Self Driving Cars are Going Nowhere

Yes, it’s called Failed to Cross the Chasm! the fate of many, many, many innovations!

The Captain

3 Likes

I like the cell phone comparison because it illustrates that the tech (the cell phone) needed a road (cell bandwidth) in order to function much like cars of the early 20th century needed standardized roads for mass adoption.

There were a number of bills in the early 1900s that made the mass adoption of cars much much easier and without such, large areas of our country would likely have gone carless for decades.

The cars (the tech) was “perfected” long before the method of travel was. The same will likely the problem with SD tech. We don’t have standardized roads - even on US highways. Each state often will have slight variations in how they pave, mark, repair, etc. their roads. That lack of standardization (coupled with all the “black swan” events - like an actual swan in the road), leads me to conclude that - outside of a well defined corridors of transportation, full self-driving for personal transportation is a long way off.

1 Like

How old are you, Andy?

Bob,

Not that you are asking me, but I am 103.

1 Like

Well, clearly Tesla is proceeding on the basis - probably correctly - that none of this will happen.

First, the Tesla materials regarding Full Self-Driving (FSD) are loaded with disclaimers that you the driver has to maintain full attention and control over the vehicle. So if one decides to “go hands free,” Tesla has a good defense against liability.

Second, Tesla insurance almost certainly includes the possibility that some folks might misuse the FSD as part of their actuarial data that they use to calculate their premiums. So while FSD misuse might happen from time to time, the Tesla insurance should have the resources to cover the resulting injuries.

Third, it’s not especially hard to successfully insulate a legitimate company from corporate veil-piercing. I don’t know if Tesla is even trying to do that. They probably have separate special-purpose entities that are the insuring companies because of the way state insurance laws work, but they could just be doing that for compliance and running the actual insurance as a division of the main company. However, if they were trying to insulate the insurance division from the main company, they’ll be able to do it.

Insurance and legal issues won’t be much of a barrier to truly autonomous cars driving. There’s already a very well-established body of law dealing with what happens when accidents result from a defect in the vehicle if the vehicle is the cause of the accident, rather than the driver. And that works pretty well, despite the fact that supply chains for cars are already very diversified and dispersed with hundreds of parts manufacturers. If we ever get to the point where it is possible to make a car that is a true Level 4 or Level 5 AV, legal and insurance issues won’t pose much of a barrier to adoption.

1 Like

Blame technology?

A few years ago, one of those “history” shows stated that the first “peaked arch” was built in an obscure church in Spain 500 CE (IIRC). The next such arch was built 700 years later, due to the technique NOT being spread.

Hand printing, general lack of writing skills, etc.
Then the printing press, etc sped things up.

Today, if a person in Terre Haute succeeds at some “thing”, she reports it in a research preprint, or a popular science news site, and others in a dozen countries repeat her effort. The “knowledge” spreads ASAP around the globe.

As others have mentioned, more and more infrastructure is now in place, that further supports “things happening faster and faster”.

To bring this back to Tesla and FSD- every time a Tesla is charged at a Tesla charger, the car connects to “the Tesla Mothership”. Yes? Does that connection include transfer of diagnostics? Ie does the car get a “health check up” at every charge, and if a problem is developing, a service call is scheduled?

The FSD beta tester cars certainly report back quickly? Every day?

:alien:
ralph

2 Likes

[albaby1]

First, the Tesla materials regarding Full Self-Driving (FSD) are loaded with disclaimers that you the driver has to maintain full attention and control over the vehicle. So if one decides to “go hands free,” Tesla has a good defense against liability.

Exactly what I expected. Posted upthread:

The probabilities that all the insurance companies will agree to keep the liability on the driver are pretty good. So, the companies advertise “autonomous” cars, with a big asterisk, and a footnote that says the driver is responsible for remaining awake and alert, and constantly monitoring the situation. Yeah, that’s the ticket, claim autonomy, but say the driver still bears responsibility for anticipating and pro-actively correcting errors by the system.

So all the advertising about “autonomous” operation is a fraud. But we can’t “burden” the “JCs” with making them tell the truth, now can we?

Steve

1 Like

I am 60 years old DrBob. But my Grandmother had 14 children and my Mother was the youngest of the 14. She is 88 years old and my Grandmother lived to 101. I have pictures of a Flathead Indian Chief on a horse in front of my Grandfathers Ford Garage.

Andy

4 Likes

No one’s advertising “autonomous” operation, AFAIK. Tesla gets close to the line with labelling its suite of options “Full Self Driving,” even though that suite still does not contain all the elements necessary to be fully autonomous. Still, though, they have a lot of disclaimers about it in their materials.

Once there actually is Level 4 or 5 autonomy, though, liability will be fixed on whoever manufactured the car (in situations where a human isn’t driving). This isn’t a tough situation. Accidents today are sometimes caused by a problem with the car, rather than human error (a part fails, a tire blows out, something is installed incorrectly, etc.) - and the legal and insurance systems deal with that quite easily.

Automobiles - like nearly all consumer products - are subject to strict liability. Meaning that if the vehicle “fails” in some way (a manufacturing or design defect that causes damage), the entity that sold the vehicle is liable for that damage - regardless of whether it was their fault, or the fault of one of their suppliers or someone else in the production chain. If the AI screws up and causes an accident, the company that sold you the AI - meaning the company who sold you the car - will be held liability. It’s not hard to deal with damages caused by something other than the driver; autonomy just expands the universe of accidents that are “car-caused” rather than “owner-caused.”

If a subcontractor or supplier or some other third party was the one who caused the defect, the injured party(ies) can sue them, too - and there will be joint and severable liability of the various responsible defendants. That’s a complicated concept with lots of nuances, but the TL;DR version is the plaintiff(s) get to sue all of them, and then they get to fight among themselves later in their own lawsuits how liability should be allocated among them. But again, this happens all the time in our current system when there are accidents caused by vehicle problems other than driver error.

2 Likes

Given that, what manufacturer in their right mind actually wants to sell an FSD car then? Why expose yourself to that liability? Only if you are making money off the FSD capability (like a taxi company, or delivery company) does that even remotely make sense.

This is why I think self driving passenger cars is a ridiculous goal.

2 Likes

Which is why Musk said that once they have developed L5, Tesla will stop selling cars. The real money is in Transportation As A Service, which is what Barra (head of GM) also said.

1 Like

It doesn’t work like this. Tesla cars contain cellular data service and even if you never go to a Supercharger and / or only ever plug in to recharge with a common 120v outlet Tesla gets updates from the car and the car downloads updates. You can also opt to use WiFi, for example when parked at home or at a Tesla service center. There are some options for what data is sent out.

Mike

2 Likes

Hi Albaby1 - I haven’t been here for decades (my original user name was soldat4800) - with the news that Musk is going to start charging people for the blue check marks on Twitter it made me think of the Fool trying to monetize the message boards - apparently it didn’t go well. How long did they try to charge people to post? How did they climb back down from that? And do you have any idea how many users they lost along the way?

Thanks

1 Like

Al at least quote me in full. Don’t always have a legal mind. :rofl: :rofl: :rofl:

Al, I have no legal mind…zero you would normally win but you over state things as fact.

The problem is if Tesla insurance goes bankrupt for whatever reason in the future…then there are millions of uninsured Teslas. Now the liability going forward for accidents reverts to the manufacturer if there is no driver and no insurer. It becomes a defective product.

They all would. For the same reason that they sell cars with cruise control, and adaptive cruise control. Because consumers find the features useful, and will pay enough for the feature to cover the costs of providing the feature and whatever small amount is required to cover the risk of failure.

Again - this is not a new issue. It applies to any advanced feature that might affect safety.
Say you’re going to sell a car with adaptive cruise control. It might fail! If it fails, there might be damages. And if there are damages, the manufacturer will probably be liable. But consumers will pay more for a car with adaptive cruise control than one without it. So you just charge enough for that option to cover your costs of the feature and some amount to cover the likely cost of damage payments associated with that feature.

Presumably, the cost of damages associated with AI failure will be lower than the cost of damages associated with driver failure. Else the AI wouldn’t be permitted to drive the car. Which means presumably the cost of “insuring” the AI should be less than insuring the cost of the driver. Which means that the price the consumer is willing to pay for the AI will be high enough to cover the cost of that AI insurance - the consumer will save more in foregone “driver” insurance than they spend in “AI” insurance.

The liability involved here is not mind-bogglingly large. Rounding heavily to simplify the math, the total auto insurance industry is about $300 billion, and there are about 300 million light vehicles in the U.S. - roughly $1,000 per vehicle per year. Presumably that will be far, far lower for an Level 5 AI vehicle.

You can either tack a few extra thousand dollars onto the vehicle price upfront, or require an annual or monthly “insurance” charge to keep the AV feature active on the vehicle. Either way, just like with any other feature, there’s certainly a price that you can charge that will make you a profit net of your damage liabilities.

2 Likes

God man none of them do. They work right now as FSD better than human drivers. None of them but Tesla is playing with it. The product is defective from day one. Just blame the driver or the manufacturer?

This is both true and not a new situation. Ironically, this is probably less of a problem if Tesla spins out a truly independent insurance provider. Insurance is tightly regulated by the states. To get licensed to sell auto insurance, the Tesla insurance entity will have to demonstrate adequate capital reserves and sufficient premiums to make sure that they retain the capacity to pay out claims. Again, that system isn’t perfect, but also isn’t a new problem - sometimes insurance companies fail, with bad results. Regulators try to keep it from happening.

If the Tesla insurer went out of business, though, injured parties would still have recourse against Tesla itself for any accidents that were caused by the AI - just as they do now for injuries or damages caused by a vehicle that malfunctions.

1 Like

The full quote makes a big difference.

I write concisely. It is easy enough to quote me. I am not rambling.

This would be more inclusive because usually independent insurers do a lot of the lifting.

Musk does have a brilliant goal here. If the FSD cars get so excellent there are no accidents or miniscule numbers of accidents then there is no need for insurance.

1 Like

We’ve had this conversation before. No one has developed Level 5 autonomy that is safer than a human driver on U.S. streets over the full range of driving situations. Tesla doesn’t have a Level 5 autonomy system - their system isn’t Level 5 being sold with disclaimers, it is literally not a Level 5 system yet. There are a few Level 3 systems out there being tried in Europe. There are a few Level 4 systems that are being piloted in California. But no one has a Level 5 system. It doesn’t exist yet, despite what your cousin a few years out of MIT (or whoever it was) might have told you.

3 Likes