Poll will be open for 31 days (you can change your vote until then) and votes will be displayed publicly. Sharing your reasoning here is welcomed.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Poll: what company is leading in driving automation?
- Thread starter shrineofchance
- Start date
Tam
Well-Known Member
Don't quote me and do quote Tesla telling California DMV about its FSD, and not just the beta, it's the "final release":
“As such, a final release of City Streets will continue to be an SAE Level 2, advanced driver-assistance feature.”
In the meantime, Waymo is doing level 4 in 50 squared miles near Phoenix, Arizona with no deaths in its entire history from Google to Waymo.
Waymo is leading the pack in the California Disengagement report:
![dr2020-miles-per-disengagement[1].png](https://teslamotorsclub.com/tmc/attachments/dr2020-miles-per-disengagement-1-png.646176/ "dr2020-miles-per-disengagement[1].png")
The problem is, Waymo doesn't sell its level 4 to me so I just have to get level 2 from Tesla.
“As such, a final release of City Streets will continue to be an SAE Level 2, advanced driver-assistance feature.”
In the meantime, Waymo is doing level 4 in 50 squared miles near Phoenix, Arizona with no deaths in its entire history from Google to Waymo.
Waymo is leading the pack in the California Disengagement report:
The problem is, Waymo doesn't sell its level 4 to me so I just have to get level 2 from Tesla.
diplomat33
Average guy who loves autonomous vehicles
Just to be more specific, Waymo's L4 is only available to the public in a 100 sq mi area in Phoenix right now. But Waymo is "doing L4" in many more areas. The Waymo Driver is L4. So anytime, a Waymo vehicle is in autonomous mode, Waymo is "doing L4", whether there is a safety driver or not. So the CA disengagement report comes from Waymo "doing L4" in CA. Waymo has trucks "doing L4" in Texas. Waymo also has test vehicles "doing L4" in Seattle, Detroit, and Orlando. So Waymo is "doing L4" in many areas outside of Phoenix, AZ. The 100 sq mi area in Phoenix is just where the L4 is available to the public.
Tam
Well-Known Member
In 10/08/2020 article:
"Waymo operates in about a 100-square-mile area. The driverless or “rider only” service area that will be offered to Waymo One members is about 50 square miles, Krafcik said."
That sounds like 100 square miles are for an operation with safety driver onboard while 50 square miles are for driverless without any human safety driver on board.
But thanks for clarifying that whether there's a safety driver on board or not, it's still L4 and not just in Phoenix but also other locations too.
Last edited:
DanCar
Active Member
Tesla is leading, not from a technology perspective, but from an ability to take risk. Waymo in its current form is a dead duck, losing billions each year. Waymo won't widely release until there is no blood risk. And there will always be risk. Google has been selling it's shares of Waymo via funding rounds. Once Waymo no longer is controlled by Alphabet, and someone willing to take risk takes over, then Waymo has a chance to succeed. In its current form, Waymo has no chance of success and will remain an excellent technology showcase.
Mobileye is great but also suffers from a lack of ability to take risk. Will be a couple of years ahead of Tesla in the lab, but a couple years behind Tesla in what is readily available to us here in the U.S.
Mobileye is great but also suffers from a lack of ability to take risk. Will be a couple of years ahead of Tesla in the lab, but a couple years behind Tesla in what is readily available to us here in the U.S.
Last edited:
Tesla is the Dyson & Bose of the car automation world. They have a product that they hyper-market so most people have heard of them. They all have a rabid fanbase. If there are deficiencies in their products, that gets glossed over by the marketing because people "know" the brand and hear it's a premium brand with super exciting articles being written.
That's not to say their products ARE better or worse than their competitors, just they are over-marketed to appear better, or at least appear everywhere. Tesla has captured imaginations and they do make a fun product.
That being said, I don't know much about all the autonomous brands. Joe Public hears about Tesla from media hype so that's all we really know. From the reality of YouTube drives, the extreme breathless predictions from Elon, and the random attention from crashes.
Waymo appears to be ahead in the game in the USA if the metric is actual public access to autonomous cars, at least in their geo-fenced area. Most everyone else is just working on it. When I look at videos of Waymo, Mobileye and Cruise, they are making good driving decisions which I consider important. Tesla is making terrible driving decisions.
I want to see Tesla solve everyday driving, safe merging, and not heading for trouble. Do that before working on edge cases. Show me they can actually do that. Some videos claim this is happening, then we see countless videos with just terrible driving. Confusing.
Here's a ridiculous cherry-picked video from Cruise. Very messy driving from everyone but seems they can do it, I guess. Before Tesla drives like this I want to see them handle regular driving without screwing up.
That's not to say their products ARE better or worse than their competitors, just they are over-marketed to appear better, or at least appear everywhere. Tesla has captured imaginations and they do make a fun product.
That being said, I don't know much about all the autonomous brands. Joe Public hears about Tesla from media hype so that's all we really know. From the reality of YouTube drives, the extreme breathless predictions from Elon, and the random attention from crashes.
Waymo appears to be ahead in the game in the USA if the metric is actual public access to autonomous cars, at least in their geo-fenced area. Most everyone else is just working on it. When I look at videos of Waymo, Mobileye and Cruise, they are making good driving decisions which I consider important. Tesla is making terrible driving decisions.
I want to see Tesla solve everyday driving, safe merging, and not heading for trouble. Do that before working on edge cases. Show me they can actually do that. Some videos claim this is happening, then we see countless videos with just terrible driving. Confusing.
Here's a ridiculous cherry-picked video from Cruise. Very messy driving from everyone but seems they can do it, I guess. Before Tesla drives like this I want to see them handle regular driving without screwing up.
Last edited:
You're assuming Tesla ever had the lead.
Look at what I posted in 2017 at Autonomous Car Progress. Notice the Waymo (the current name for Google/Alphabet's self-driving efforts) Phoenix deployments in 2017? Notice the publicity stunt video posted in 2012? Google/Waymo's been at it for a long time.
Notice what Nissan (not at all a leader in the space) did in a video posted March 30, 2017? Look at the CA autonomous miles driven and miles per disengagement at Disengagement Report 2017 – The Good, The Bad, The Ugly for Google/Waymo for 2015 thru 2017?
Since the CA DMV removed a lot of their content and redid their pages, if you want to see what the reports to the CA DMV for 2015 and 2016, see these:
Although not exactly apples to apples, compare to what you see posted in the past few weeks at FSD Beta Videos (and questions for FSD Beta drivers) and FSD Beta Videos (and questions for FSD Beta drivers).
Compare to what Cruise Automation has posted at https://www.youtube.com/channel/UCP1rvCYiruh4SDHyPqcxlJw/videos and the types of cases they've posted at Why testing self-driving cars in SF is challenging but necessary (in Oct 2017).
powertoold
Active Member
How do we even define "leading" when the endpoint is unclear?
This sort of open-ended question has led to heated discussions based on biased metrics.
Also, the question in this poll is poorly worded and focused. What does it mean "driving automation"? If that's referring to the SAE levels, then clearly Waymo is leading. If it means "which company is leading in driving automation deployment," then that's Tesla (most widely deployed automation features).
This sort of open-ended question has led to heated discussions based on biased metrics.
Also, the question in this poll is poorly worded and focused. What does it mean "driving automation"? If that's referring to the SAE levels, then clearly Waymo is leading. If it means "which company is leading in driving automation deployment," then that's Tesla (most widely deployed automation features).
Tam
Well-Known Member
I don't think L4 in a geo-fenced area is not that bad of a deal at all.
That's because most people drive mostly on a very repeatable route such as from home to work, to school, to the supermarket...
So if someone can sell an L4 with limited customized routes, that would be a great start!
Last edited:
Tam
Well-Known Member
That's one perspective.
L0 (manual driving) is widely available to consumers, but I would not call that's leading.
If L2 is widely available to consumers still with crashes and deaths, I would not call that leading.
I think in medicine or science, leading is when it's verifiable for its most achievement in the set standards. If a vaccine with an efficacy of 10% and another 95%, then it doesn't matter whether which vaccines are available, better numbers mean better numbers after verified by others.
In Self-driving, it's SAE levels. Google was able to prove that its Prius machine without a driver could drive a 95% blind person in 2012, 2 years after its inception, that's leading:
In 2016, it again drove a legally blind passenger around the Texas Capitol for 10 minutes without a steering wheel or pedals. That's leading!
It's not available to all consumers but it's verifiable. It has earned an L4, that's leading.
Last edited:
Tam
Well-Known Member
I've hardly heard any probes against L4 Waymo but for L2 Tesla:
"The U.S. auto safety agency disclosed on Thursday it has opened 27 investigations into crashes of Tesla vehicles, 23 of which remain active, and at least three of the crashes occurred in recent weeks."
Tesla has a long-term concept of how L5 robotaxi will be achieved and especially without the crutch of LIDAR.
However, even at L2, Tesla can't make the grades in reigning in the accidents.
Tesla has a long-term concept of how L5 smart summon will be achieved so that we won't even need a human valet attendant anymore.
It's a good long-term goal, but currently, smart summon is so unreliable to get a passing grade for competency.
Manual driving isnt an autonomous system. It seems...well.... stupid to use that as an example of why using how widely an autonomous driving feature is available is flawed. Nice try tho.
If a vaccine is only available to one person who cares how effective it is, the rest of us will all be dead.
The entire topic, including the topic of this thread, has been discussed, and re discussed, and re re re re re discussed by the same like 9 people in this forum for 2 years.
DanCar
Active Member
Do we get a gold star after we've discussed it for 10 years?
Tam
Well-Known Member
That's the whole point. Which one is proving that L4 is achievable? The one with Tesla L2 that drivers should be onboard or accidents and deaths will continue to happen or the one with L4 with blind people onboard with minor accidents and no deaths since 2010?
Which ones prove L4 is possible? The ones with drivers who passed the vision test or the ones with no drivers and its occupants are either blind or don't have a driver's license to drive?
Last edited:
Similar threads
