Cybercabs have a 12.5 times higher crash rate per mile driven than humans. Yet Musk tells the Tesla faithful and media that Tesla is "safer than himans". The data doesn't lie. Tesla has been caught numerous times hiding accident data and blatantly hiding accidents and injuries to OSHA. Then add in he likes to pick fights with other ceos but when they have a snappy come back Musk doesn't think it's so amusing.
Tesla Basic Autopilot Suite Now a Paid Subscription Only.
- Thread starter E365
- Start date
If you have a link to your information that would be great, doing a search and I am not getting the same response as your post. I am seeing Waymo involved with 91% fewer crashes than humans and Tesla possibly more safer than that. I think self driving will and possibly is being much safer than humans but of course would like to see your source. Im talking serious injuries.
https://gvwire.com/2025/12/03/dont-fear-self-driving-cars-the-data-shows-they-save-lives/
https://www.nextbigfuture.com/2025/...-of-safe-robotaxi-is-4x-safer-than-waymo.html
To expand on above AV vehicles have to reports literally every little incident, bump and scrape. I am now seeing reports of what you maybe talking about. ITs a tough statistic to narrow down until more data. Serious accident vs a bump. I will say though, after reading it you might have a point as right now, much data from the AVs in cities where speeds are slower.
Last edited:
I have a 2021 Model 3 and have as much automatic driving stuff turned off as I can.
I turned off the autosteer the first time it refused to let me make a quick lane change without signalling. In that case there was simply no time to signal. So now the steering wheel only vibrates if I make an un-signalled lane change or touch a traffic line. And that's a good feature.
I mostly like the cruise control that will slow to maintain a gap with the vehicle in front. The part I don't like is when a vehicle in front is just going a bit slower or gradually slowing and I suddenly realize I'm traveling below my intended speed.
The part I really dislike is the unnecessary "braking" (some call it shadowbraking) when a vehicle crosses my path (often) some distance ahead. It's been known to "brake" for a shadow on the road, or an overpass too. It often "braked" at a the same slightly complicated entrance to the highway but hasn't done that recently. Soon after we got the car we were driving on a rural highway during a hard cross wind and it kept "braking" for waving grass at the side of the roadway too. It doesn't actually brake it just rapidly decelerates by regenerating but it feels the same and has much the same effect as braking. I suppose it could save my bacon by braking for a real danger some day, but I think it's more likely to get me rear ended or put me out of control on a slippery road.
I tried "full self driving" a couple of times over a few minutes and didn't like it. I might try it again if I was going on a really long highway trip where the monthly subsription plan (for a single month) could work out well.
I turned off the autosteer the first time it refused to let me make a quick lane change without signalling. In that case there was simply no time to signal. So now the steering wheel only vibrates if I make an un-signalled lane change or touch a traffic line. And that's a good feature.
I mostly like the cruise control that will slow to maintain a gap with the vehicle in front. The part I don't like is when a vehicle in front is just going a bit slower or gradually slowing and I suddenly realize I'm traveling below my intended speed.
The part I really dislike is the unnecessary "braking" (some call it shadowbraking) when a vehicle crosses my path (often) some distance ahead. It's been known to "brake" for a shadow on the road, or an overpass too. It often "braked" at a the same slightly complicated entrance to the highway but hasn't done that recently. Soon after we got the car we were driving on a rural highway during a hard cross wind and it kept "braking" for waving grass at the side of the roadway too. It doesn't actually brake it just rapidly decelerates by regenerating but it feels the same and has much the same effect as braking. I suppose it could save my bacon by braking for a real danger some day, but I think it's more likely to get me rear ended or put me out of control on a slippery road.
I tried "full self driving" a couple of times over a few minutes and didn't like it. I might try it again if I was going on a really long highway trip where the monthly subsription plan (for a single month) could work out well.
Electrec
Tesla hides accident data
"
Elektrek also crunched some numbers based on data released by Tesla last month and estimated that the Tesla Robotaxis are involved in a crash for every 40,000 miles they drive. For comparison, the publication reported, cars driven by humans crash about once every 500,000 miles, meaning the Robotaxis so far have crashed 12.5 times more frequently than human-driven cars.
All of the Robotaxi crashes so far have occurred with human safety monitors—who have been trained to take control of the car in the event of a software error—present in the vehicles."
Tesla has "unsupervised " cybercabs. Musk didn't disclose that they are still being tailed by chase vehicles they aren't unsupervised.
People are calling out Teslas garbage math and false claims. Last year Musk even said that Tesla needs far fewer miles between disengagements.
According to Forbes
"These two events have caused many to declare that Tesla has “done it" but a deeper examination suggests this is quite unlikely, and in fact, it’s fairly consistent with Tesla still needing to improve by a factor of at least 100 to 1,000 times to have actually done it. That difference is quite extreme, almost incredible, but the principles are fairly simple."
We didn't break the sound barrier two months after the Wright Brothers flew. It took decades.
Look up yahoo or various other links. Tesla was caught red handed leading to a huge fine due to hiding accident data.
"
"Tesla said it didn’t have key data in a fatal crash. Then a hacker found it."
Should have been a stop sale and jail time.Yeah I get this but the word "crash" includes even bumper scrapes and very minor incidents that typically would not be recorded by the public. So the word "crash" is an overstatement compared to what the public perceives as a crash.
Im completely neutral on MANY subjects but I am an advocate of reading between the lines of sensationalist so called news/media stores. Media clarity which is almost none existence in today's social media blasts.
Give me numbers of serious injuries and accidents of AV's vs the public drivers. Then I am good, but until then ....
So from June 2025 to Nov 2025 the entire fleet had 7 incidences with no injuries. True, in this report is says Waymo is better, then in other reports it says Tesla is better. Thing is we dont know the amount of reported public crashes since so many or not reported.
I mean one of these "crashes" is with an animal https://electrek.co/2025/11/17/tesla-robotaxi-had-3-more-crashes-now-7-total/
I wonder how many animal crashes are reported when a human is driving? Another a bicyclist... but what were the circumstances? Did the bicyclist run out into the road? Others were "fixed objects" ok, what was that? A construction cone in the road?
All I am saying is I dont trust "media" reports without facts, the stories in today's world are typically biased reporters. It makes a great eye catching click to read about 7 "crashes" until the public finds out what is a "crash" is all I am saying.
Last edited:
OK Forbes ran the accident rates with real compiled data.
"A recent study on the “Most Dangerous Cars on the Road” by iSeeCars reveals a sharper picture of the Tesla accident rate compared to other cars. The analysis reports Tesla vehicles with a fatal crash rate of 5.6 deaths per billion miles driven, compared to a national average of 2.8 deaths per billion miles."
Forbes compiled data by make. February 2025 is the most recent but note the amount of accidents Teslas had per 1,000 people went up.
Vehicle accident rates.
Robotaxis aren’t driving highway miles. It’s all city=more accidents.
These are not AV's which is what this thread is about. But even then, give me the rates of all muscle cars NOT national averages of everything. Makes a great headline but give me muscle cars, not a generic average of everything on the road. It's almost silly this kind of reporting.
Let's compare Tesla death rates to these vehicles. This would be a fair comparison mostly, and even all these cars are not high horsepower like a Tesla.
"INDIANAPOLIS — The Insurance Institute for Highway Safety, famous for their impact tests and accident research, recently found muscle cars to be among the deadliest vehicles. Six of the 21 vehicles with the highest driver death rates are variants of the Chevy Camaro, Dodge Challenger/Charger, and Ford Mustang."
https://www.wthr.com/article/money/...tang/531-f0229ab1-6a08-4e44-8290-ffef8626c195
Not debating, just saying we have to compare Apples with Apples. Tesla is a very high horsepower vehicle and MANY including in this forum enjoy the exhilarating acceleration and speed. Much like the muscle cars they compare them to.
So lets compare the death numbers of muscle cars to Tesla not to a mix of cars like Toyota Corolla's, Ford Fiestas and Chevy Suburbans
Last edited:
I see it in a slightly different way. I'm sure money is a big factor but based on what I can tell their current FSD hardware is not really ever going to be up to date for the latest FSD. They likely have to keep upgrading the hardware like camera, GPU, adding ram or flash memory, and maybe one day add lidar if they are cheap enough. Being on a subscription they can just tell people it is not supported anymore whereas selling them and say it will get better, then suddenly say they can't because they found a bug or tech limitation, would trigger class action lawsuit.
If I lease a Tesla "maybe" I will try it for a couple months here and there when I am doing road trip, but I wouldn't pay $100 a month to just commute 15 days a month when I can just drive each way manually for 20 mins following other cars in traffic. Insurance company likely won't give you a discount unless it is permanent. Nobody will notify insurance whether FSD subscription is on or off by the day and insurance won't charge you by the day of it on/off.
You can always flip the logics around to your benefit and enable "on demand" like Amazon Prime subscription I guess?
If I lease a Tesla "maybe" I will try it for a couple months here and there when I am doing road trip, but I wouldn't pay $100 a month to just commute 15 days a month when I can just drive each way manually for 20 mins following other cars in traffic. Insurance company likely won't give you a discount unless it is permanent. Nobody will notify insurance whether FSD subscription is on or off by the day and insurance won't charge you by the day of it on/off.
You can always flip the logics around to your benefit and enable "on demand" like Amazon Prime subscription I guess?
That's why leasing is so common with Teslas. Sure, other EVs are often leased as well, but for most EVs, as long as you're happy with the range, charging speed, and features when you take delivery, it don't matter that much. But as autonomous features evolve the hardware and software need to improve together.
It's like an iPhone, where you still get software updates for many years and it continues to work and improve, but over time the latest features are eventually not available for older units anymore.
Even my 2025 Model 3 with AI4 (aka HW4) is technically already outdated... Yes, I'm using outdated in the same sentence as a vehicle that's less than one year old... but that's just how it is. Not only is it missing the front bumper camera that is now shipping, but some people are reporting that some newer production 2026 model year Teslas are quietly coming with AI4.5 which supposedly has three SoCs (processors) not the two that AI4 has. Theories so far is that the third chip will either be used to test new FSD versions in the background, to improve stability in edge cases where the two existing stacks disagree and a third stack will be useful to make a decision, or that they just need more compute. Who really knows...
AI5 is supposed to go into mass production and put in vehicles by the end of 2027. I REALLY hope that it's out by May 2027 when my lease is up so I can get a car with that, but I am not sure yet. But a lot changes in a year and we will see what is going on when we get there.