but on the flip side ground transportation is much more complex to automate.

An automated air system just needs air traffic control guidance as well as radar for collision avoidance. There is much less in the way of obstacles up there. Get down to the ground and things get much more complicated. You have to stay on roads for one, the roads may be ill defined with or without lines, etc. And there are TONS of obstacles.

Other vehicles, pedestrians, barriers, stop lights, pot holes, etc. etc.

 

[sadsteve]

Couldn't they just ban food and drink consumption in the car and ban anyone who has a certain limit of alcohol in their system using the onboard Breathalyzer ? Bear in mind all autonomous lease cars will have the same policy so your own private transport could limit certain activities under the car TOS / EULA.

Like how everyone follows the EULA (or even reads them) that comes with software.

 

[travisty]

Here we have a myopic city-dweller who has no concept of hilly and curvy country backroads, many unpaved.

Nope, I hate cities

 

[sadsteve]

Oh i'm well aware of the main predictions that have been incorrect. Do some research on who founded Singularity University - Ray Kurzweil and Peter Diamandis - and learn what's happening around us today and you'll have a better understanding of what's different today than the 60's. Hint: What the 60's thought could happen is now happening.



Some predictions are saying 2025 is when car ownership is dead.

Heh, my 2003 Honda Civic only has ~117,000 miles on it, I've got at least another 12 years before I hit the buy a new car button (250,000 miles).

 

[WhoMe]

Some of us actually enjoy driving cars, going fast, and taking risks. While this may be lumped in with risk taking, I'm still going to get the package that has the driving wheel on my car

Take up sky diving or free climbing.

 

[WhoMe]

I laughed when I read this. You must be young. I have heard way too many of these types of claims in the past that turned out to be dead wrong to believe them now. Let's revisit this topic in another two years and discuss why your prediction failed.

Just 15 years ago AI couldn't steer a vehicle through a simple obstacle course. Tech moves fast, AI is moving very fast. Glad I'm not still not using my TRS-80.

 

[MV75]

Cars as a service, no thanks. There are already too many subscriptions that are nickel and diming us into oblivion with nothing to show for it.

Saving time? More like working us longer. You're expected to work on the way to work, do a full day as that work in the car is extra, and then work on the way home. The whole city will be a school pickup zone as everyones personal autonomous car will be there to personally whisk them away from the front door in fairy land to some other magical point to point destination that is a straight line. Yea, you and the other 50000 people there buddy.

Well, there's a thing called car sickness that a fair number of people have. Riding in an enclosed box with my wife facing me would end up with me covered with puke. She's also claustrophobic so there needs to be windows.

Backwards facing seats are supposed to actually help with that. Does choosing the back facing seats in trains help? They generally do for car sick people.

 

[GNUse_the_force]

Like how everyone follows the EULA (or even reads them) that comes with software.

Just playing devils advocate here. Won't all Ai cars come with internal cameras and multiple microphones always connected to the cloud ? It's not like there won't be crystal clear audio & photographic evidence of peoples behavior. It may even be required to keep insurance companies happy.

Maybe we should stop calling them cars and call them drone-cabs.. or johnny cabs

 

[Krenum]

Simply put, this is the future they want you to have.

Fixed.

 

[KarsusTG]

Automated vehicles have a LONG way to go. I don't see one replacing our work truck pulling materials out onto the oil fields, or on our ranch where half the places only have tractor paths. Much less operating in the north when there is 3ft of snow on the ground, or 1ft of mud to navigate through..

Ya, happening tomorrow in no place near you.

 

[WhoMe]

Automated vehicles have a LONG way to go. I don't see one replacing our work truck pulling materials out onto the oil fields, or on our ranch where half the places only have tractor paths. Much less operating in the north when there is 3ft of snow on the ground, or 1ft of mud to navigate through..

Ya, happening tomorrow in no place near you.

Living in a deep snow area that also has tourists, I can safely say most people can't drive in snow anyway, tech will be able to handle it better soon-- it's a learned skill. Not even my lifted pickup 4x4 with chains and very aggressive tires can handle 3' of snow, it can handle two though (down hill). It also does well on 1.5' of frozen slush like I had last year when two of the local plows broke along with the rotary plow...but no car could handle that, exceptional circumstances don't mean most won't be able to use a self-driving car. I'd love one, but yeah my 18 year old pickup will probably become a 30 year old back-up for extreme cases (or if I need to go off road to get away from a forest fire) but 99% of the time I won't need it. Special cases like snow plows will requite human drivers a bit longer, but again these things are training and experience and AI will catch up.

 

[KarsusTG]

Living in a deep snow area that also has tourists, I can safely say most people can't drive in snow anyway, tech will be able to handle it better soon-- it's a learned skill. Not even my lifted pickup 4x4 with chains and very aggressive tires can handle 3' of snow, it can handle two though (down hill). It also does well on 1.5' of frozen slush like I had last year when two of the local plows broke along with the rotary plow...but no car could handle that, exceptional circumstances don't mean most won't be able to use a self-driving car. I'd love one, but yeah my 18 year old pickup will probably become a 30 year old back-up for extreme cases (or if I need to go off road to get away from a forest fire) but 99% of the time I won't need it. Special cases like snow plows will requite human drivers a bit longer, but again these things are training and experience and AI will catch up.

I will admit that I have done some stupid stuff. I lived in Northern Indiana and that first winter I learned a LOT about what I could and couldn't get through. But honestly my point was, as a human you can judge the capabilities of your vehicle, and even push the limit if you need to. Not only is an automated vehicle not going to do that, the overwhelming majority of the country is rural roads, many of which never get plowed or are poorly maintained.

 

[WhoMe]

I will admit that I have done some stupid stuff. I lived in Northern Indiana and that first winter I learned a LOT about what I could and couldn't get through. But honestly my point was, as a human you can judge the capabilities of your vehicle, and even push the limit if you need to. Not only is an automated vehicle not going to do that, the overwhelming majority of the country is rural roads, many of which never get plowed or are poorly maintained.

But a computer will know far more about its car and the current conditions than a human ever could, computers can react far faster than human and deep learning etc. mean they will be able to learn what humans can and they make far less mistakes. I was going uphill in blizzard one day, I took my foot of the gas when I saw some headlights coming down, it was an instinctive reaction, and that was it lost momentum and ended up parking in someone's driveway until the plow came. A computer would have been able to see that the lights were on a skip loader plowing a driveway and no threat, lidar or IR imagery would have shown that in real time.

 

[KarsusTG]

But a computer will know far more about its car and the current conditions than a human ever could, computers can react far faster than human and deep learning etc. mean they will be able to learn what humans can and they make far less mistakes. I was going uphill in blizzard one day, I took my foot of the gas when I saw some headlights coming down, it was an instinctive reaction, and that was it lost momentum and ended up parking in someone's driveway until the plow came. A computer would have been able to see that the lights were on a skip loader plowing a driveway and no threat, lidar or IR imagery would have shown that in real time.

I don't disagree, I just think we have a LOT of hurdles that are going to have to be overcome. We can't even get cars to auto parallel park themselves reliably. They can do it, but have like a 5% failure rate. Even the standard 99.99% is 52 minutes of guaranteed clusterfuckery which may or may not happen at 80mph with your child in the car.

Honestly, the cars doing these tests are meticulously maintained. I have yet to see one with 2 years of dirt and grime on all their sensors, 1 working headlight, and brakes that are well past their prime because their owners don't give a shit. This is basically half the taxi's in the world right now. Have you ever seen a clean taxi?