RJ: Rivian ADAS will expand to Tesla style FSD - full self driving on highway and roads (from driveway to destination)

[pamalabama]

In a related development, Waymo has "recalled" 1200 vehicles due to safety failures. Waymo has 250,000 rides weekly with years of actual experience with driverless vehicles. Tesla has a long way to go to catch up. I'd love to see what happens with Tesla's latest FSD with a 1" chain spanning an entrance to a parking area.

Tesla does not see chains. In practice it has not been a problem yet because they are not unsupervised

Sponsored

 

[EVtowing]

If you were not ignorant about FSD you would know that FSD has exponentially improved with each version. The reason that the miles per disengagement are the same is because tesla does not fix easy issues.

Sure that makes sense; don’t fix things that are easy to fix.

https://electrek.co/2025/05/14/tesl...i-service-without-driver-weeks-before-launch/

…Tesla’s ‘Supervised Full Self-Driving’ program, which is currently achieving about 500 miles between critical disengagements fleet-wide, according to the latest crowdsourced data.

Elon has said FSD need to be many times safer than the average human driver, and currently FSD is orders of magnitude worse.

Tesla is actually behind in autonomous driving and that Musk’s latest claims that Tesla will have “millions of robotaxi on the road” in 2026 are just the same as when he claimed it would happen in 2025, 2024, 2023, 2022, 2021, 2020, and 2019: corporate puffery.

 

[pamalabama]

Sure that makes sense; don’t fix things that are easy to fix.

https://electrek.co/2025/05/14/tesl...i-service-without-driver-weeks-before-launch/

Easy to fix is relative. Mobileye owns a patent on reading road signs. Easy could involve integrating all offending road signs into maps.

You don't need to test robotaxi without a driver. Having a driver in the seat is no different from driverless if the driver does not intervene. Tesla could be using remote operators already and the safety driver does nothing.

Unlike waymo, tesla is not hiding anything and is willing to let people ride driverless from day one. Waymo tests and you are required to sign an NDA

 

[Donald Stanfield]

I'm reading a one-sided argument and trying to keep up here because I must have the other side on ignore. It's pretty interesting trying to guess what the person I have ignored is saying by the context clues of all the people replying.

 

[Zoidz]

Tesla does not see chains. In practice it has not been a problem yet because they are not unsupervised

Then it's likely going to be a problem later this year when they go live with the Robotaxi.

 

[pricedm]

FSD is tech that should be made available for people with special needs...

Tell that to the 40,000+ "safe drivers" who die each year on USA roads, and the many more with life-altering injuries.

Any type of driver aids, starting with daytime running lights, are a boon to safety and lower injuries. "Tesla-style" FSD is long overdue, from Rivian and others (Lucid) offering advanced vehicles. Same with GM and Ford's offerings...

 

[SANZC02]

This is an interesting read….

 

[pamalabama]

This is an interesting read….

it's worth mentioning that tesla is one of the few brands reporting crashes using autopilot because they have the technology to do it

In other vehicles, if the person dies there is no record of proof of an ADAS system being used

https://www.nhtsa.gov/laws-regulations/standing-general-order-crash-reporting#level-2-adas


If you look at the crash data for level 2 systems, the Telematics data pretty much only comes from tesla. No one else

If you look at rivian crashes they look user reported, meaning that rivian is not reporting their crash data. At least for gen 1

 

[pamalabama]

Then it's likely going to be a problem later this year when they go live with the Robotaxi.

if you say so. I have used FSD for a long time and I'm not even 100% sure it won't hit a chain because I've never encountered one.

I've only seen anecdotes of autopark ignoring parking spots that have a chain in front

FSD definitely sees a chain link fence

 

[SANZC02]

it's worth mentioning that tesla is one of the few brands reporting crashes using autopilot because they have the technology to do it

In other vehicles, if the person dies there is no record of proof of an ADAS system being used

https://www.nhtsa.gov/laws-regulations/standing-general-order-crash-reporting#level-2-adas


If you look at the crash data for level 2 systems, the Telematics data pretty much only comes from tesla. No one else

If you look at rivian crashes they look user reported, meaning that rivian is not reporting their crash data. At least for gen 1

As I stated it was an interesting read, the part that should concern most people is the below quote. When the reaction time to these systems is measured in milliseconds, 5-10 seconds is a literal eternity.

From the article:

”In 59 crashes examined by NHTSA, the agency found that Tesla drivers had enough time, “five or more seconds,” prior to crashing into another object in which to react. In 19 of those crashes, the hazard was visible for 10 or more seconds before the collision.”

 

[pamalabama]

As I stated it was an interesting read, the part that should concern most people is the below quote. When the reaction time to these systems is measured in milliseconds, 5-10 seconds is a literal eternity.

From the article:

”In 59 crashes examined by NHTSA, the agency found that Tesla drivers had enough time, “five or more seconds,” prior to crashing into another object in which to react. In 19 of those crashes, the hazard was visible for 10 or more seconds before the collision.”

That's because people are abusing the system.

If you use that metric then 99% of ADAS systems do not brake for stationary objects

Tesla is actually in the minority here that it can see a stationary object and slow down at a natural pace

Outside of tesla, BMW seems like the only brand that can do it without lidar

Which means the amount of crashes tesla would have prevented despite the system being abused has to be very high

 

[SANZC02]

That's because people are abusing the system.

If you use that metric then 99% of ADAS systems do not brake for stationary objects

Tesla is actually in the minority here that it can see a stationary object and slow down at a natural pace


Outside of tesla, BMW seems like the only brand that can do it without lidar

More of a concern to me when they start allowing eyes off on systems. Especially since most of the regulatory government agency controlling this have been decimated. You know companies will push these systems out before they are ready for prime time trying to stay ahead of the curve.

Just my opinion but when you have a 5-8k pound missile going down the road even 99% reliability can be catastrophic in that remaining 1% case.

 

[pamalabama]

More of a concern to me when they start allowing eyes off on systems. Especially since most of the regulatory government agency controlling this have been decimated. You know companies will push these systems out before they are ready for prime time trying to stay ahead of the curve.

Just my opinion but when you have a 5-8k pound missile going down the road even 99% reliability can be catastrophic in that remaining 1% case.

Those crashes are from mostly autopilot and not FSD. 99% reliability is absolutely nothing. FSD easily has 99% reliability now if you look at the miles per disengagement

 

[EVtowing]

FSD easily has 99% reliability now if you look at the miles per disengagement

Please provide a source for that assertion, since Tesla does not release data on FSD miles driven per disengagement.

 

[ohseedee]

What I find intriguing is how people form opinions about the effectiveness of FSD based on outdated experiences. FSD software and hardware are not static, and unless you’ve recently used the latest version of both, your assessment is likely limited. I received a free trial of FSD on my MY with hardware 4 about 1.5 years ago. FSD was flawed, with frequent disengagements and I found it awkward. Consequently, I decided it wasn’t worth paying for it when my trial expired. However, they released a significant update earlier this year, accompanied by another free trial. I found it remarkably close to perfect, with zero required disengagements. I could simply enter my destination, hold the button on the screen, and just sit there until it was time to park 100 miles away. Ultimately, I decided to pay for FSD because I genuinely believe that supervised driving with FSD is currently safer and more relaxed than without. this is my wife’s primary car, and I feel it significantly reduces the risk of accidents. Initially, I had planned to replace my MY with my R2 reservation, but that decision has been reversed due to my desire to have FSD. My perspective is based on the hardware version 4 and the latest software, and I find it incredibly impressive and worth $99 per month. If it were based on the version released earlier in 2024, my opinion would be that it’s somewhat cool but not worth paying for.

Sponsored