Out of Spec roadtrips Gravity and DDP2 has some issues

Yes they show happy people dancing in the car, right before they get T boned because the car didnt see the truck failing to stop. Or the idiot in rush hour cut in front of you while being tailgated.

Tesla accident avoidance with FSD is superhuman already. The car will easily see a truck failing to stop.



Look at these accident avoidance scenarios and with FSD 14 it's even better.

If you compare this to chinese or nvidia or whatever, the accident avoidance of those systems will be nonexistent by comparison. There's a reason tesla operates 200+ cars in texas unsupervised using the same technology that is on consumer cars. Because they have accident avoidance at levels they can do this.

Other robotaxis have sensors mounted really high that can see over cars. Tesla does not have that and overcame this limitation with fast reaction time and good accident avoidance. Just having lidar/radar is not enough.
 
Tesla accident avoidance with FSD is superhuman already. The car will easily see a truck failing to stop.



Look at these accident avoidance scenarios and with FSD 14 it's even better.

If you compare this to chinese or nvidia or whatever, the accident avoidance of those systems will be nonexistent by comparison. There's a reason tesla operates 200+ cars in texas unsupervised using the same technology that is on consumer cars. Because they have accident avoidance at levels they can do this.

Other robotaxis have sensors mounted really high that can see over cars. Tesla does not have that and overcame this limitation with fast reaction time and good accident avoidance. Just having lidar/radar is not enough.
And yet Teslas on the latest FSD continue to make dangerous decisions every day, as reported by the many people that use it on Reddit. That also explains why their autonomous operations in California are non-existent. California has far more stringent testing and reporting requirements than Texas, where AV regulations are kind of a joke. Let's also not forget that every one of their Cybercabs isn't actually road legal, as they did not get an exception from the NHTSA like Zoox did. They instead self-certified, which is why there is an ongoing probe.

We are no longer compute-limited with Nvidia's latest chips (Thor). We USED to be, which probably explains why Tesla decided to proceed with a camera-only configuration. Unfortunately for Tesla, reverting back to a more sensor-heavy hardware stack would set them back a few years, essentially eliminating them from the driverless vehicle "race". IMO, they're doing the broader public a disservice by risking lives with a camera-only system when it's clear they CAN increase safety with the inclusion of radar and lidar at the cost of their ADAS "leadership".
 
Right now, Tesla FSD is orders of magnitude better than any other “Currently Available” consumer ADAS. Hopefully they will get some significant competition from Lucid, Mercedes, Rivian (and possibly some others) in the next year. But until those systems go live, it is just speculation. My guess is that these new FSD Like systems will be more comparable to FSD 13 and take a couple years to catch up to Tesla (remember that as these new competition systems are refined, so will FSD). Use of new NVIDIA Thor processors will help accelerate the competition. The good news is that within 5 years, there should be multiple manufacturers with cars that can safely do point to point self driving. AI will accelerate this process.
 
And yet Teslas on the latest FSD continue to make dangerous decisions every day, as reported by the many people that use it on Reddit. That also explains why their autonomous operations in California are non-existent. California has far more stringent testing and reporting requirements than Texas, where AV regulations are kind of a joke.
You do realize FSD is being used for 15 billlion miles with about 1 billion miles every month? 100% of these miles are recorded with video. Of course you will see errors. Secondly, not everyone is on HW4. A lot of the bad FSD stuff is HW3, and not every bad thing you see HW4 do is going to lead to an accident.

Then you have the poster above who claims FSD made many safety critical errors across 600 miles yet he is too ignorant to see that all of his suupposed "critical errors" were just mapping issues. You can literally report these mapping errors and then your car will not make this mistake anymore. Tesla drives with really poor maps but they just drive so well that they hide a lot of the mapping issues. When you get your hands on nvidia alpanmayo you will see how bad the generalized driving is.
We are no longer compute-limited with Nvidia's latest chips (Thor). We USED to be, which probably explains why Tesla decided to proceed with a camera-only configurationIMO, they're doing the broader public a disservice by risking lives with a camera-only system when it's clear they CAN increase safety with the inclusion of radar and lidar at the cost of their ADAS "leadership".
The nvidia thor chips are weak where it counts. Memory bandwidth. Tesla actually has very high memory bandwidth compared to nvidia, and the more sensors you have, the more memory bandwidth you need. Not to mention AI5 has over 2000 tops for just one chip, and in 2 of them t hat absolutely smokes nvidia thor.

Anyone that is not tesla running a robotaxi, is using effectively server grade chips that consume 1000s of watts. They have terabytes of memory bandwidth and they have ridiculous amounts of memory as well. As you can imagine it's not possible to put these in cars without reducing your range by 50% or more

Tesla is the only company that has demonstrated that it is possible to operate without HD maps and with consumer level compute.

Unfortunately for Tesla, reverting back to a more sensor-heavy hardware stack would set them back a few years, essentially eliminating them from the driverless vehicle "race".
This is false. Tesla safety with robotaxi is very good. The only issue is whether not having lidar or the placement of their cameras is an issue scaling to 100-200K vehicles or a larger ODD. At that point, yes they could add lidar, different camera positions, etc. but that would not set them back as nobody else has a technology that scales to that many vehicles.

Tesla would never end up behind, because that would imply somebody passes them, which is clearly not the case, or that anyone is close to doing it

Even if they added more sensors, they would drive 99.99% like they already do and use the sensors to fill in missing areas, which would not take years and/or set them back at all.

Let's also not forget that every one of their Cybercabs isn't actually road legal, as they did not get an exception from the NHTSA like Zoox did. They instead self-certified, which is why there is an ongoing probe.
This is 100% false and you need to stop believing the nonsense you read on the internet. Cybercab was literally 99% FMVSS compliant. The only "missing" feature was is the foot activated brake control. Turns out the internet was wrong and cybercab does have a foot activated brake control.

There's a button underneath the carpet that you can very clearly hear in one spot of the car if you press on it

The zoox story is different. Zoox self-certified their car as compliant like tesla did. They used a BS excuse that they could design alternative test procedures for missing components of the car (because they did not apply). Therefore zoox self-certified a non compliant car. Zoox was penalized because they were told that they could not use alternative test procedures.

Tesla did not do this. they built a 100% FMVSS compliant car by including every function required under the FMVSS. Turns out FMVSS does not require a steering wheel or accelerator pedla, and only a foot activated brake control is required. They complied with this by putting a hidden button underneath the carpet.

There are other required functions like wiper control, turn signals, lights, shift indicator, speedometer etc. but that stuff is all right there on the center screen
 
Last edited:
Tesla accident avoidance with FSD is superhuman already. The car will easily see a truck failing to stop.



Look at these accident avoidance scenarios and with FSD 14 it's even better.

If you compare this to chinese or nvidia or whatever, the accident avoidance of those systems will be nonexistent by comparison. There's a reason tesla operates 200+ cars in texas unsupervised using the same technology that is on consumer cars. Because they have accident avoidance at levels they can do this.

Other robotaxis have sensors mounted really high that can see over cars. Tesla does not have that and overcame this limitation with fast reaction time and good accident avoidance. Just having lidar/radar is not enough.
Yes I did not stop where there no stop sign for me. And the truck blew through the intersection

And yes I have had drivers do that. And Tesla is not autonomous like everyone keeps saying. Just sit in the back and have it safely drive you through traffic and then we’ll talk. But until then it remains a supervised driving mode. Because you know eleon (yes I know the spelling) really cares about you, and your health, specifically.

Is Tesla supervised driving better than Lucids today, yes. Will it always be, I think not.
 
Tesla accident avoidance with FSD is superhuman already. The car will easily see a truck failing to stop.



Look at these accident avoidance scenarios and with FSD 14 it's even better.

If you compare this to chinese or nvidia or whatever, the accident avoidance of those systems will be nonexistent by comparison. There's a reason tesla operates 200+ cars in texas unsupervised using the same technology that is on consumer cars. Because they have accident avoidance at levels they can do this.

Other robotaxis have sensors mounted really high that can see over cars. Tesla does not have that and overcame this limitation with fast reaction time and good accident avoidance. Just having lidar/radar is not enough.
This is a joke right?

The car passed by at like 10-15 MPH after a multi second delay to finally pass in front of the stopped tesla. I can't be bothered to read the rest of the context but this was not, at all, an impressive save to me. This is bare minimum to be considered FSD. Great that it worked but let's not pretend this is 'super human'. Had the Tesla not stopped at the sign like a kid on his drivers license test and then accelerated like Grandma it would have been through the intersection before the other car even got there. The other car approached fast, saw the intersection was taken, slowed down, and then chose to continue because the tesla had come to a complete stop. If the driver feels like this would have resulted in a life ending crash had he/she been driving then they have no business behind the wheel. Perhaps the exact person who needs FSD in the first place....

To be clear, I agree FSD is light years ahead of any other right now and that is to be applauded.
 
Last edited:
Wow I agree if fsd saved your life in that situation then get off the road as you have no capacity for driving.
 
This is a joke right?

The car passed by at like 10-15 MPH after a multi second delay to finally pass in front of the stopped tesla. I can't be bothered to read the rest of the context but this was not, at all, an impressive save to me. This is bare minimum to be considered FSD. Great that it worked but let's not pretend this is 'super human'. Had the Tesla not stopped at the sign like a kid on his drivers license test and then accelerated like Grandma it would have been through the intersection before the other car even got there. The other car approached fast, saw the intersection was taken, slowed down, and then chose to continue because the tesla had come to a complete stop. If the driver feels like this would have resulted in a life ending crash had he/she been driving then they have no business behind the wheel. Perhaps the exact person who needs FSD in the first place....

To be clear, I agree FSD is light years ahead of any other right now and that is to be applauded.
I'm not talking about the first post. That post is not even me. I'm talking about the accident avoidance in the second video. And those are from fsd v13

My point is FSD has real accident avoidance. If someone cuts you off or runs a red light/stop sign it will see it and brake
 
S
I'm not talking about the first post. That post is not even me. I'm talking about the accident avoidance in the second video. And those are from fsd v13

My point is FSD has real accident avoidance. If someone cuts you off or runs a red light/stop sign it will see it and brake
ti not seeing anything near as good as a person should be doing!
I don’t care what anyone thinks the law says there’s no reason to pull into the middle of the highway and stop for a car you never should have been near.
The car clearly created the situation in the first place, so it saved you from its own FUBAR

Now that’s the funniest twist I’ve seen yet. Sorry bu no
 
Is Tesla supervised driving better than Lucids today, yes. Will it always be, I think not.
People get this wrong but it 100% always will be.

What people don't understand is self driving is not just a software/hardware issue. It is a complete package.

What stuff is actually needed for a car to become unsupervised?

- You need camera cleaning for every camera. Looks like it is trivial for tesla to add this when they have a tech that can scale to consumer cars.
- you need as close to perfect data connection as possible in every region of the U.S. and of course cellular is not enough.
- you need 24/7 remote support capable of handling millions of vehicles
-you need real time accurate and updated maps of the U.S. to provide the safest possible prerequisite knowledge
-you need vehicle to vehicle communication to enhance safety
-software will need to be updated to improve driving, enhance ODD, et c.


All of this costs billions of dollars a year, which means that this only works if you have enough vehicles to offset the cost. This means that it's likely not a licensed system that ever becomes #1 and there will be a monopoly where one system is 10x or more cheaper than the nearest competitor
 
Have you seen Tesla's sales or their order backlog? Both are strong and growing despite, primarily, the widespread hatred for Elon Musk, increasing choice in the marketplace and Tesla not really offering anything new in 5 years. Their HW4 equipped used cars have nearly no depreciation. Why is this? It's because FSD is now so very effective.
 
People get this wrong but it 100% always will be.

What people don't understand is self driving is not just a software/hardware issue. It is a complete package.

What stuff is actually needed for a car to become unsupervised?

- You need camera cleaning for every camera. Looks like it is trivial for tesla to add this when they have a tech that can scale to consumer cars.
- you need as close to perfect data connection as possible in every region of the U.S. and of course cellular is not enough.
- you need 24/7 remote support capable of handling millions of vehicles
-you need real time accurate and updated maps of the U.S. to provide the safest possible prerequisite knowledge
-you need vehicle to vehicle communication to enhance safety
-software will need to be updated to improve driving, enhance ODD, et c.


All of this costs billions of dollars a year, which means that this only works if you have enough vehicles to offset the cost. This means that it's likely not a licensed system that ever becomes #1 and there will be a monopoly where one system is 10x or more cheaper than the nearest competitor
Welcome Tesla fanboi to the Lucid channel.
 
Welcome Tesla fanboi to the Lucid channel.
I'm just not ignorant about other's technology.

There's a reason tesla is operating ~400 unsupervised vehicles in austin today. That's more than waymo. The same technology in those cars is the same technology in a tesla.

There are so many self driving companies using HD maps, geofenced driving, 360 lidar, sensors, server-grade compute, etc. and they are nowhere close to ever doing something like this. Most of them are nowhere being able to do any level of driverless at all.

There are self driving consumer cars in china and the safety of the best systems is worse than FSD 12. There was a comparison of hauwei ADS 5 (with lidar) which performed worse than FSD 14 and the livestream was cut halfway through.

If self driving is a sensors issue, it clearly doesn't look that way.

There are a suprising number of people all over the internet that literally believe tesla is remote driving these vehicles. Like 100% of the time remote driving them.
 
I bought my 2017 new (old HW3 hardware, v12 FSD) and have used their FSD since its inception. A few thoughts:

1. FSD with an alert driver is safer than alert driver alone. Same with DreamDrive, I'm sure.

2. FSD (non RT version) is nowhere near autonomous, even on newer models and the latest hardware.

3. The goals of the Tesla RT business has overlaps with privately owned L2++ system, but there are significant differences too. RT's rarely park, for example while L2++ needs to do that well these days to be competitive. There are other examples. Feels to me like Tesla is not much interested in L2++ systems and would rather put those resources to RT.

4. Even the Tesla RT is seriously crippled. Selective easy routes, no highways, limited scope. Perhaps with those serious restrictions, it could be considered safe. I am not an RT fan or user.

5. Even with my old hardware and no more Tesla support, FSD does the vast majority of my driving. It's maybe 98% accurate and its relatively easy to manually intervene. Its also predictable and I feel very safe. I have no trips that are intervention free.

6. Best I can tell, the L2++ systems being delivered today by almost all is what I had in 2017 - freeway lane keep and lane change on demand.

7. Not sure I care what kind of ADAS sensors a car has. I'm more interested in what it can do.

8. Lucid obviously makes an incredible vehicle. But I'm not buying another vehicle based on "this year for sure". Been there, done that, never again.

9. What I'm waiting for is a premium vehicle that can do L2++ supervised circa Tesla 2020 or better - supervised, city streets, follow nav, recognize lights/signs. Lots of talk from lots of manufacturers in that regard but nothing but vaporware to date.

I'm still waiting. My wish is that Lucid shows some real progress in this area soon.
 
There's a reason tesla is operating ~400 unsupervised vehicles in austin today. That's more than waymo
What are you talking about? Waymo has almost 4000 cars on the road with over 500 in Austin alone, with about 3000 of those unsupervised. This is the problem with die hard FSD fanatics, they skew the truth constantly.
 
What are you talking about? Waymo has almost 4000 cars on the road with over 500 in Austin alone, with about 3000 of those unsupervised. This is the problem with die hard FSD fanatics, they skew the truth constantly.

Robotaxi tracker creator (ethan mckenna) tracked public cameras in Austin. Visible cybercabs on the road relative to waymo vehicles, and per 1000 vehicles, cybercabs are almost 1:1 waymo vehicles.

Tesla was typically running about 50% cybercab, 50% model Y which means either waymo is not running 500 vehicles, or tesla is running over 600. Tesla in a public X post said they were running 300+ cybercabs today

Also keep in mind waymo does not run nearly as much as they used to. They pretty much stop in all levels of rain. The issue is that they drive into flooded waters, stop moving, and their remote software does not let them move the car. They have a remote driving feature that uses keyboard and mouse but they claim they have never used it and are too proud to ever use it. After one washed away for 4 days in san antonio they seriously stopped running in rain.

What's funny is waymo hid this deficiency for years, I'm guessing by geofencing the vehicles around flooded areas. For whatever reason they don't do that anymore.
 
Have you seen Tesla's sales or their order backlog? Both are strong and growing despite, primarily, the widespread hatred for Elon Musk, increasing choice in the marketplace and Tesla not really offering anything new in 5 years. Their HW4 equipped used cars have nearly no depreciation. Why is this? It's because FSD is now so very effective.
That being said I'm looking for a replacement for my Tesla since my wife will take it away from me. I ordered a Cayenne S that got delayed so many times I canceled my reservation. The Cayenne came in last week and when I drove it I was a bit underwhelmed for a car costing $144k. It has great steering and the absolute worst console ever conceived. The '27 Air reveal disappointed me because of the insistence on black front seats and the 50 kW Tesla charging. The search continues.
 
Robotaxi tracker creator (ethan mckenna) tracked public cameras in Austin. Visible cybercabs on the road relative to waymo vehicles, and per 1000 vehicles, cybercabs are almost 1:1 waymo vehicles.
Brb…. Need to stand on the side of the road for an hour and count how many Lucid’s drive past to determine the total number of Lucid’s in town 😂
 
Brb…. Need to stand on the side of the road for an hour and count how many Lucid’s drive past to determine the total number of Lucid’s in town 😂
He's not tracking absolute number of cybercabs. He is tracking visibility of each vehicle relative to every 1000 vehicles he captures. There are public cameras all over austin so this is very useful.


That means the visibility of cybercab and waymo is the same. tesla officially gives a number of 300+ cybercabs running today

if we assume they're not all out at once, waymo is not even running 300

And in tesla's case they run an entirely seperate model (model Y) which they run quite a lot

And we are only discussing driverless.

If you include test vehicles, tesla has manufactured well over 3000 cybercabs based on vin numbers
 
Last edited:
That being said I'm looking for a replacement for my Tesla since my wife will take it away from me. I ordered a Cayenne S that got delayed so many times I canceled my reservation. The Cayenne came in last week and when I drove it I was a bit underwhelmed for a car costing $144k. It has great steering and the absolute worst console ever conceived. The '27 Air reveal disappointed me because of the insistence on black front seats and the 50 kW Tesla charging. The search continues.
iX5?
 
Back
Top