Tesla self-driving is so much safer, you get an insurance discount for using it

[SwampNut]

You don’t have to have it drive in the edge cases.

Crazy talk! It's all or nothing. If you use it at all, you are legally required to let it drive you off a cliff. Cost of doing business.

Sponsored

 

[COdogman]

Where can we see those statistics?

Where can we find the statistics proving it's safer than a human driver?

See how that works?

 

[therealcmj]

Where can we find the statistics proving it's safer than a human driver?

"Trust me bro!"
-Tesla

 

[UnsungZero_OldTimeAdMan]

As if Lemonade has a overwhelmingly positive reputation. Go ahead and Google search reviews. Just another outfit advertising low rates for sake of intake, but be as difficult as possible when it's time to payout.

ADAS is safer? Safer than shit drivers who are worse than it. ADAS, like any other tech, works until it doesn't. The question is will you be alert and ready to assume control when it doesn't. How many wrongful death suits does Tesla have to-date? were those drivers or "victims" alert and ready? Rhetorical, obviously. Waymo's Krafcik claim their system is better. But, better is not the same as perfect. They have had their own issues with fatalities. The fanbois think these systems are perfect. Reality points otherwise.

 

[CharonPDX]

Unironically yes. This is already a thing.

https://www.thedrive.com/news/tesla-insurance-charges-you-more-if-you-drive-at-night

Yep. Most of the "monitor your driving to give you a discount" services penalize you for driving at night.

I had Progressive, and they have three 'scores' - "Phone use", "Hard braking", and "speeding." I scored 100 in all three. And yet my 'overall score' was 69. of course, the big discount threshold was at 70.

Why was it so low? "unspecified reasons". I called in to find out, it was because I regularly drove at night.

 

[UnsungZero_OldTimeAdMan]

Yep. Most of the "monitor your driving to give you a discount" services penalize you for driving at night.

I had Progressive, and they have three 'scores' - "Phone use", "Hard braking", and "speeding." I scored 100 in all three. And yet my 'overall score' was 69. of course, the big discount threshold was at 70.

Why was it so low? "unspecified reasons". I called in to find out, it was because I regularly drove at night.

Yup, just another marketing scheme to lure in new customers (suckers?) and grow revenue.

 

[Luxus]

The issue I have with that is over time you will be lulled into a false sense of security. The SD does seem to work most of the time so it's easy to see how someone could come to think concerns are overblown. Asking someone to pay full attention to the road constantly while the car does the all work is a very big ask IMHO.

 

[renderpaz]

I'm very grateful to the overuse of AI by desperate startups, it makes it very easy to separate out the garbage.

That said I do fully expect ADAS/FSD use to dramatically reduce insurance rates once it is widespread. For now there is no AI solution for a moron in a 2005 F150 smashing into my truck.

 

[TexasBob]

Where can we see those statistics?

What is the frequency with which an driver must take over FSD and has a "critical disengagement"? 1,454 on FSD 14 late 2025 data, right? The very most optimistic version is once every 9,200 miles.

What does that tell you? That once every 1,454 - 9,200 miles (giving FSD every benefit of the doubt here) it would get into a collision - minor or major - if left to its own device.

What is the accident rate for a teen driver?

• 16-Year-Olds: ~34.5 crashes per 1 million miles (every 28,985 miles)
• 17-Year-Olds: ~20.2 crashes per 1 million miles (every 48,780 miles)
• 18-Year-Olds: ~13.8 crashes per 1 million miles (every 72,464 miles)
• Average: ~22.8 crashes per 1 million miles (every 43,859 miles)

So FSD is something between 4.8x and 30x more likely to have an accident than an average teenager. I am extrapolating that a drunken teenager is something between 5x - 30x more likely to crash than a regular teenager. IDK, maybe I should have said a tipsy teenager? A slightly stoned teenager?

 

[CharonPDX]

False, and I already posted the evidence. Did you read the thread or rush to reply?

Please provide a link to where you provided evidence that an automaker is shipping true self driving.

You posted one photo of one person's Tesla saying that it has driven ~5000 miles on FSD, that being 100% of their miles driven since updating to FSD 14.2.

That is not the same as "evidence that a carmaker is shipping true self driving."

Telsa is shipping level 2 driver assistance that they call "Full Self Driving - Supervised". The driver still has all legal responsibility for the actions of the vehicle, and must be ready to take over control with zero notice at any time. The fact that some people are successfully using it without intervention for decently long periods of time/distance is not evidence that the tech is truly ready for *ALL* users to use it unsupervised.

Which would be needed to meet everyone but a Tesla-stan's definition of "true self driving."

 

[s4wrxttcs]

Bottom line actual truth: The Model Y has a significantly higher fatal accident rate and nearly double the insurance cost of a comparable RAV4.

This is not necessarily at odds with having a 50% discount for every mile on FSD.

The take rate of FSD is actually pretty low so the statistic you posted was likely largely manual driving or using AP. In fact it could be that Tesla drivers are so terrible that FSD usage results in fewer crashes.

Plus a lot of fatality crashes are due to reckless driving. It's probably one of the worst cars to allow a teenage driver to drive since its so quick.

I say this as someone who thinks FSD is a TERRIBLE idea. As it gets better it actually gets more dangerous. The reason it gets more dangerous is the human in charge of overseeing it becomes more complacent.

Insurance doesn't really care about good ideas or bad ideas. They're essentially the barometer of the real truth. So I think the question is whether Lemonade is being honest with this offer. Or is it a bait and switch to attract new customers? Or does it come with additional restrictions. Maybe its less costly because FSD simply won't engage in bad conditions.

 

[s4wrxttcs]

Please provide a link to where you provided evidence that an automaker is shipping true self driving.

You posted one photo of one person's Tesla saying that it has driven ~5000 miles on FSD, that being 100% of their miles driven since updating to FSD 14.2.

That is not the same as "evidence that a carmaker is shipping true self driving."

Telsa is shipping level 2 driver assistance that they call "Full Self Driving - Supervised". The driver still has all legal responsibility for the actions of the vehicle, and must be ready to take over control with zero notice at any time. The fact that some people are successfully using it without intervention for decently long periods of time/distance is not evidence that the tech is truly ready for *ALL* users to use it unsupervised.

Which would be needed to meet everyone but a Tesla-stan's definition of "true self driving."

How is any of that relevant to Lemonades discount for FSD driven miles? FSD that is currently shipping in FSD Supervised so its an Level 2 ADAS system.

If it was L3 or L4 the FSD driven miles would be covered by the automaker or fleet owner.

They do use the autonomous language but its just marketing speak.

It's an interesting position to take for a insurance company as it seems contradictory to what they'd want. Somehow its cheaper to have a teenager (I think that's being generous to FSD) drive the car while be supervised by an adult than to have an adult do the driving?

I think the devil is likely in the details in whatever agreement they have with Tesla.

I also don't think we should be too quick to dismiss this as Rivian is ramping up their ADAS efforts as well. So in the future this might also apply to Rivian vehicles being driven with Driver+ but supervised.

 

[mkg3]

"Model S completed the 3,081-mile trip entirely without human intervention."

https://www.thedrive.com/news/a-tes...hlid=e176cc74b7a0d4dbb52649a02f0592e38a901ca5

Alex Roy and a team of independent autonomy experts just completed a Tesla FSD "Cannonball Run" with zero disengagements.

 

[CharonPDX]

"Model S completed the 3,081-mile trip entirely without human intervention."

https://www.thedrive.com/news/a-tes...hlid=e176cc74b7a0d4dbb52649a02f0592e38a901ca5

Alex Roy and a team of independent autonomy experts just completed a Tesla FSD "Cannonball Run" with zero disengagements.

Meanwhile, a route I take at least once a week, FSD tries to change *OUT* of the exit lane I need to be in after the line has become solid on a well-traveled major freeway intersection.

And going to my local (frequently full) Supercharger, where there is a right-turn lane, it tries to take the straight-ahead lane and turn right from it. Two blocks from the Supercharger - at the end of the freeway offramp for a major Interstate. A path driven by hundreds of Teslas daily.

The fact that these "I drove a long distance with no takeovers!" are news shows how infrequent they are.

Yes, someday the tech will be good enough to achieve full L4 or L5 autonomy. The problem is that Musk has been promising "later this year!" every year for over a decade now. And people keep believing it.

 

[jemenake]

Where can we find the statistics proving it's safer than a human driver?

See how that works?

If someone were to claim that FSD is “statistically safer” than humans (in a way meant to convince others or influence a debate), then, yeah, they should have to show which statistics led them to that conclusion. But that’s not what I’m seeing. I see an insurance company offering discounts for using FSD, so _they_ think that _their_ statistics show that, but they’re not trying to convince anybody of that notion. _They_ believe it and they’re willing to put their money where their beliefs are.

Sponsored