Waymo brings in $2.25 billion from outside investors, Alphabet

[malcolm]

I mean "autonomous driving". They all have autonomous driving, maybe geofenced, but it still counts as real autonomous driving. Within the geofence, the car is driving fully autonomously.

Believe it or not, I am open minded. After all, I own a Model 3 with FSD. I would love nothing more than to be able to come here and say that Tesla has achieved full autonomous driving. And if Tesla releases "City NOA" and it's good, I will give Tesla credit for it. But I am not going to say that Tesla will release L5 autonomy in 2 years based on the current state of AP.

Would it be fair to say that a key difference for you is that the Tesla is bought and paid for /leased and you have to rely on it to get you from A to B.

So, of course, Ford, Waymo, Baidu, Cruise etc seem more impressive autonomous offerings simply because you don't have to put up with any of their shortcomings.

 

[diplomat33]

Would it be fair to say that a key difference for you is that the Tesla is bought and paid for /leased and you have to rely on it to get you from A to B.

So, of course, Ford, Waymo, Baidu, Cruise etc seem more impressive autonomous offerings simply because you don't have to put up with any of their shortcomings.

Not really. The reason I think Waymo et all have more impressive autonomous driving is because their cars can do a lot more autonomous stuff than Teslas can do. They can drive autonomous in busy city driving. They can stop at red lights. They can make right turns. They can do unprotected left turns with busy traffic. They can pull over for emergency vehicles. They can stop for a construction worker holding up a stop sign. They can navigate a narrow street and move over to let an incoming car pass before resuming driving. They can let an incoming car pass and then navigate around a double parked car. And much more... And they can do all those things autonomously with no driver intervention. Our Teslas cannot do those things autonomously. Based on that, I am going to rank those cars as further ahead in autonomous driving than Tesla since they can do more autonomously than Tesla can.

But I will agree that I probably have less patience with Tesla because I own a Tesla. I have expectations on Tesla since it's the car I use to get from A to B. And Elon has been promising FSD for 3.5 years now. So when I see a Waymo or Cruise car drive completely autonomously and my Model 3 still even can't stop at a red light autonomously, I get a little frustrated at Tesla. And yes, I probably get a little frustrated at "Tesla fanboys" yapping about how camera vision is all you need and Tesla is going to solve autonomous driving any day now when my car is no where near autonomous yet.

 

[MP3Mike]

Not really. The reason I think Waymo et all have more impressive autonomous driving is because their cars can do a lot more autonomous stuff than Teslas can do. They can drive autonomous in busy city driving. They can stop at red lights. They can make right turns. They can do unprotected left turns with busy traffic. They can pull over for emergency vehicles. They can stop for a construction worker holding up a stop sign. They can navigate a narrow street and move over to let an incoming car pass before resuming driving. They can let an incoming car pass and then navigate around a double parked car. And much more... And they can do all those things autonomously with no driver intervention. Our Teslas cannot do those things autonomously. Based on that, I am going to rank those cars as further ahead in autonomous driving than Tesla since they can do more autonomously than Tesla can.

So you are giving Waymo credit for what they demo but haven't released to anybody to use themselves, but you only give Tesla credit for what they have released for you to use on your own car. And you give Tesla no credit for their demo videos or the demo rides that they gave people in their prototype FSD cars. (That you claim they don't have.)

 

[diplomat33]

So you are giving Waymo credit for what they demo but haven't released to anybody to use themselves, but you only give Tesla credit for what they have released for you to use on your own car. And you give Tesla no credit for their demo videos or the demo rides that they gave people in their prototype FSD cars. (That you claim they don't have.)

No. First of all, Waymo robotaxis are available to the public in some areas. So Waymo did release their autonomous driving in the form of ride sharing.

You mention the autonomy day demos. But Tesla never released videos of the demo rides that that gave investors on autonomy day. So there is no video evidence of how good the FSD was. So Tesla won't show us their FSD development. The fact is that I will happily give Tesla credit for autonomous driving if they released videos of their cars driving autonomously on city streets. Heck, if Tesla released a video of their FSD development cars doing what Cruise cars can do, I'd probably geek out like a fanboy again.

Tesla did release this demo:

In comparision, Cruise released several videos of their autonomous car driving for an entire hour in city driving:

Comparing the two videos, I would say the Cruise video is better FSD.
.

 

[malcolm]

So we're back to comparing videos and press releases.

Which was the basis of last year's Navigant Research / Guidehouse / Veritas Capital report into automated driving.

Tesla doesn't spend enough on marketing.

Oh well.

 

[diplomat33]

So we're back to comparing videos and press releases.

Which was the basis of last year's Navigant Research / Guidehouse / Veritas Capital report into automated driving.

Tesla doesn't spend enough on marketing.

Oh well.

Tesla sells a ton of cars with no advertising. So I don't think marketing is the issue. Like I said, Tesla could easily put out their own videos on their FSD development and it would go along way to convince people of their FSD approach and progress.

 

[Daniel in SD]

Tesla doesn't spend enough on marketing.

How much do you think it costs to release an hour of unedited autonomous city driving?

 

[malcolm]

Comparing the two videos, I would say the Cruise video is better FSD

Seriously?

With the camera mounted on top of the car no one has any idea what the driver is doing.

 

[malcolm]

How much do you think it costs to release an hour of unedited autonomous city driving?

I know, right? Cruise does it and everyone believes that they are looking at a self-driving miracle.

I'm going to stick a camera on top of my POS ICE and post the footage as an FSD breakthrough.

Throw in some forward-looking press releases - job done.

My point is that Tesla haven't done it and neither has anyone else.

Anyone can manufacture a slight lead / advantage. Pick a location, ringfence some roads, Lidar scan the crap out of a neighborhood.

Whatever.

 

[Daniel in SD]

Seriously?

With the camera mounted on top of the car no one has any idea what the driver is doing.

So you’re claiming that Cruise is a hoax? Ok. Then there is nothing to discuss.

 

[malcolm]

So you’re claiming that Cruise is a hoax? Ok. Then there is nothing to discuss.

Nope, not at all. What is needed is an equality of scepticism.

We know that Tesla's AP claims do not measure up because people like diplomat33 own the cars and experience the problems for themselves.

But when companies like Cruise or Waymo post a few videos or press releases, suddenly everyone believes them.

The same thing happens to the guys over at Navigant Research. Too easily impressed by glossy websites and curated videos.

Hilarious.

Everyone, including Tesla, crafts videos to emphasize niche advantages of their systems. No one is close to solving the problem on the basis of the evidence presented.

 

[malcolm]

The Tesla video that diplomat33 posted above doesn't tell us anything because we have no idea how much assistance the system had or didn't have to learn the route.

Then what about driving the same route in the rain or in the dark, or even just driving the same route in the opposite direction?

You can (and really should) apply the same questions to self-driving videos from Cruise, Waymo, Mobileye or whoever.

 

[timewasted]

No, there is no need for an equality of skepticism because the companies are not equal in their progress. Cruise had 54 different VINs that drove a combined 831,039.9 miles in the 2019 California autonomous disengagement reports. Tesla, however, had 1 VIN that drove 12.2 miles. Several orders of magnitude difference is not equal.

There is, however, a need to not just disregard facts because it paints your favorite team in a bad light.

 

[MXLRplus]

Some of us want better autonomous operation in our cars ASAP.
And it doesn't matter which brand is selling it.
When Tesla showed me what FSD did as of March 2020, I cancelled the FSD option. It's not worth $7k.

 

[malcolm]

Ah yes disengagement reports.

If you choose your routes, traffic levels, weather conditions, average vehicle speed you can pass any test.

As we've all learned from the emissions scandal even third party testing can be undermined if the manufacturer knows in advance what the testers are looking for.

ATM nobody's autopilot system is worth the money.

 

[malcolm]

There is, however, a need to not just disregard facts because it paints your favorite team in a bad light.

Don't have a favorite team.

Everyone's self-driving system has some niche advantage but too many limitations / things they aren't showing us or telling us.

None are worth the money.

 

[Daniel in SD]

Nope, not at all. What is needed is an equality of scepticism.

I’m as big a skeptic as anyone. A one hour video is pretty useless for proving that a car is capable of self driving as well as a human. Realistically you’d probably need a 1 million hour video to prove that.

Ah yes disengagement reports.

If you choose your routes, traffic levels, weather conditions, average vehicle speed you can pass any test.

As we've all learned from the emissions scandal even third party testing can be undermined if the manufacturer knows in advance what the testers are looking for.

ATM nobody's autopilot system is worth the money.

If you live in San Francisco you can get a pretty good idea of what Cruise is doing because their cars are everywhere. If you can drive around San Francisco and only disengage every 10k miles that seems like they’re making real progress. Obviously it’s only proven when they remove the test drivers. Hopefully someone does that soon so we can stop these silly arguments about who is “ahead”.

 

[Soda Popinski]

There are a number of reporters who have reviewed Waymo's robotaxies. Spend a little time Googling or YouTubing it, and you'll see it's not all press releases and curated videos from Waymo. Credit where credit is due.

 

[diplomat33]

I’m as big a skeptic as anyone. A one hour video is pretty useless for proving that a car is capable of self driving as well as a human. Realistically you’d probably need a 1 million hour video to prove that.

True. But I don't think any demo video is really intended to be a final conclusive proof that an autonomous car is ready to be deployed without a human driver. That's not what they are designed to do. Demo videos are simply meant to show off one or more "features". Basically, they are designed to say "look what I can do autonomously!" For example, we can look at the Cruise videos as basically demonstrating how the cars handle various city driving scenarios. So they do give us some nice visual info on what the car can do. They can also give us hints about how the car behaves in certain situations. For example, we might watch a video and notice that the car is too cautious or too aggressive. But they can't tell us everything the car can or can't do.

 

[DanCar]

There are a number of reporters who have reviewed Waymo's robotaxies. Spend a little time Googling or YouTubing it, and you'll see it's not all press releases and curated videos from Waymo. Credit where credit is due.

Waymo is willing to risk a ride for reporters to show off (in broad daylight, perfect weather conditions, preplanned route), but they are not willing to place the system in real use. They are too afraid of losing millions in a lawsuit. Plus the bad publicity would cause millions more harm to Google. The system is good, but it won't be launching anytime soon, since there will always be risk. Are there any unprotected left turns at busy intersections?