> Does anyone have good data on comparing deaths per mile driven between various auto brands?

Not really. Most of the data is based on comparing all sorts of cars against all sorts of other cars.

Even ignoring how many years other brands have been making cars compared to Tesla, car models have very specific demographics they cater to.

Take the Corvette for example, one of the highest deaths per mile, and it's not because the car is inherently unsafe... it just attracts a very specific type of buyer with a very specific purpose in driving.

-

Adding in Autopilot makes the comparison even more hopeless since it works best in the absolute easiest situation (basic highways) doesn't work in the hardest ("stroads") and limits speeding.

You could try comparing only cars that have AEB/LKA/etc. but no one seems to have done that (Tesla certainly won't since it doesn't sell their narrative)

FSD would be slightly more interesting, but again, will work best on highways (so people will enable it there more) and also Tesla hasn't shared FSD numbers separately

>Does anyone have good data on comparing deaths per mile driven between various auto brands?

Hopefully this evaluation will force Tesla to share the data as the other companies, hopefully the army of fanboys have nothing against transparency.

USA drives 3.225 trillion miles per year, and kills 38000 people, that's about 1 death per 100 million miles. According to the anti Tesla site above, Tesla cars kill 1 person per 200 million miles. That's not accounting for deaths from pollution though

It is easy to use statistics bad. For example you can compare all kind of drivers vs Tesla drivers , you don't have teens driving Teslas so stats are not the same.

You can compare all cars vs Tesla and you still have bad stats since Tesla are new and expensive cars, you can't compare with old cheap cars.

You can compare autopilot vs human drivers and is still bad stats because most of autopilot miles are on good conditions on big hihgways so you would need to compare on the exact same roads.

TL:DR you need stats for the similar driver population, similar car prices and age and drive assist, similar roads. The TFA asks for more transparency so we would get more data and maybe we get better stats or more bad statistics, that we need to debug.

Do all the other car companies provide those stats?

No because:

1 they are not alpha/beta testing on public roads

2 not all cars track you, my brother Ford won't call to the homebase with crash info when he hit a dog.

So what's the relevance of the site, that's for non self driving Teslas.

Why would the model/price/class of the car affect the crash rate of a self driving car?

They do not sell a self-drivibg feature

That's....honestly bad.

1 death in 100 million miles driven? that's amazingly good!

Since Tesla is a very young and upscale brand deaths per mile would be very misleading. You need to also account for price range, car type, and age of the vehicle.

There is quite the difference getting into an accident with a 20 year old Honda Accord and a 3 year old Model 3.

Ok, still, try to compare it to equivalent brands and socio-economic groups.

My point is that anti Tesla self driving sentiment is almost exclusively driven by emotion, not hard data.

Show me the data that justifies why I need to be afraid of public FSD.

It’s a new technology which no one else is testing in the real world with actual drivers at such a scale. Onus is on Tesla to provide data for others to verify the claim that it’s safe.

Isn't it a catch 22? How do you provide the data that it is safe without testing it to collect the data?

It would be but Tesla charged ahead and put its vehicles on the road.

Well yes, that is my point. How else would they ever acquire that data if they didn't charge ahead? You are asking them to accomplish an impossible task to release the data without having a way to acquire it.

I don't think this is a good faith answer; feels more like "I don't have a good answer for why this is a problem, so I'm going to punt it onto you to convince me"

Disagree. It is a good faith answer which states standard practice for safety standards.

Burden of proof is on Tesla to show that FSD is safe, not on the DMV to prove that it isn't.

You are welcome to argue that this principle is in need of reform. People may or may not agree with you. The discussion might be interesting.

But like it or not, the underlying principle for an exception to an established safety principle is that the burden of proof is on Tesla.

The only ones who have the hard data are Tesla and you are expecting somebody else to provide it. Are you serious?

It is well-known that Teslas auto-brake-check cars at highway speeds (not with that intention of course, but same outcome). This is behavior for which non-AI drivers get arrested, for good reason.

That's enough reason to be very wary of Teslas on the road. There are others, but even one is enough.

> Show me the data that justifies why I need to be afraid of public FSD.

AFAIK they have mostly resolved the 'phantom braking' issue now.

On an absolute basis, Tesla has more self-driving-related deaths than every other automaker in the world, combined. That doesn't scare you?

On a relative basis, it's even scarier: scaled up to Toyota or GM numbers (in terms of vehicles sold), hundreds of people would be dying each year from self-driving crashes.

A few hundred people have died on a 5km stretch of road near me in the last couple of years. None of them driving Teslas. Most of them drunk or driving far too fast (50km/h is safe, people often drive 150km/h because shortcut and the stretch literally drives like a race course).

That sounds a lot better than the tens of thousands that die with humans at the wheel.

Not if you start comparing against miles driven it doesn't.

So far FSD beta has killed 0 people, and other FSD software has killed 1 person. So they seem to be ahead.

3000 people die every day from Human Driven Cars

Until teslas don't require constant human attention and frequent interventions, this comparison doesn't matter. A human is still technically in control of the car.

So they should work on doing that as soon as possible?

But aren’t 10s of thousands of people dying from self-driving crashes?

https://www.tesladeaths.com/miles.html https://www.tesladeaths.com/vehicleyears.html

Tesla doesn't release total yearly mileage data anymore so there isn't much to work with. There is some IIHS data and from that Tesla models do worse than comparable vehicles without a doubt (https://www.tesladeaths.com/index.html#FAQ).

And also road setting. Highways (where self driving tech is most commonly used) are the safest per mile.

Tesla has to provide data that its technology works, not the other way around. Public doesn't have to prove or provide data for shit.

Just like drug companies have to prove that their medicines work before FDA approves them to be provided to patients, similarly Tesla has to provide data and prove that it works before it should have been unleashed on public roads since both, drugs and FSD/autopilot, have the capacity to kill if they don't work as designed/promised.

I've found this series illuminating. https://blog.piekniewski.info/2020/03/01/av-safety-myths-and...

Absolute data points like this are pretty much useless, what would be more helpful is a percentage of deaths compared to Teslas out on the road vs. other automakers' cars. Still though, there's so many confounders (weather, types/subsets of customers, etc) that would make it nearly impossible to use this data to judge whether Teslas are safer or more dangerous than Car X or Car Y.

> what would be more helpful is a percentage of deaths compared to Teslas out on the road vs. other automakers' cars

...yeah, that's exactly what I asked for in the first sentence.

Right, but what I'm saying is that that still wouldn't paint the entire picture, there's too many overlapping variables and too few Tesla vehicles relative to market share of other brands to make any sort of sound or critical judgment of their safety.

I'm not defending Tesla here, just stating the obvious. Not a fan of their cars for quite a few reasons

The problem is deaths are so infrequent that it takes many years to get a statistically representative sample. Perhaps at-fault collisions per mile would be a better metric?

There's data on the other parts of the website. Use the links.

> Does human life have intrinsic value?

Yes? I don't understand your point.

We have limited problem solving resources, which means we need to be smart about how we use those resources to solve the problems that legitimately have the biggest impact on human suffering.

Getting bogged down on issues that are negligible in a relativistic sense, but trigger greater emotional response, is a recipe for failure. See the anti nuclear power movement for example, https://ourworldindata.org/safest-sources-of-energy

> We have limited problem solving resources, which means we need to be smart about how we use those resources to solve the problems that legitimately have the biggest impact on human suffering.

Sounds great in principle, but how do self-driving cars fit into that picture?

It fits because humans are objectively shit drives who cause massive amounts of death, suffering, and property damage through car crashes all the time.

Unless there's hard data that FSD is worse than the status quo, attempts to curtail it's use is not only a waste of time, but potentially actively contributing to unnecessary suffering if in fact it's contributing to fewer accidents.

Whatever Tesla is offering, it's not even remotely full self driving. It can't handle much beyond a highway on a sunny day, and even then, they have an unfortunate habit of crashing into stopped emergency vehicles.

Reduction of total cars on the road, in favor of competent mass transit, would do much more to actually reduce suffering. But Musk wants money, not improved quality of life for those who can't afford Teslas. This isn't some rinkydink charity, it's a fat slice of a trillion dollar industry.

Why should the burden of proof be on the critics? Carmakers shouldn't be able to unleash experimental technology that jeopardizes the lives of pedestrians without credible evidence that it is at least as safe as human drivers

>It fits because humans are objectively shit drives who cause massive amounts of death,

Why not maybe don't allow bad drivers to drive at all? Like don't give a license without 40 hours of driving school with a professor, have a test with traffic rules, then have a final test with an official. Have the drivers provide medical proof they are capable, Each time a driver does something dangerous like speeding remove his driving license for a few months, etc. Btw don't give driving licenses to kids could also work.

It is not like we the ONLY solution is AI driven cars, there are a lot of more cheaper solutions, even tech ones like identifying drunk drivers, speed limiting cars to match the road limits etc.

The trouble is it's still quite hard to model who will have a collision next, because collisions each year are still far fewer than the number of drivers. Although from a social standpoint they are too frequent, from a purely statistical point of view they are rare - in the sense of being the sort of event you might model with a poisson distribution. Sure, driving history can give an indication, but ultimately anyone could have one.

This is why when I had over 15 years of perfect driving record I still had to pay £350 in insurance. And the following year I caused a collision (fortunately without injuries). We are all human and fallible.

Sure but statistics show some things, like drunk driving is bad so is AI the ONLY solution to drunk driving? Speeding is also aproblem is AI the ONLy solution to speeding? Texting or losing focus is a problem is Ai driving the ONLY solution?

No, there is more then 1 solution , tech can be used to help with preventing drunk driving, speeding and losing focus. Tech could be used for drive assist to help on tricky situation. Reducing the mass of cars could also help,

IMO the idea that AI driving is the only solution and we shouyld

1 do nothing else and wait for AI

2 rush the AI deployment be

is really bad because

1 we should implement other solutions too

2 we should not rush AI deployment , Tesla fanboys will hate it when some non-Tesla would beta test on their roads and keep the stats secret.

Ah right, I think we're kind of on the same page. Full AI control of driving is likely 20 years away, and we should plan to have human drivers on the roads still for 30+ years. Anything we can do in the interim to improve our societal safety/utility tradeoff is good.

Cars are the primary cause of death for certain age groups! Technologies that meaningfully change the fatality rate (up or down) are of extreme importance to reducing human suffering

You can argue than yes, you can put a price on human life. It’s definitely the case for road safety and Covid-19.

https://www.aeaweb.org/articles?id=10.1257/jep.28.4.213

https://en.m.wikipedia.org/wiki/Value_of_life

In France today, it’s about 3 000 000€