Autonomous Car Progress

[Matias]

If you don't think there's really people being paid to badmouth Tesla let me introduce you to one Gordon Johnson

There might be, but I don’t think that they post anonymously on TMC.

 

[Bladerskb]

For distance? Why would it need to?

You don't design a driving system to always stop 0.000001% from hitting the car in front of it after all. So if LIDAR can tell me the car is EXACTLY 22.6748991 meters away... but vision can "only" tell me it's only about 22.67 meters away, that's still PLENTY accurate for the task at hand.


You want to design it to stop with a MUCH larger distance between objects than that- both because you can't be 100% certain of stopping distance given variable conditions, and because you don't want to be shoved into the car in front if the guy behind didn't brake in time.

So once again not using mm-level requirements makes the system generally easier to scale and not require otherwise unneeded and expensive sensors.

who said anything about distance? I and every single logical person is talking about the entire system. There are no ML system in any category that is even close to the failure rate required of a sdc system. Even with limitless training data from users (Google, MS, Apple, Amazon, etc)

 

[Doggydogworld]

If only they hadn't wasted all that money on LIDAR and HD maps

Recurring cost of LIDAR and maps is 0.23 cents/mile*. A rounding error for a service that charges 50-100 cents/mile.

*not an actual calculation, but you get the idea

Reliable vision-only RTs, in low-TCO EVs, it'd be almost impossible NOT to make money on- and in significantly larger amounts than currently profitable taxi offerings in a lot more areas.

Taxis are a niche business. And again, vision-only is just cult marketing. It works for consumer vehicles where every penny matters and bulbous protrusions are shunned, but it's pure downside for a Robotaxi business.

people think the google car thing is to make money?

you dont know google very well, then. they are in it for the long game. they are NOT a product company, they are a r/d company.

Google is a scale company. But even they can't scale Robos with high operating losses. And they've made no meaningful attempts to find a business model they can scale.

 

[powertoold]

people think the google car thing is to make money?

you dont know google very well, then. they are in it for the long game. they are NOT a product company, they are a r/d company.

Killed by Google

Killed by Google is the open source list of dead Google products, services, and devices. It serves as a tribute and memorial of beloved services and products killed by Google.

killedbygoogle.com

 

[Daniel in SD]

Taxis are a niche business.

Uber's gross bookings are over $60 billion a year and growing. That's a pretty big market.

 

[linux-works]

Google is a scale company. But even they can't scale Robos with high operating losses. And they've made no meaningful attempts to find a business model they can scale.

business model, other than advertisements (the ONLY thing they make money at, and they make more than god himself at it) - they are not a company in the usual sense. they are not pressured to 'capture market dominance' or any other nonsense like that.

again, long game, research, become an expert and see where it can go.

they have all the profit they need to keep running, in general (google) and what drives other companies is not what drives them.

I dont know why people refuse to see that. they are NOT competing with tesla and such; they are just doing their own thing in their own way.

its not a race even though the sports fans here seem to WANT to make it into a vendor A vs vendor B thing.

not everything in life is like that.

 

[Knightshade]

who said anything about distance?

LITERALLY THE POST YOU WERE REPLYING TO DID.

Do you not read the posts before replying to them?

That would explain a lot

The primary thing lidar provides, that previously was not being done with vision, is providing accurate distance to the objects.

Tesla is now doing that with vision.

If they can obtain accuracy needed for safe driving then LIDAR no longer adds any value at all.

you then replied asking if it was 99.99999% accurate.

I said "For distance? why would it need to be"


And suddenly you had no idea what distance had to do with this?

I and every single logical person is talking about the entire system. There are no ML system in any category that is even close to the failure rate required of a sdc system. Even with limitless training data from users (Google, MS, Apple, Amazon, etc)

You're making two mistakes here:

1) Including yourself in the set of "logical people"

2) Confusing "is the distance or speed data derived by the vision system 99.9999% accurate- even though there's no reason it needs to be nearly that accurate" and "What is the FAILURE rate (where it fails in a way that actually matters to the driving task)"

Those are apples and hand grenades.

A "logical" person would see the obvious difference

 

[Knightshade]

Recurring cost of LIDAR and maps is 0.23 cents/mile*. A rounding error for a service that charges 50-100 cents/mile.

*not an actual calculation, but you get the idea

No, I don't.... if you have an actual calculation that supports your claims then we might.

Or it might not end up actually supporting them which would be somewhat awkward for you.

Taxis are a niche business

Taxi revenue just in the US was about 15 billion dollars in 2019.

Ridehailing revenue is many times larger than that.

Some niche.

And RTs of course allow you to expand it massively.

Having a taxi sitting around with a human driver waiting for a call in a rural area is nonsensical, economically.

But an EV owner who lives in a rural area letting his car run fares the 12-16 hours a day he's not using it suddenly, massively, expands your targetable market.

. And again, vision-only is just cult marketing. It works for consumer vehicles where every penny matters and bulbous protrusions are shunned, but it's pure downside for a Robotaxi business.

.... can you unscramble this word omlette a bit?


How is a system currently being testing broadly by a fleet of over a million cars, and narrowly (city streets) by ~2000 cars, "cult marketing"?

How is it, assuming reliably can be developed and proven safe, a "downside" to have much lower per-vehicle costs for your taxis?

There might be, but I don’t think that they post anonymously on TMC.

Do you not read a lot of threads here?

There's tons of em where you've got someone who just joined, has 0 previous posts, then posts about "OMG MY TESLA RAN OVER MY DOG THEN EXPLODED MY HOUSE AND TESLA TOLD ME THEY ARE NOT RESPONSIBLE THEN ELON MUSK KICKED MY BABY!"

(I exaggerate- but not by much)

And you see the same stuff in the comments of many many elektrek, teslararti, etc stories-- rando anon comments spreading FUD about Tesla (and often EVs in general)

 

[diplomat33]

Funny you don't hold your favorite company to that same exact standard. Except in Chandler, AZ, which is hardly NYC or SF (the latter apparently now being driven by your favored company with safety drivers).

I do hold Waymo to that standard. Waymo has removed driver supervision for the public in Chandler. Tesla has not removed driver supervision in any area yet. So based on who has removed driver supervision the most, Waymo is ahead.

If it takes years of test driving and running a tiny service to gather data in each new city then Waymo might as well give up now.

Yes, it took Waymo years to develop real autonomous driving from scratch and deploy the first service to the public with no driver. Go figure! And it always take longer when you are the first to develop a brand new technology. And nobody has been able to match Waymo yet.

Developing real autonomous driving and deploying it safely is insanely hard. It takes time. You have to make sure your technology is incredibly safe and can reliably handle all driving with no driver supervision at all. If you deploy too soon, you could literally kill people. It's not some easy thing as Elon would like us to believe. But Waymo hopes to accelerate the process with Simulation City so that it does not continue to take years to deploy a service in each new geofenced area:

Bringing our technology to more people in more places
SimulationCity is especially beneficial as we expand our operating domains. Even if we haven't autonomously driven in a specific location, we can upload a map, insert a Waymo vehicle into the scene, and combine the millions of miles we have driven with other data to inform what other roads users could behave like in areas we currently don't have behavioral data on. This process enables us to refine our overall driving abilities and rider experience in a new area before having a single rider in a vehicle. It also means that we can put a lot more vehicles "on the road" – with no constraints on when, where or the number of vehicles we can simulate, and how quickly we can simulate them, we can rapidly speed up our learning and development.

I hate to keep repeating myself, but Waymo's problem is their business model. Forget the billions they've sunk into R&D and whatever -- their service itself loses money every day. They haven't found a path forward to it being able to make money. And they lack entrepreneurial leaders who iterate rapidly until they find that path forward. Assuming the path even exists -- despite the hopes of autonomy fans it's possible Robotaxis are simply a small niche market.

IMO, if Waymo fails, it will be because of their business model, not their tech. Waymo has the best FSD tech by far. But I do acknowledge that having the best tech is not good enough if your business model fails.

If only they hadn't wasted all that money on LIDAR and HD maps

Lidar and HD maps is what helped Waymo deploy L4 first. If Waymo had not used lidar and HD maps, they would be far behind, just like Tesla is.

 

[rxlawdude]

Lidar and HD maps is what helped Waymo deploy L4 first. If Waymo had not used lidar and HD maps, they would be far behind, just like Tesla is.

Yet that's so reliable that Waymo won't make unprotected left turns on intersections like Chuck Cook's videos of FSD, and certainly not without a safety driver.

Your continued assertion that Waymo is so far ahead is getting rather pitiful, unless you show me total autonomy in SF, NYC, or LA. You can't because it doesn't exist. Just like the unicorn of L5 FSD. If the technology is so good at protecting safety, why not have their vehicles drive autonomously with no one in the vehicle when testing it?

Waymo has the advantage of limitless resources. But responsible, fiduciary management at Alphabet will have to determine the ultimate feasibility of scaling it to actually make money. My prediction: Google spins off and sells Waymo before deploying it in any difficult urban environment.

 

[Matias]

Do you not read a lot of threads here?

There's tons of em where you've got someone who just joined, has 0 previous posts, then posts about "OMG MY TESLA RAN OVER MY DOG THEN EXPLODED MY HOUSE AND TESLA TOLD ME THEY ARE NOT RESPONSIBLE THEN ELON MUSK KICKED MY BABY!"

(I exaggerate- but not by much)

And you see the same stuff in the comments of many many elektrek, teslararti, etc stories-- rando anon comments spreading FUD about Tesla (and often EVs in general)

Some of them might be trying to move the stock (which is btw not possible with some random john doe post, but of course someone might try it..). But some are more likely just normal trolls.

I see a lot of misinformed/troll posts in Finnish ev news comments and I’m sure they’re not trying to move the stock. You would be pretty desperate to post Fud in Finnish forums with the idea of trying to move big stocks

And still one point, that people have honestly different opinions. Not everything is FUD if it is not pro Tesla.

 

[linux-works]

Waymo has the advantage of limitless resources. But responsible, fiduciary management at Alphabet will have to determine the ultimate feasibility of scaling it to actually make money. My prediction: Google spins off and sells Waymo before deploying it in any difficult urban environment.

and we all know that elon is just scraping by with ramen noodles. poor guy. someone should loan him some money.

sheesh.

tesla has more money than god. what the hell are you blathering on about again this time?

 

[diplomat33]

Your continued assertion that Waymo is so far ahead is getting rather pitiful, unless you show me total autonomy in SF, NYC, or LA. You can't because it doesn't exist.

Yes, it does exist. I've shown you the CA DMV disengagement data that shows their fully autonomous driving in SF and LA. and there are Waymo autonomous vehicles driving around SF all the time.

https://twitter.com/x/status/1424262852475461633

If the technology is so good at protecting safety, why not have their vehicles drive autonomously with no one in the vehicle when testing it?

Because that is not how good testing is done. You always have a safety driver when testing.

 

[rxlawdude]

Yes, it does exist. I've shown you the CA DMV disengagement data that shows their fully autonomous driving in SF and LA. and there are Waymo autonomous vehicles driving around SF all the time.


Because that is not how good testing is done. You always have a safety driver when testing.

Yet you continuously hold Tesla to a different standard for their beta test, where the driver is the safety driver.

And as we've discussed, uncomplementary data has been known to be carved out of statistics. In multiple industries.

How long did GM make cars with defective ignition switches without federal action?

 

[diplomat33]

Yet you continuously hold Tesla to a different standard for their beta test, where the driver is the safety driver.

No, I don't think so. Tesla is L2, Waymo is L4. There is a big difference between L2 with a driver and L4 with a safety driver.

 

[powertoold]

There's Waymo, and then there's all the other fsd developers using hd maps, who are catching up to Waymo. Waymo is the current upper bound on performance with HD maps.

 

[rxlawdude]

No I am not. Tesla is L2, Waymo is L4. There is a big difference between L2 with a driver and L4 with a safety driver.

Ah, we agree. Tesla IS L2 (beta FSD included). Waymo is L4 in narrowly defined geofenced areas, yet still requires someone behind the wheel? And again, let's see how it does without safety drivers in those urban environments we talked about. If your value proposition for Waymo includes a "safety driver" making presumably minimum wage (here, $14/hr), you've shot the economic model.

And would you want some minimum wage earner being a "safety driver?" Next, you'll say the safety drivers are paid more than that. Which further weakens the economic viability of Waymo.

 

[Daniel in SD]

No, I don't think so. Tesla is L2, Waymo is L4. There is a big difference between L2 with a driver and L4 with a safety driver.

I disagree. Tesla is prototype L5 despite what they tell the CA DMV.

 

[linux-works]

I disagree. Tesla is prototype L5 despite what they tell the CA DMV.

which tesla is a prototype of level 5?

what production cars do tesla have that even come close to asil-d rating?

 

[rxlawdude]

I disagree. Tesla is prototype L5 despite what they tell the CA DMV.

Nope. Not even close.