It isn't fully autonomous yet. For any future system sold as level 5 (or level 4?), I agree with your contention -- the manufacturer of the level 5 autonomous system is the one who bears primary liability and therefore should insure. "FSD" isn't even level 3.

(Though, there is still an element of owner/operator maintenance for level 4/5 vehicles -- e.g., if the owner fails to replace tires below 4/32", continues to operate the vehicle, and it causes an injury, that is partially the owner/operator's fault.)

Generally speaking, liability for a thing falls on the owner/operator. That person can sue the manufacturer to recover the damages if they want. At some point, I expect it to become somewhat routine for insurures to pay out, then sue the manufacturer to recover.

Because that's the law of the land currently.

The product you buy is called "FSD Supervised". It clearly states you're liable and must supervise the system.

I don't think there's law that would allow Tesla (or anyone else) to sell a passenger car with unsupervised system.

If you take Waymo or Tesla Robotaxi in Austin, you are not liable for accidents, Google or Tesla is.

That's because they operate on limited state laws that allow them to provide such service but the law doesn't allow selling such cars to people.

That's changing. Quite likely this year we will have federal law that will allow selling cars with fully unsupervised self-driving, in which case the insurance/liability will obviously land on the maker of the system, not person present in the car.

Seems like the role of the human operator in the age of AI is to be the entity they can throw in jail if the machine fails (e.g. driver, pilot)

Because the operator is liable? Tesla as a company isn't driving the car, it's a ML model running on something like HW4 on bare metal in the car itself. Would that make the silicon die legally liable?

Not an expert here, but I recall reading that certain European countries (Spain???) allow liability to be put on the autonomous driving system, not the person in the car. Does anyone know more about this?

It’s because you bought it. Don’t buy it if you don’t want to insure.

Tesla is cooperating with Lemonade on this by providing them necessary user driving data.

If Tesla didn't want Lemonade to provide this, they could block them.

Strategically, Tesla doesn't want to be an insurer. They started the insurance product years ago, before Lemonade also offered this, to make FSD more attractive to buyers.

But the expansion stalled, maybe because the state bureaucracy or maybe because Tesla shifted priority to other things.

In conclusion: Tesla is happy that Lemonade offers this. It makes Tesla cars more attractive to buyers without Tesla doing the work of starting an insurance company in every state.

You'd be very surprised. Distribution works wonders. You could have a large carrier taking over Tesla's own vehicles. The difference then would be loss ratios on the data collection, like does LIDAR data really beat Progressive Snapshot?

I would be surprised if that was what they were actually looking at. They are an established insurance company with their own data and the actuaries to analyze it. I can't imagine them doing this without at least validating a substantial drop in claims relating to FSD capable cars.

Now that they are offering this program, they should start getting much better data by being able to correlate claims with actual FSD usage. They might be viewing this program partially as a data acquisition project to help them insure autonomous vehicles more broadly in the future.

> "Tesla’s data shows that Full Self-Driving miles are twice as safe as manual driving"

Teslas only do FSD on motorways where you tend to have far fewer accidents per mile.

Also, they switch to manual driving if they can't cope, and because the driver isn't paying attention this usually results in a crash. But hey, it's in manual driving, not FSD, so they get to claim FSD is safer.

FSD is not and never will be safer than a human driver.

For the vast majority of people who own a car, continuing to own the car will remain the better deal. Most people need their car during "rush hour", so there isn't any savings from sharing, and worse some people have "high standards" and so will demand the rental be a clean car nicer than you would accept - thus raising the costs (particularly if you drive used cars) Any remaining argument for a shared car dies when you realize that you can leave your things in the car, and you never have to wait.

For the rest - many of them live in a place where not enough others will follow the same system and so they will be forced to own a car just like today. If you live in a not dense area but still manage to walk/bike almost everywhere (as I do), renting a car is on paper cheaper the few times when you need a car - but in practice you don't know about that need several weeks in advance and so they don't have one they can rent to you. Even if you know you will need the car weeks in advance, sometimes they don't have one when you arrive.

If you live in a very dense area such that you almost regularly use transit (but sometimes walk, bike), but need a car for something a few times per year, then not owning a car makes sense. In this case the density means shared cars can be a viable business model despite not being used very much.

In short what you say sound insightful, but reality of how cars are used means it won't happen for most car owners.

Imagine this for a whole neighborhood! Maybe it'd be more efficient for the transport to come at regular intervals though. And while we're at it, let's pick up other people along the way, you'll need a bigger vehicle though, perhaps bus-sized...

Half-jokes aside, if you don't own it, you'll end up paying more to the robotaxi company than you would have paid to own the car. This is all but guaranteed based on all SaaS services so far.

This is the nightmare scenario for me. A forever subscription for the usage of a car.

Subscription for self driving will almost be a given with so many bad actors in tech nowadays, but never even being allowed to own the car is even worse.

[flagged]

So does AEB in any modern car.

Just curious about this, this was Lemonade's integrated insurance to the Tesla right? How's Geico like for you? Probably just fine right? Any differences?

No it does not. A 50% discount and the insurance still having industry average profit, or at least being profitable at all, would tell you that. Selling at a loss does not indicate your costs are actually lower. You need to wait until we learn if it is actually at a loss.

We don't know if 50% makes it actually cheaper than other car insurance companies, or the coverage is comparable, or if they have comparable service. Or if they sell your location information to marketers.

Assuming this discount is offered broadly and indefinitely. Otherwise these might just be marketing dollars.

Or a price hike if the fleet API tattles on you for negative driving behaviors.

It may not be on the marketing copy but it’s almost certainly present in the contract.

It seems the answer is yes. From their web site:

> Fair prices, based on how you drive [...] Get a discount, and earn a lower premium as you drive better.

I've had their Home Insurance since they started up and grabbed their car insurance a couple years ago. Competitive price, excellent customer service, no notes.

I think there's a narrow unregulated space where this could be true. I'm exercising my creativity trying to imagine it - where automations are built with the outcome of obscured responsibility in mind. And I could understand profit as a possible driving factor for that outcome.

As an extreme end of a spectrum example, there's been worry and debate for decades over automating military capabilities to the point where it becomes "push button to win war". There used to be, and hopefully still is, lots of restraint towards heading in that direction - in recognition of the need for ethics validation in automated judgements. The topic comes up now and then around Tesla's, and impossible decisions that FSD will have to make.

So at a certain point, and it may be right around the point of serious physical harm, the design decision to have or not have human-in-the-middle accountability seems to run into ethical constraints. In reality it's the ruthless bottom line focused corps - that don't seem to be the norm, but may have an outsized impact - that actually push up against ethical constraints. But even then, I would be wary as an executive documenting a decision to disregard potential harms at one of them shops. That line is being tested, but it's still there.

In my actual experience with automations, they've always been derived from laziness / reducing effort for everyone, or "because we can", and sometimes a need to reduce human error.

You're not making any sense. In terms of civil liability, fault is attached to the vehicle regardless of what autonomous systems might have been in use at the time of a collision.

One might imagine that lower courts won’t determine fault, one would be wrong.

> and then one sees that the primary purpose of most automations is to obscure responsibility.

Are you saying that the investments in FSD by tesla have been with the goal of letting drivers get a way with accidents? The law is black and white