Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Waymo

This site may earn commission on affiliate links.
@diplomat33 has shared some posts from Warren Craddock and it is a really good place to start when understanding ADS. Most of us aren't experts in the field and is good to have a basic understanding so we can communicate effectively. There are some things I had the wrong understanding on that is cleared up by some of his post so I will share some especially because I recently had an unpleasant discussion about perception, sensor fusion and how it is not the most difficult part of solving autonomous driving.

AV Myth #7: "If two sensors disagree, there's no way to figure out which one to believe." Fact: The disagreements are the *entire point* of using multiple types of sensors. If they never disagreed, then they'd be 100% redundant, and would offer no additional value. 1/n

Radar does well in fog, because fog is essentially transparent at radio wavelengths. Cameras do well in daylight, because the sun is bright. Lidar does well at night, because it produces its own light. Each sensor sees the world differently, and they disagree all the time. 2/n

The disagreements between sensors are intentional and desirable. The disagreements mean the AV gets much richer data than it would from any individual sensor. The sensors are much stronger together than they are individually. 3/n

Sensor fusion is a mature field. Medical imaging is a prominent example, where PET and CT scanners are bundled together. The PET system detects positrons, but can't see the body itself. The CT system sees the body, but can't detect positrons. 4/n
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1987321/

Sensor fusion is never about trusting one sensor over another. Instead, data from multiple sensors is combined and fed into a single neural net, which learns to make the best use of the salient features provided by all the sensors. 5/n https://arxiv.org/pdf/2005.09202.pdf

An aside: If you believe neural nets are sufficient to drive cars more safely than humans, then you must also believe that neural nets are able to competently combine inputs from multiple sensors -- a vastly simpler task. 6/n

Some of the most exciting work in the AI research community today is explicitly multi-modal, e.g. DeepMind's Gato. An AI system can be much stronger when it is given text, images, sound, etc. all associated with the same event. 7/n
https://www.deepmind.com/publications/a-generalist-agent

This is of course also true for humans: we learn more effectively when we're taught something in a way that uses all our senses. Even kindergarteners understand that. 8/n

In summary, we use multiple sensing modalities (cameras, radars, lidars) precisely *because* they sometimes disagree. Together, they provide a fuller and more complete picture of the world than they do individually. And we put it all into the same neural nets, anyway! 9/9
Craddock's points are irrelevant -- he got 6 likes and 3 replies. Musk tweets some random gibberish and gets eleventy billion likes and replies. That's how you win in today's world. That's why Tesla collects billions in FSD revenue while Waymo collects $118.53.
 
Myth #0 : Only employees of Google and other Tesla competitors are totally unbiased real experts. They should be taken at absolute face value without any questions. Anything Tesla or its employees say should be discounted.

Don't be silly. Nobody is saying to automatically discount Tesla experts or that Tesla competitors are perfect unbiased experts. Warren is correctly pointing out things that many Tesla fans say that are objectively false. And if Tesla employees repeat these myths, then yes, we should point it out. We are simply saying that things that are false should be discounted.

And guys like Dolgov, Anguelov, Jesse Levinson and others have years of experience with AVs and ML. They are experts, even pioneers in autonomous driving. You can't just ignore them because it does not fit your Tesla narrative! It seems to me that Tesla fans want to ignore any experts that don't parrot Tesla FSD talking points.
 
Last edited:
  • Like
Reactions: 2101Guy
For now maybe. But Waymo is testing in way more areas though. The driverless areas will expand over time. Eventually driverless cars will operate everywhere.
Already more and more cities are designating car-free areas. The future belongs to self-driving trucks and buses (not so sexy eh?).

No one, including Waymo, is trying to lower the cost of self-driving cars — they’re not a scalable solution.

Ultimately, self-driving vehicles will bring the end of private cars except, perhaps, for the ultra elite. Not unlike how the automobile brought the end of private horse-drawn carriages except for the Royal Family.
 
Yes but you are trying to change the subject.
Not really. Just exposing the hypocrisy of "critics".

Every curated Waymo (and Mobileye etc) video is presented as evidence of their great self-driving while every 3rd party 1x raw video of Tesla FSD is discounted and people called "shills".

Also, let's not turn this thread into bashing TMC posters that you feel are "anti-Tesla". This thread is not the place for that. Every TMC should be able to express whatever opinions they want.
Oh really ?! How do you feel about people sh*t posting in Tesla threads ?
 
Waymo like to be very secretive about everything. I don't think the data will shed a favorable light on its service. Remember the time we learned from a safety driver that there were usually less than 5 cars available for rides in Chandler (correct me if I'm wrong on that, but I posted the vid of it)?:

“If Waymo’s competitors obtained access to Waymo’s trip-level data, they could analyze the data to gain valuable insights into Waymo’s customer base, fleet utilization rate and optimization capabilities, marketing strategies, and other critical aspects of its business that Waymo does not publicly disclose, causing Waymo irreparable harm,” the company wrote in its July 14 confidentiality request to CPUC.

 
  • Funny
Reactions: diplomat33
Waymo like to be very secretive about everything. I don't think the data will shed a favorable light on its service. Remember the time we learned from a safety driver that there were usually less than 5 cars available for rides in Chandler (correct me if I'm wrong on that, but I posted the vid of it)?:

“If Waymo’s competitors obtained access to Waymo’s trip-level data, they could analyze the data to gain valuable insights into Waymo’s customer base, fleet utilization rate and optimization capabilities, marketing strategies, and other critical aspects of its business that Waymo does not publicly disclose, causing Waymo irreparable harm,” the company wrote in its July 14 confidentiality request to CPUC.

Tesla doesn’t provide FSD beta trip level data either. I would be very curious to see Tesla‘s disengagement data too.
In Waymo‘s safety report they only reported 60k driverless miles so 5 cars sounds plausible. Number of miles is a better metric than number of cars.
 
  • Like
Reactions: Doggydogworld
Remember the time we learned from a safety driver that there were usually less than 5 cars available for rides in Chandler (correct me if I'm wrong on that, but I posted the vid of it)?:
I think it was 5-10. Even that was massive overkill because they had virtually no customers. Actual trips provided only require 1-2 cars, but wait times would be too high if you tried to cover 50 square miles with only 1-2 cars. They stationed a car every couple miles to keep wait times down, but those cars mostly sat around waiting for calls that never came.

They also had hundreds of spare Pacificas available, enough to handle 100x as many customers as they actually got. They thought they'd need all those spare vans just in Chandler, and would need the 80,000+ additional Pacificas and iPaces they ordered to cover the rest of Phoenix. But the customers stayed away in droves. That's why I keep saying the business model failed.
 
  • Funny
Reactions: enemji
Waymo like to be very secretive about everything. I don't think the data will shed a favorable light on its service.

Waymo did withhold a CA DMV report that they say would have revealed trade secrets. That's because their report to the CA DMV had answers to questions about HOW their tech works. But Waymo shares the annual CA DMV disengagement report. Waymo published a detailed safety report that analyzed every accident their autonomous driving had in Phoenix over an entire year period. They also did a detailed analysis of how their AV would have handled real human accidents. We also have the CA DMV accident report every year which details every single collision that Waymo encounters both in autonomous mode and manual mode. That is quite a bit of data. And more importantly, Waymo has done over 500k driverless miles which shows confidence in their AVs. They would not be able to do that many driverless miles if their AVs had serious flaws.

Frankly, I think it is becoming more and more obvious that Waymo is leading in terms of actual reliable driverless since Waymo is letting the public ride in fully driverless cars in more and more locations and expanding. So you are resorting to wishful thinking and speculation. You are hoping that there is some "secret" data that will reveal some major problem.
 
In which locations or areas has Waymo driverless expanded to the general public recently?

Downtown Phoenix. Yes, they are "early access" but they are members of the general public, ie not employees. So it counts. And as soon as Waymo is content with the results, they will open up driverless to everybody. It's inevitable.