Reuters: US safety agency contacts Tesla on robotaxi issues seen in online videos

[s4wrxttcs]

I don’t recall saying that.

Waymo doesn’t drive on highways, though. Correct? This inherently lowers risk of airbag deployment occurrences.

Here is the AI overview (based on available data) with Waymo

• Overall Injury Reduction: Waymo's vehicles exhibit a significantly lower rate of crashes resulting in any injury.
• Serious Injury Reduction: Studies suggest an 85% reduction in crashes involving suspected serious injuries.
• Reduced Vulnerable Road User Injuries: Waymo vehicles show a dramatic decrease in crashes injuring pedestrians (92% reduction) and cyclists/motorcyclists (82% reduction).
• Fatal Crash Statistics: As of March 2025, Waymo has been involved in one fatal crash, where the Waymo vehicle was stationary and not at fault. The fatality was the result of a high-speed collision initiated by another vehicle, in which the Waymo car was one of several vehicles struck

So I think the evidence is there that absolutely confirms that Waymo style autonomous driving absolutely reduces injuries and fatalities.

But, that's Waymo style of being utterly cautious with their roll out.

The times I've been in a Waymo were not mistake free. Like once it stopped well ahead of where I was, and had to walk to it without any perfectly good reason.

Despite the small mistakes I felt like the Waymo ride was superior than Uber so from now on I'm always going to ride a Waymo when possible. The biggest drawback was they avoided the highways but I believe they're started to go on the highways recently.

I'm not even going to talk about Tesla's Self-driving as I don't believe they meet my own personal minimum safety requirement on equipment needed. At best its small experiment on seeing whether vision only can work for L4 level driving.

Tesla self-driving is just not equivalent in any shape or form to Waymo's.

Sponsored

 

[s4wrxttcs]

I only care if the technology is safer than human drivers. The faster we can take humans out of the driving equation the better for everyone.

I think this ultimately comes down to a Trolley car problem.

Do we allow Tesla FSD vehicles to mow down people in the short term for the long term payoff of eventually saving lives?

Or do we choose the Waymo style rollout that's painfully slow where we wait decades for any payoff in terms of lives saved.

The answer is probably somewhere in the middle.

I do think all autonomous vehicles should be put through a pretty tough test before they're allowed to carry passengers.

 

[pamalabama]

I think this ultimately comes down to a Trolley car problem.

Do we allow Tesla FSD vehicles to mow down people in the short term for the long term payoff of eventually saving lives?

Or do we choose the Waymo style rollout that's painfully slow where we wait decades for any payoff in terms of lives saved.

The answer is probably somewhere in the middle.

I do think all autonomous vehicles should be put through a pretty tough test before they're allowed to carry passengers.

Have tesla had a record of mowing down people? No.

That's just ignorant talk

1 person was mowed down over 4 billion miles and that was an obvious mistake where the driver was not paying attention

We already have a payoff in terms of lives saved whereas waymo has not driven even the number of miles where we'd statistically see 1 death

These companies like waymo take forever to roll out (not for safety) but because they are much further behind than you think they are. They make you think they have something but they are nowhere near where what they advertise.

Mobileye has been letting people experience their self driving under controlled conditions and the driving is horrible. It makes tesla look ridiculously good.

 

[COdogman]

Have tesla had a record of mowing down people? No.

That's just ignorant talk

1 person was mowed down over 4 billion miles and that was an obvious mistake where the driver was not paying attention

We already have a payoff in terms of lives saved whereas waymo has not driven even the number of miles where we'd statistically see 1 death

These companies like waymo take forever to roll out (not for safety) but because they are much further behind than you think they are. They make you think they have something but they are nowhere near where what they advertise.

Mobileye has been letting people experience their self driving under controlled conditions and the driving is horrible. It makes tesla look ridiculously good.

Exactly how many lives are you claiming Tesla has "saved"?

 

[pamalabama]

Exactly how many lives are you claiming Tesla has "saved"?

1.26 deaths per 100 million vehicle miles traveled

4 billion miles would be 50.4 deaths.

If tesla vehicles are supposedly more dangerous than the average vehicle that would put the death rate higher than 50.4

So 2 deaths with FSD is statistically very good

Saving lives sounds wrong because 1.26 deaths for every 100 million miles is not a lot. So FSD could save a life and you'd not notice it. Because you can assume many of these deaths are freak accidents, tiredness or drunk driving. Someone could use FSD while tired or drunk and you'd never know

By comparison you drive tired or drunk with any other ADAS system and you're dead bcause the car doesn't see stationary vehicles or vulnerable road users. Or it cuts off because lane markings disappear or the curve is too steep

 

[COdogman]

1.26 deaths per 100 million vehicle miles traveled

4 billion miles would be 50.4 deaths.

If tesla vehicles are supposedly more dangerous than the average vehicle that would put the death rate higher than 50.4

So 2 deaths with FSD is statistically very good

Again, those are Tesla's claimed* numbers. There is absolutely zero supporting evidence that is accurate. It cannot be verified because they refuse to share any actual data. You already know that...

 

[s4wrxttcs]

Have tesla had a record of mowing down people? No.

That's just ignorant talk

1 person was mowed down over 4 billion miles and that was an obvious mistake where the driver was not paying attention

We already have a payoff in terms of lives saved whereas waymo has not driven even the number of miles where we'd statistically see 1 death

These companies like waymo take forever to roll out (not for safety) but because they are much further behind than you think they are. They make you think they have something but they are nowhere near where what they advertise.

Mobileye has been letting people experience their self driving under controlled conditions and the driving is horrible. It makes tesla look ridiculously good.

I think you missed what I was saying in your need to defend Tesla.

Tesla has the most aggressive plans for self-driving so I used Tesla as an example.

I don't know of any other North American based organization that has a more aggressive plan for self-driving.

So obviously the danger factor goes up.

But is that necessarily a bad thing? If it saves lives long term?

You need to stop defending Tesla and answer the hard questions.

 

[DuoRivians]

Again, those are Tesla's claimed* numbers. There is absolutely zero supporting evidence that is accurate. It cannot be verified because they refuse to share any actual data. You already know that...

Tesla notoriously turns off fsd right before an accident. Teslas also have the highest number of fatality rate of any auto brand.

https://www.roadandtrack.com/news/a...fatal-accident-rate-of-all-auto-brands-study/

 

[s4wrxttcs]

So 2 deaths with FSD is statistically very good

There are two FSD systems

FSD supervised - What all the miles are on
FSD unsupervised - Has very few miles.

We simply don't know how effective of a killing machine FSD unsupervised will be.

 

[pamalabama]

I think you missed what I was saying in your need to defend Tesla.

Tesla has the most aggressive plans for self-driving so I used Tesla as an example.

I don't know of any other North American based organization that has a more aggressive plan for self-driving.

So obviously the danger factor goes up.

But is that necessarily a bad thing? If it saves lives long term?

You need to stop defending Tesla and answer the hard questions.

Aggressive only because they are doing it at a scale that no one else is. Waymo doesn't have a lot of vehicles because they need to buy them and retrofit. They need to map and test. Tesla so far is only doing small scale robotaxi and large scale supervised.

it doesn't seem like tesla is anywhere near large scale supervised

 

[pamalabama]

Tesla notoriously turns off fsd right before an accident. Teslas also have the highest number of fatality rate of any auto brand.

https://www.roadandtrack.com/news/a...fatal-accident-rate-of-all-auto-brands-study/

Tesla includes any accident where adas or autopilot is engaged 5 seconds prior to the crash. it's part of the NHTSA requirement

That guy who flipped his car over 100% on his own. Tesla reported it because it was required

IIHS says that tesla model 3 and Y are below average in terms of fatalities. That aligns with a car that has both a high crash test rating and the world's best auto emergency braking. Hard to get in an accident when your car prevents it

 

[pamalabama]

Again, those are Tesla's claimed* numbers. There is absolutely zero supporting evidence that is accurate. It cannot be verified because they refuse to share any actual data. You already know that...

Tesla shares their data. It is in the FARS reporting system. All accidents are reported according to the standing general order.


Tesla is the only brand reporting their accidents

Tesla's data is purely referring to accidents and deaths. If someone is not in an accident or death what does it matter?

This data doesn't work because how many close calls do you see every day? Lots. Are they reported anywhere? No.

The only stats we have are for real verifiable accidents. We don't know about any close calls.

you can just turn on FSD in the heaviest rain and see how well it visualizes the road markings. It's freaky because it sees them better than you do

 

[COdogman]

Tesla shares their data. It is in the FARS reporting system. All accidents are reported according to the standing general order.


Tesla is the only brand reporting their accidents

Tesla's data is purely referring to accidents and deaths. If someone is not in an accident or death what does it matter?

This data doesn't work because how many close calls do you see every day? Lots. Are they reported anywhere? No.

The only stats we have are for real verifiable accidents. We don't know about any close calls.

you can just turn on FSD in the heaviest rain and see how well it visualizes the road markings. It's freaky because it sees them better than you do

Absolutely false and you know it because we have been through this EXACT discussion multiple times You have even repeatedly tried to send me to a 3rd party website due to the fact that Tesla refuses to release any worthwhile data.

Tesla releases a “safety report” that conveniently combines Autopilot and FSD data so that no one actually knows what the specific statistics are for FSD, despite their claims and yours. No raw data is provided or independently verified.

The funniest part is that whenever someone else combines AP and FSD stats, you whine about it. But it’s totally fine and dandy

 

[pamalabama]

Absolutely false and you know it because we have been through this EXACT discussion multiple times You have even repeatedly tried to send me to a 3rd party website due to the fact that Tesla refuses to release any worthwhile data.

Tesla releases a “safety report” that conveniently combines Autopilot and FSD data so that no one actually knows what the specific statistics are for FSD, despite their claims and yours. No raw data is provided or independently verified.

The funniest part is that whenever someone else combines AP and FSD stats, you whine about it. But it’s totally fine and dandy

The website tesladeaths gets the data from the same place it is reported. The FARS reporting system

You can also see tesla ADAS crashes under the standing general order crash data

you have to manually sift through it as many crashes are reported multiple times.

Tesla is actually the most transparent automaker in that regard as other automakers choose to not report their crashes. Rivian has to manually get the crash data and they don't remotely obtain it like tesla using the telematics control unit

Tesla reports their crashes because they have the advanced telematics capability to do it and they design their own systems. They could have conveniently designed their car in a way in which they don't have access to that data, but they did not. And they choose to report it because it is legally required.

Your ignorant logic is that tesla should report every mistake the car makes. If it is a mistake, how do you know about it? Car doesn't make mistakes on purpose

 

[COdogman]

The website tesladeaths gets the data from the same place it is reported. The FARS reporting system

You can also see tesla ADAS crashes under the standing general order crash data

you have to manually sift through it as many crashes are reported multiple times.

Tesla is actually the most transparent automaker in that regard as other automakers choose to not report their crashes. Rivian has to manually get the crash data and they don't remotely obtain it like tesla using the telematics control unit

Tesla reports their crashes because they have the advanced telematics capability to do it and they design their own systems. They could have conveniently designed their car in a way in which they don't have access to that data, but they did not. And they choose to report it because it is legally required.

Your ignorant logic is that tesla should report every mistake the car makes. If it is a mistake, how do you know about it? Car doesn't make mistakes on purpose

Sponsored