Here's What's Missing from Self-Driving Cars: TRUST

[McRat]

@NerdUno, you have said repeatedly that GM's Supercruise will be better than Autopilot when it is released.

But GM concealed a faulty ignition switch for over 10 years, resulting in an estimated 124 deaths, possibly more.

Since GM hid defects leading to at least 124 deaths, by your logic in this thread one should never "TRUST" GM. Yet you somehow manage to "promise" that GM's Supercruise system, when it is finally released, is going to be better than Autopilot, Move Over Tesla. Here Comes Cadillac.

Here's how Wikipedia describes the results of GM's misdeeds:

Quantifying fatalities
The faulty ignitions have been linked (by GM itself) to 124 deaths.[3] GM originally only linked the failures to 13 deaths and 31 crashes.[12] The company only counted incidents which resulted in head-on collisions in which the airbags did not deploy. It did not include, for example, an incident where after a car's ignition switch failed, the car "spun out, hydroplaned, hit an oncoming vehicle and rolled off the road, dropping 15 feet into a creek".[13] In a collision in which two young women in a Chevrolet Cobalt were killed when the ignition switch shut off the engine, GM only counted the death of the woman in the front seat, because the death of the woman in the back seat was not caused by the failure of the airbag to deploy.[14] Most of the victims were under age 25.[15] On June 3, 2014, Reuters published an analysis concluding that the faulty switches were responsible for 74 deaths, based on Fatality Analysis Reporting System data. General Motors disputed its conclusion, and stood by their original figure of 13 deaths after the report was released as of June 3, 2014.[16]By the end of September, Reuters stated in an article that 153 deaths were linked to the faulty ignition switch.[17] As of March 2015, GM had offered compensation for 11 Category One injuries, along with 115 Category Two injuries.[18] In April 2015, GM officially noted that the death toll was believed to have reached 87, higher than the previous number of 74 they reported in March 2015.[19] Upon its completion, the compensation fund established by GM had offered compensation for 124 deaths, nearly 10 times more than the 13 deaths GM executives reported in April 2014.[3][20] However, the true number of deaths resulting from the ignition switch is likely higher as GM's compensation fund rejected more than 90% of claims and it did not include claims that are part of the ongoing Multidistrict Litigation.[21][3] General Motors ignition switch recalls - Wikipedia​

Anyone who wants to make a rational judgment on whether to trust the safety of Tesla's Autopilot need look no further than Tesla vehicles' outstanding safety record -- most recently Model X was rated the safest SUV ever tested by the NHTSA -- as well as the NHTSA's finding that there was a 40% reduction in airbag inducing accidents after AP1 was activated. I expect comparable or better numbers from AP2, with the safety benefits continuing to increase as time goes on and the Full-Self Driving system is activated.

That was written by the class action law firm that collected more than 1/2 the settlement. Part of their advertising budget.

There are many other incidences where heavy key chains have turned ignition switches off and disabled airbags on all brands. This issue predates airbags. So why did the suit get filed so late?

When the government mandated airbags, automakers demanded tort protection due to the dangerous nature of airbags. The ignition switch torque memo was a way to bypass the airbag tort rule. People were dying due to airbags not firing, but the lawyers could not sue until they found a way around it.

Trivia - The NHTSA killed more people due to airbag issues than GM and Takata combined. 290 people were killed in low speed survivable collisions by the airbag deployment before the NHTSA allowed automakers to use 2-speed airbags. IIRC, it was 9 years after automakers told the NHTSA that current airbags were killing people unnecessarily before any action was taken? But the same tort protection that protected automakers protected incompetent and lazy NHTSA officials from doing the right thing.

 

[AnxietyRanger]

@EinSV

I am not understanding how a report of a broken accelerator pedal relates to autonomy. I would say it has absolutely nothing to do with it.

Are you saying that even though the Model X is the safest SUV ever rated by the NHTSA, you would recommend buying something else because stuff may crop up in the future and the NHTSA might turn out to be wrong? If so, that does not seem like sound logic to me.

Again, not seeing what this has to do with autonomy, which is supposedly the topic of this thread.

Well, you brought up ignition switches as an example of GM's untrustworthiness relating to autonomous. No?

@NerdUno, you have said repeatedly that GM's Supercruise will be better than Autopilot when it is released.

But GM concealed a faulty ignition switch for over 10 years, resulting in an estimated 124 deaths, possibly more.

My point was similar - a response to GM's on-going ignition switch woes that took years to uncover, by noting that Tesla's potential woes may not be quite known yet.

Over time our understanding of what the features of Tesla cars are gets better, as evidenced by counter gate, DC peak rate throttling and breaking accelerator pedals. So, as Tesla's are getting older, we are learning about them and all is not peachy.

If we are to compare GM woes from a decade or so, we need to acknowledge Tesla is not old enough yet for that comparison, as we don't know "everything" the same way we know about a decade old GM...

Time teaches.

But be that as it may, it is also quite possible that any woes in GM's ignition or Tesla's accelerator pedals - or even the company's handling - bear no relevance to their autonomous solution comparison. Those teams and the history and future of those projects are probably quite separate....

 

[EinSV]

You realize that was written by the class action law firm that collected 1/2 the settlement, right? Part of their advertising budget.

The reality is that there are many other incidences where heavy key chains have turned ignition switches off and disabled airbags on all brands. This is actually an issue that predates airbags by half a century. <snip>

GM paid a $900 million penalty. And it was GM who decided to pay out claims to the 124 victims. $900 Million Penalty for G.M.’s Deadly Defect Leaves Many Cold

I know you can be counted on to defend GM to your last breath, but there is no getting around it on this one.

 

[EinSV]

@EinSV

Well, you brought up ignition switches as an example of GM's untrustworthiness relating to autonomous. No? <snip>.

No. I was pointing out the logical disconnect in @NerdUno's posts.

 

[McRat]

GM paid a $900 million penalty. And it was GM who decided to pay out claims to the 124 victims. $900 Million Penalty for G.M.’s Deadly Defect Leaves Many Cold

I know you can be counted on to defend GM to your last breath, but there is no getting around it on this one.

Welcome to our legal system.

Create a villain, create a crime, create billions.

But if you are all that interested in automotive technology you should not read class action advertisements. Lawyers are not engineers, nor are they necessary honest.

The infamous Rocket Motor Blowing Up Truck Trick was alleged to have been funded by a legal firm, much like the Vaccines Cause Autism debacle.

The fact you believe lawyers more than engineers and scientists shows how effective the law firm marketing has become. Soon, history will be written by lawyers for profit.

 

[electracity]

Similar here. In 2014 other than being a bit miffed at Tesla for disabling my SAS for few months, I was telling everybody who asked how great their products. Today, I give cautionary tales, warning about early adopter issues, about overhyped features and recommending buying the low end without EAP of FSD.

There is one more thing that worries me about Tesla self-driving. They gamble a lot by selling Beta software (of their own admission) to the public. The problem is, if they screw up and cause some shocking public disaster, this could set the entire autonomous car industry behind by 10+ years, if people freak out and force the legislators to ban autonomous driving. This can happen. If an autonomous Tesla semi plows into a school bus somewhere killing kids, it's game over for all manufacturers for a while.

I don't think Tesla can set autonomous driving back ten years anymore. The effort has grown too big.

I think the autonomous drive situation with Tesla is separate from other self inflicted wounds. Tesla is competing for talent, which is a reason for the bold claims. Musk desperately wants to hold onto the perception of autonomous leadership established with AP1.

But the unforced errors, such as the horsepower claim, are unfathomable. There's no upside to the many exaggerations and unfulfilled promises. These efforts don't sell more cars, they just sow distrust. Which is the OPs point.

Tesla's fate now rests with execution, not new strategy. I expect Musk doesn't like the day to day much anymore but has learned he has to stop making car specific claims to reduce stress. So I expect this situation to improve.

 

[NerdUno]

@NerdUno, you have said repeatedly that GM's Supercruise will be better than Autopilot when it is released.

Just for the record... that's not exactly correct. What I said was "Betting the Caddy won't be slamming on the brakes at overpasses and drifting out of its lane." I am confident that Cadillac will thoroughly test its software before releasing it rather than pushing out a half-baked product, labeling it as beta, and leaving it up to customers to find all the shortcomings on their own. Time will tell whether Supercruise is functionally better, but I do believe from a reliability standpoint (for what it does) it will do it in a safer and more predictable way than what we have now with AP2. YMMV!

 

[boonedocks]

I have only owned one Ford in my life, Lincoln Navigator excluded, and it was a newly released Ford C-Max. I have a long daily commute and it is almost all expressway at 75mph speeds. The Monroney sticker had an advertised MPG that the C-Max could never achieve in any normal course of driving. Ford sent me 2 different checks specifically stating that the stated MPG was incorrect. The total amount % to the car was more than I would have ever expected. Can't remember the exact details, but it was significantly more than I would have expected and both checks were completely unexpected.

Question for the group: as a similar example, do you feel you were delivered 100% of what was promised with AP1. Not where it is but what the original display of capabilities was when originally advertised for sale? One small example I remember was being dropped off in a parking lot and the car parking itself and then being able to summons it back to the curb. I do believe in the earliest of the Tesla online ordering it was listed somewhat that way. (didn't keep screen shots, I have however kept every screenshot from day one of what my HW2 EAP / FSD, both purchased, has been advertised to do)

Do you feel Tesla should repay some of what you paid for AP1 HW1 due to it never being able to accomplish some of it's advertised capabilities. I sure wish I had original screenshots from everything that was originally shown and the changes made month to month where features simply vanished.

 

[croman]

Welcome to our legal system.

Create a villain, create a crime, create billions.

But if you are all that interested in automotive technology you should not read class action advertisements. Lawyers are not engineers, nor are they necessary honest.

The infamous Rocket Motor Blowing Up Truck Trick was alleged to have been funded by a legal firm, much like the Vaccines Cause Autism debacle.

The fact you believe lawyers more than engineers and scientists shows how effective the law firm marketing has become. Soon, history will be written by lawyers for profit.

Oh lawyers are the boogeymen. Lets ignore the fact lawyers also are protecting our Constitution from the crazies, protecting innocent refugees from discrimination, and otherwise obtain justice for injured individuals (and hold corporations accountable for duping consumers).

No profession is all roses. Doctors maim, injure, and kill patients through negligence. Priests abuse their flock.

Lets just agree that politicians are the one group that does nothing nowadays to help people except themselves?

 

[EinSV]

Just for the record... that's not exactly correct. What I said was "Betting the Caddy won't be slamming on the brakes at overpasses and drifting out of its lane." I am confident that Cadillac will thoroughly test its software before releasing it rather than pushing out a half-baked product, labeling it as beta, and leaving it up to customers to find all the shortcomings on their own. Time will tell whether Supercruise is functionally better, but I do believe from a reliability standpoint (for what it does) it will do it in a safer and more predictable way than what we have now with AP2.

Well, just for the record, you said a lot of things. The kinds of things many people here on TMC like to call "promises."

Here are a couple recent examples. Will be interesting to see what you have to say in a few months if/when Cadillac does not live up to your "promises," which I think it's fair to say have included that this fall Supercruise will be better than AP2.

I'm puzzled by Tesla's continuing use of the "subject to regulatory approval" language. That must not be such a big deal if Cadillac is already announcing a hands-free vehicle for interstates beginning this fall. Their cars use the same highways and are subject to the same "regulations" unless I've missed something. Didn't see those weasel words in Cadillac's announcement. Of course, we all know that Tesla had barely started coding when all of their pie-in-the-sky promises were being made, i.e. the shortcomings we all are enjoying with our AP2 vehicles had little or nothing to do with regulatory hurdles.

So long as Tesla continues to monitor hands on the wheel, FSD could be turned on tomorrow. But wait. They've forgotten to build it. The rest of the regulatory approval disclaimer is pure B.S. Ask Cadillac. They'll have FSD on interstates this fall with hands in your lap so long as your eyes are glued to the road ahead.

 

[boonedocks]

Oh lawyers are the boogeymen. Lets ignore the fact lawyers also are protecting our Constitution from the crazies, protecting innocent refugees from discrimination, and otherwise obtain justice for injured individuals (and hold corporations accountable for duping consumers).

No profession is all roses. Doctors maim, injure, and kill patients through negligence. Priests abuse their flock.

Lets just agree that politicians are the one group that does nothing nowadays to help people except themselves?

That's why we need to drain the swamp.

 

[AnxietyRanger]

No. I was pointing out the logical disconnect in @NerdUno's posts.

If so, I don't see the disconnect by your own logic. You claim that non-autonomous related faults (Tesla accelerator pedal breaking off) are not relevant to autonomous performance in a Tesla, yet seem to be suggesting @NerdUno is wrong in predicting Supercruise performance being better than Autopilot because of that. How is GM's ignition switch issue somehow relevant to GM's Supercruise performance, if Tesla's issues other are not relevant to Autopilot performance?

@NerdUno, you have said repeatedly that GM's Supercruise will be better than Autopilot when it is released.

But GM concealed a faulty ignition switch for over 10 years, resulting in an estimated 124 deaths, possibly more.

...

Anyone who wants to make a rational judgment on whether to trust the safety of Tesla's Autopilot need look no further than Tesla vehicles' outstanding safety record

 

[SomeJoe7777]

Model S does apparently have the breaking acceleator pedal issue.

"The breaking accelerator issue." ??? There is no such thing.

Neither you nor the owner of that car know why that accelerator broke. You're jumping to the speculative conclusion that somehow it's a design issue, when there are dozens of other issues that could have caused it:

• Defect in the material before the part was ever installed.
• Damaged during installation.
• Installed incorrectly.
• Incorrect part for the car.
• Damaged in the car prior to this incident.

Your emotions have caused you to lose objectivity. You're seeing everything as evidence of Tesla's wrongdoing, even when no such conclusion can be drawn.


"When the only tool you have is a hammer, then everything looks like a nail."

 

[AnxietyRanger]

@SomeJoe7777

You are missing my point. I was demonstrating to @EinSV how Tesla's are young and we have not learnt about all the possible defects they are prone to. After all, @EinSV was using a decade-long episode as evidence against GM's lack of safety. I was demonstrating how aging Tesla's are only now starting to show those symptoms that eventually can possibly become known as flaws. Comparing GM and Tesla like this is apples and oranges, because the latter (Model S) does not have ten years of data, hardly even five.

I fully concur the accelerator pedal breakage may be limited in scope, but given that we now have a few such episodes reported on older Teslas, does suggest a wear and tear issue that is not evident in new cars, but takes time to become visible... Which was my point. It takes time to see what the full level of Tesla safety and/or issues are - just like it has taken time with GM products to see certain issues not evident from the start.

 

[EinSV]

@SomeJoe7777

You are missing my point. I was demonstrating to @EinSV how Tesla's are young and we have not learnt about all the possible defects they are prone to. After all, @EinSV was using a decade-long episode as evidence against GM's lack of safety. I was demonstrating how aging Tesla's are only now starting to show those symptoms that eventually can possibly become known as flaws. Comparing GM and Tesla like this is apples and oranges, because the latter (Model S) does not have ten years of data, hardly even five.

I fully concur the accelerator pedal breakage may be limited in scope, but given that we now have a few such episodes reported on older Teslas, does suggest a wear and tear issue that is not evident in new cars, but takes time to become visible... Which was my point. It takes time to see what the full level of Tesla safety and/or issues are - just like it has taken time with GM products to see certain issues not evident from the start.

But in autonomous driving, Tesla has far more real world experience with cars in the hands of actual, real-world drivers than anybody. So if you think experience matters, you should count that as a point in Tesla's favor.

 

[AnxietyRanger]

But in autonomous driving, Tesla has far more real world experience with cars in the hands of actual, real-world drivers than anybody. So if you think experience matters, you should count that as a point in Tesla's favor.

Remains to be seen. Tesla's greatest asset is their experience and delivery system in software updates. That I freely give them credit for. They are also pushing the envelope more than the competition. But as for delivering a reliable, trustworthy self-driving system, the history is quite checkered...

I actually, personally, believe Audi will be out with a reliable self-driving system before Tesla. You know, the kind that allows you to read a book at least some of the time. Having followed both closely for years now, that is my hunch. Audi will fly under the radar as far as the consumer is concerned for longer, but when it releases something, I expect it will be significantly better than what Tesla will be able to show this year.

We shall see. Maybe someone else surprises both of the companies I mentioned...

 

[Pruitt]

There is one more thing that worries me about Tesla self-driving. They gamble a lot by selling Beta software (of their own admission) to the public. The problem is, if they screw up and cause some shocking public disaster, this could set the entire autonomous car industry behind by 10+ years, if people freak out and force the legislators to ban autonomous driving. This can happen. If an autonomous Tesla semi plows into a school bus somewhere killing kids, it's game over for all manufacturers for a while.

So you're saying self-driving cars should not be sold because an accident might result in the banning self-driving cars?

If, when electrical power was being introduced to homes, it had to face the regulatory and legislative bureaucracy innovators face today, it would never have been allowed. We'd all still be sitting in the dark.

 

[boonedocks]

If so, I don't see the disconnect by your own logic. You claim that non-autonomous related faults (Tesla accelerator pedal breaking off) are not relevant to autonomous performance in a Tesla, yet seem to be suggesting @NerdUno is wrong in predicting Supercruise performance being better than Autopilot because of that. How is GM's ignition switch issue somehow relevant to GM's Supercruise performance, if Tesla's issues other are not relevant to Autopilot performance?

Dang, that is one severe way for Tesla to control too many launches!!!! I thought limiting the battery and motors was severe....

 

[AnxietyRanger]

So you're saying self-driving cars should not be sold because an accident might result in the banning self-driving cars?

If, when electrical power was being introduced to homes, it had to face the regulatory and legislative bureaucracy innovators face today, it would never have been allowed. We'd all still be sitting in the dark.

I guess the question comes down to the right balance.

Tesla's approach is very different from the rest of the industry on what they ship to the end-customer and when.

It is not completely unreasonable to think a risk realized because of Tesla's more risky approach could have negative repercussions.

MobilEye made this case during the Autopilot 1 fatal accident, of course, that Tesla is pushing the tech too far given its ability and maturity.

That said, I don't believe in the 10 year setback. Autonomous is coming. It remains to be seen which approach gets there first.

I believe Audi is a likely leader, Level 3 "read a book" possibly this year.

 

[McRat]

Just for the record... that's not exactly correct. What I said was "Betting the Caddy won't be slamming on the brakes at overpasses and drifting out of its lane." I am confident that Cadillac will thoroughly test its software before releasing it rather than pushing out a half-baked product, labeling it as beta, and leaving it up to customers to find all the shortcomings on their own. Time will tell whether Supercruise is functionally better, but I do believe from a reliability standpoint (for what it does) it will do it in a safer and more predictable way than what we have now with AP2. YMMV!

I can tell you the AEB in Caddys is very good. Certainly not beta, perhaps the world's leader in AEB technology since it also uses thermal imaging in the system, not just multiple radar and cameras. The ACC is also high quality.