-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Autonomous Car Progress
- Thread starter stopcrazypp
- Start date
-
- Tags
- Autonomous Vehicles
oh come now, while I'm not worried about the Tesla vs Waymo bickering (I'm happy to sit back and watch them BOTH improve), I think this is a bit of a circular argument. Waymo has autonomous driving (an assertion), so editing a video is harmless. But on what can your assertion be based? That they have videos showing the car driving autonomously! Hmmmmm.
It all comes down to cherry-picking and/or confirmation bias. Those who hate Tesla are happy to point to some of the (also heavily edited) videos showing FSD messing up, those who are fans of Tesla point to counter examples. I've no doubt Waymo prefer showing success video of their cars and not those where the car does less well. That's hardly shocking.
Tesla are attempting a path to make a car drive reasonably well everywhere, Waymo a path to make a car driver very well in limited areas. Both have their merits and demerits, and both, ultimately, are going to end up in the same place but via different routes.
Doggydogworld
Active Member
The assertion is based on the fact that any of us can go to the east valley area of Phoenix and hail an autonomous Waymo robotaxi.
Sometime in the next year (or two, ha) you'll be able to hail one in downtown Phoenix and San Francisco as well. Despite Musk's pie-in-the-sky BS, Tesla is far from having robotaxis anywhere. More to the point, they have no reason to even pursue robotaxis. Their current business model of selling FSD at 12k a pop is vastly superior.
diplomat33
Average guy who loves autonomous vehicles
Waymo has driverless robotaxis that the public can use. So I think it is a bit more than just an assertion. Again, I am not saying Waymo has solved autonomous driving or has autonomous driving everywhere, but they do have autonomous driving.
It probably depends on the nature of the editing. If you have a perfect driverless ride and you simply edit out the boring parts for time, that's fine. If you edit out all the driver interventions to make the autonomous drive look perfect, that's problematic. Of course, with an edited video, it is impossible to know if it is the former or the latter. My real point is that I felt that the other poster was trying to distract and attack by focusing on the edited nature of the video rather than the substance of the video.
I want both Tesla and Waymo to succeed. I am a Tesla owner and a Tesla fan. I am also a big fan of Waymo's tech. I think there is a place for both approaches, Tesla in the consumer car space and Waymo in the robotaxi space. I just wish more people would admit that Waymo has really good autonomous driving. And stop making it us versus them. You are not betraying Tesla for liking Waymo's tech. And you are not betraying Waymo if you like Teslas. I like both Tesla and Waymo for different reasons. I also really like Mobileye's FSD approach.
I agree. I've said before that Waymo is pursuing a "tall" approach while Tesla is pursuing a "wide" approach. Tesla is focused on getting autonomy to work everywhere first, and then making it great, while Waymo is focused on getting autonomy to work great in one area first and then expanding to everywhere. If you want to own a great EV that can drive itself anywhere but will likely need some supervision, Tesla is for you. If you live in a city with Waymos and you want to be able to hop in the backseat and have the car chauffeur you around like an expert driver, Waymo is for you.
I think we are getting way off topic. Debating who is a shill or brainwashed is not what this forum should be about.
Last edited:
EVNow
Well-Known Member
Absolutely right.
But we can’t discuss stuff if there are trolls who work for Tesla competitors running around here shilling for Waymo and abusing Tesla fans.
Doing so is absolutely unethical and supporting such shills is no better.
ps : If someone trolls like this in the investment thread they would be banned from TMC in 5 minutes.
I was very active in various forums when I worked at Microsoft for 15 years and I didn’t write one word about its products or competitors. If I went on PlayStation forums and repeatedly trashed PS while hyping XBox, I’d be called a shill for XBox and banned.
Last edited:
For good and for bad we give a longer leash. Still, the astute will notice that calling someone a troll or shill should assure your post will get snipped. Let’s be civil.
Statements on how awesome Tesla or Waymo will obviously be in a year I let stand. Even though I cringe when I read them. Stick to the facts please.
You think that it is a good thing? Because of that, the investor thread is very biased.
EVNow
Well-Known Member
You think TMC should give a platform for competitor shills & trolls ?
I think a happy medium can be achieved.
As I keep saying - if someone is working for a competitor of Tesla and only writes anti-Tesla posts here - without disclosing their conflict of interest - it is downright unethical and IMO should not be tolerated.
We have to be just truthful.
Last edited:
Data likely shows Teslas on Autopilot crash more than rivals
The government plans soon to release data on collisions involving vehicles with autonomous or partially automated driving system that will likely single out Teslas for a disproportionately high number of such crashes.
apnews.com
Daniel in SD
(supervised)
Not autonomous. This should be in one of the many "speculating about Autopilot safety" threads.Data likely shows Teslas on Autopilot crash more than rivals
The government plans soon to release data on collisions involving vehicles with autonomous or partially automated driving system that will likely single out Teslas for a disproportionately high number of such crashes.
diplomat33
Average guy who loves autonomous vehicles
diplomat33
Average guy who loves autonomous vehicles
New clip of Mobileye's autonomous driving in Paris.
It is great to see Mobileye testing in so many diverse cities around the world from Munich, Jerusalem, Paris to Orlando. And yes, I am fully aware that a 15 second clip does not prove anything. The purpose of the clip is not to prove that Mobileye has "solved FSD" or anything, but just to show that Mobileye is scaling their testing in lots of different places.
It is great to see Mobileye testing in so many diverse cities around the world from Munich, Jerusalem, Paris to Orlando. And yes, I am fully aware that a 15 second clip does not prove anything. The purpose of the clip is not to prove that Mobileye has "solved FSD" or anything, but just to show that Mobileye is scaling their testing in lots of different places.
Last edited:
It seems that the object being classified are far less stable than with FSD. As the car drives, pedestrians appears instantly in the middle of the road. A car changes into a semi, then into a SUV and motorbike – at much higher rate than with the latest FSD. The rest of the scene is table, but as far as I know, it's because of their HD map (see pedestrians crossings and lanes drawn long before the camera can see them). After watching many dozens of FSD videos on YT, I'm tempted to believe Tesla keep getting ahead.
Adding countries/cities doesn't seem such a big advantage when the technologies isn't ready. Better trained for non US roads when the models are good enough.
diplomat33
Average guy who loves autonomous vehicles
You can't judge FSD on just visualizations. Also, keep in mind that clip is just Mobileye's vision-only stack that Mobileye only uses for L2. Mobileye combines the vision-only stack with a radar/lidar stack for their actual L4 autonomy. Also, the objects bouncing around could be glitches in the visualization and not a product of the actual camera vision. Mobileye has state of the art camera vision.
Mobileye is able to do L2 FSD with vision-only, just like FSD Beta, that is as good or better than FSD Beta based on the unedited videos I've seen and Mobileye can do L4 driverless with their full stack. So I consider Mobileye to be ahead of Tesla in FSD.
Here is a full unedited trip of the full stack that Mobileye uses for L4 robotaxis:
EVNow
Well-Known Member
That they carefully chose and put out. We have been through this multiple times before.
Even I can put out great FSD Beta videos by carefully selecting the best trip - even though my disengagement rate is < 5 miles. And an individual like Omar can put out more than an hour of unedited drive. He doesn't have the resources that a multi-billion dollar company has, either.
diplomat33
Average guy who loves autonomous vehicles
I acknowledge that videos can be curated. That's why I said "based on the videos I've seen". But, I would point out that Mobileye uses AV maps which help with reliability. Also, Mobileye has a standardized driving policy, RSS, which regulates the car's behavior. Because of these things, Mobileye's driving will be more reliable and consistent across different areas. So, I think the videos would be a pretty good representation of overall driving. It is unlikely that Mobileye's driving would be substantially worse than the videos.
But I think there is evidence to support my opinion that Mobileye's FSD is better than Tesla's beyond a few curated videos. I am not judging Mobileye on just a handful of curated videos. I am trying to look at the big picture:
Mobileye's vision-only has a MTBF above 1000 hours of driving per Prof Shashua interview. Mobileye uses AV maps and RSS to increase reliability and safety. Mobileye also has increased redundancy with radar/lidar subsystem which improves reliability. And Mobileye is testing L4 robotaxis.
Tesla does not have camera vision with MTBF of 1000 hours of driving. Tesla has disengagement rate about 1 per 5 miles (your experience and mine). FSD beta makes frequent mistakes that we don't see in Mobileye videos. Tesla lacks sensor redundancy. Tesla lacks AV maps or RSS. Tesla is not testing L4 robotaxis.
Based on that information, I think Mobileye has better FSD than Tesla. But at the end of the day, we can only make a personal judgment based on the information available. Nobody has all the driving data from both Tesla and Mobileye to do an objective analysis.
If Mobileye controls the only videos demonstrating performance, then it would be exceptionally naive to accept them as being representative of the vehicle's performance.
diplomat33
Average guy who loves autonomous vehicles
I think you are missing my point. Mobileye uses AV maps which help with reliability no matter what the road is. So even on completely different roads, since Mobileye maps them first, the car's performance will be similar. And Mobileye uses RSS that guarantees consistent safety behavior regardless of the road. So even in different traffic scenarios, the car will have similar behavior since it is guided by the same policy. Plus, Mobileye's vision has a MTBF above 1000 hours of driving. So yes, I think we can expect the videos to be generally representative of vehicle's overall performance. Of course, I am sure there are routes that Mobileye handles worse than others. But I am saying is that it is highly unlikely that the videos are total outliers and that Mobileye's FSD is much worse than FSD Beta everywhere and that the videos represent the only miles that Mobileye's FSD is able to do competently. That is very unlikely for the reasons I mention (MTBF of 1000 hours of driving, AV maps, RSS).
Put simply, I am not saying the videos are a perfect representation of Mobileye's FSD everywhere. But I am disagreeing with the implication that the videos are some sort of giant PR scam because Mobileye is only showing carefully curated routes and hiding that the system is actually terrible everywhere.
Last edited:
Similar threads
- Article
