Tesla FSD Full Self Driving & Rivian ?

[Rivian Head]

Thoughts?

You posted something about Tesla in the most anti-Tesla forum on the internet. There's not going to be any actual discussion on the subject. It's just "Elon Bad".

So true, blinded by hate!

Sponsored

 

[Rivian Head]

Reading lots of comments about how terrible FSD is. I’d remind folks that the FSD in cars today is very very different than the FSD in cars at the start of this year. I’ve got over 1K miles on the latest FSD during the April trial period and I’d say it’s pretty freaking great. It is pretty clear after this past few weeks that Tesla’s vision of autonomy using cameras only was the right one. While I’ve never been in a waymo or other autonomous car, there has never been a situation in the past month that my Model Y didnt see something that lidar could have. It never disengages and steering is pretty much perfect. My only gripe is how long it stops at stop signs. Also I guess the auto park (while really really good) is too slow. If they started a robo Taxi service tomorrow I’d have no issues using it over an Uber. It drives like a regular (maybe overly cautious) human.

That said, I cant see myself paying $99/month for supervised/hands on FSD on my Model Y. I care more about drivers assistance on long highway stretches vs city streets and that’s free with standard auto pilot. While I probably also wouldn’t pay for FSD on my R1T, I certainly would rather have Tesla’s autopilot assisting me on highways over Drivers+. However, once you can go hands free / eyes off I’d absolutely pay $99 month.

Ultimately I’m not sure how interested Rivian is in autonomy. I cant see them ever invest the time and resources that Tesla has. It might make sense to leverage Tesla FSD as a way to get those features early and invest in their own solution later if they choose. Similar to how they used the Bosch motors to get started but ultimately planned on phasing those out with their own.

After the trial, we will pay for monthly subscription. It’s an amazing system that keeps on improving. It saw a man behind a big van today which I didn’t see at an intersection. It will undoubtedly save many lives. Will see what the data reveal after another billion miles of FSD V 12. Humans do not use Ultrasound like bats do. Not sure if redundancy would add much to HD camera based system. We can only speculate, but will have to wait for the data after another billion miles. But I am convinced autonomous driving is around the corner, and will save lives; 37, 000 death per year in the US, and mostly from human errors .

 

[f1racer328]

Half of the battle of more automated cars is going to be training people on how to properly manage the automation.

Automation is a huge part of my job, and half of the battle is knowing its limitations and when to disconnect it. Tesla and other auto makers do not do a good job at educating this.

My friend has FSD on a Model X and we've been using it the past few days. It has vastly improved from earlier this year, but still has a lot of progress to be made.

We watched FSD Just sit right next to another car as 2 lanes were merging. Didn't attempt to speed up or slow down, manual intervention was required. I wasn't there for this one, but it also didn't want to stop for a school bus this morning (multiple lane road)

 

[Dark-Fx]

Reading lots of comments about how terrible FSD is. I’d remind folks that the FSD in cars today is very very different than the FSD in cars at the start of this year. I’ve got over 1K miles on the latest FSD during the April trial period and I’d say it’s pretty freaking great. It is pretty clear after this past few weeks that Tesla’s vision of autonomy using cameras only was the right one.

How often are you driving in substandard conditions? At a minimum I'm curious about daytime when the sun is glaring directly into your windshield. But I'm also curious about fog, heavy rain, snow, etc. Those are the situations where I'd want the help if the computer is more capable than I am, and situations where Tesla's "vision only" doesn't receive more data about the world than I do.

IMO it's also one of the reasons I don't ever see the robotaxi system working anytime soon, especially if it's vision only. "Sorry, you can't request a ride right now because of inclement weather"

 

[CrazyOne]

No thanks. What I want is dumb cruise control with warnings for use at non highway speeds. I use it extensively to help avoid tickets. Rivian was pretty good on highway and I used it almost all the way from Vancouver to Seattle.

An imperfect system, as they all are today will have false positives or false negatives. False positives give you phantom emergency braking and false negatives giv e you dead motorcycle riders or crashes into emergency vehicles.

 

[the_mace]

Then was mention in Tesla's earnings call yesterday that they are in talks with another car manufacture about licensing their full self driving software. I wonder if Rivian may be that manufacture. Driver+ seems to be way behind others, and I would expect they could more easily make the necessary hardware changes.

Faster Path to Market:

• Reduced Development Time: Leveraging Tesla's existing FSD system would save Rivian years of development on their Driver+ assistant. This allows Rivian to offer a more competitive product sooner.
• Improved Feature Set: Tesla's FSD is constantly evolving, offering features like lane-changing and traffic light recognition. Licensing FSD would instantly upgrade Rivian's Driver+ capabilities.

Addressing a Potential Customer Need:

• Market Demand: While Rivian hasn't prioritize self-driving, some customers might desire it, especially when Tesla officially rolls it out. FSD could attract a wider audience seeking advanced driver-assistance features.
• Competitive Landscape: Tesla is a leader in self-driving technology. Offering a comparable system would help Rivian compete more effectively.

Focus on Core Strengths:

• Frees Up Resources: By using FSD, Rivian's engineering team could focus on core vehicle development and differentiation, like Rivian's off-road focus.
• Faster Iteration: Licensing avoids the need for a complete in-house solution, allowing Rivian to adapt and improve Driver+ features more quickly based on Tesla's updates.

However, there are also potential drawbacks to consider:

• Cost: Licensing FSD might be expensive, impacting Rivian's profit margins.
• Integration Challenges: Integrating Tesla's software with Rivian's hardware and systems could be complex and time-consuming.
• Strategic Dependence: Rivian would become reliant on Tesla's roadmap for future advancements in self-driving technology.

Ultimately, the decision depends on Rivian's long-term vision. If self-driving becomes a priority (or a competitive requirement), licensing FSD could be a strategic move. But if Rivian prioritizes differentiation and driver engagement, focusing on in-house development of Driver+ might be their choice (... and we are in for a long wait).

Thoughts?

So I have it in my Tesla Model S, got in early on the FSD Beta and have >30K miles driving with it through all the various versions. When I bought my Tesla (2018) it was a $3K add on. I wouldnt pay for it on the Rivian with its current capabilities. Honestly they just need to nail the divided road behavior and have it work on ALL divided roads and do lane changes/follow route. People other than Tesla have this working pretty well, although they're all super conservative with allowing you to enable it. The backroads etc are not close on Tesla and its going to be a long haul.

Overall i'm not a fan of Rivian connecting with Tesla other than the charging side. They dont need the rest.

 

[R1Thor]

I've seen how a lot of you drive.

Full Self Driving ANYTHING can't come soon enough!!

Objectively, computers can make decisions faster, calculate stopping/swerving/maneuvering in fractions of a second, and will understand the telemetry of the vehicle MUCH better than anyone ever will. And as those algorithms are honed to continue to evaluate road circumstances, traffic circumstances, environmental situations. It's only going to get better.

We will absolutely--and should embrace-- see a day where cars are doing most of the driving for us. It'll be objectively safer, AND (what a lot of people are glossing over) is it'll get us places more efficiently (if cars are talking to cars are talking to roadway infrastructure, say goodbye to traffic jams).

AS a perfect example, turn on your adaptive cruise control in any vehicle and see that, even when adjusted to the minimum following distance, it's practically 2x farther away than the typical person drives at highway speeds from the bumper of the person in front of them. This is a system that is doing linear algebra on a second by second basis to determine "if I had to stop in an emergency, how far behind would I have to be." Humans are completely incapable of fathoming the reality of this.

Now, the (current) implementation clearly has a lot of maturing to do. But it gets better with more data over time.

There will come a day that someone is cursing grandma and grampa for being luddites and refusing to turn on their FSD because THEY are the ones causing traffic to backup because of it.

Embrace change, friends. It's inevitable! (You're driving electric vehicles, for crying out loud!!!)

 

[BCondrey]

I wouldn't use it if it were free. My wife has 60 days of FSD free with her new Model 3, and we aren't even going to try it.

I like driving my truck, thats why I bought it. I can see applicability for age / mobility / handicapped folks. I hope Rivian stays well away from it.

 

[bdwalters]

Rivian is not fine with it and that is why they implemented plug and charge at Tesla super chargers. Customer info stays with Rivian. Detailed data released on this.

User data is the most valuable thing any of these new car companies have. Your driving data is more valuable than the $$$ you spent to buy the car. Every time you drive the data is being recorded and mapped. you really think anyone hands that over for free... this is why Rivian and Tesla have not implemented Android auto or CarPlay...

Rivian wouldn’t be handing the data over for free. They are getting in return a system that they can’t afford to develop on their own.

 

[iBang]

I would love to have Tesla FSD on Rivian. I currently drive Model S with FSD and their latest update nailed it. I now use FSD on 90% time and never disappoints. It drives seamlessly in busy LA downtown streets. However I don't think Rivian will go for it. I mean they don't even want to allow Apple Car Play to be integrated into their system.

 

[voxel]

Now, the (current) implementation clearly has a lot of maturing to do. But it gets better with more data over time.

Yeah not really. The current ML tech is all just a LLM. Call me unimpressed for using that for FSD where the car ends up reacting to situations rather than deep planning.

I want AI to self drive our cars but the two schools of AI - neural nets (deep learning) and knowledge rep (building models/rules of world) - have not merged. All we have are poorly reactive cars that can't handle exceptions or have to be trained on every single exception in the world... like idiots u-turning in a traffic circle.

 

[Eric9610]

Rivian wouldn’t be handing the data over for free. They are getting in return a system that they can’t afford to develop on their own.

You proved my point. Tesla is willing to give tech up for access to user data. That tells you how valuable it is.

also we don’t know how far along Rivian is and you are guessing. I think Driver + is better than autopilot where available. We just need Rivian to map more roads.

another thing to note only 1 level 3 system is approved for used in NA approved by NTHSB and it’s not teslas. MB is the first and only level 3 approved system.

I don’t see Tesla getting NTHSB approval for level 3 in current form without including more hardware. Tesla went cheap and uses mass marketing to blind people to the truth. It’s not FSD. It’s assisted Driving.

 

[Tahoe Man]

These premium features will become common everyday feature systems, like airbags. And the wow factor wears away and premium pricing declines. Licensing is good for Tesla, why not anyway. But at the end of the day, it takes a lot of capital cost to do these things and then there is the regulator hurdles. Rivian should have their system work good enough but not anymore do to cost.

 

[SwampNut]

I miss my Autopilot so very much. I let it drive 98% of the time. I knew I'd lose this with the Rivian, and only did it because I really do have use for a truck. AP was near perfect in most places, and I almost never had to drive. Now I have to drive until I get to a highway, and most of my days are on city streets where AP excelled.

 

[Rivian Head]

Unless Rivian stands out with both hardware and software, they will not make it in crowded EV market. As of 2024, Rivian doesn’t stand out like it did 3 years ago. Clearly Tesla stands out as “the technology “ company, and the Wall Street seems to buying its hype as of today. It’s a no brainer for Rivian to adopt FSD just like they did with charging network. Elon’s reference to Nokia phone is spot on. Automated driving is the next logical application of AI.

Sponsored