Autonomous Car Progress

[daniel]

I don’t think Waymo tech will ever make it to consumer cars in my lifetime - that’s why I only find it mildly interesting.

I can understand that. I don't share your pessimism, so I do find it interesting.

I’m more interested in what I can get in my car rather than what someone somewhere can experience - without it ever being available in my car.

Me, too, except that if someone somewhere else can experience it, then it's possible here, which is exciting.

Besides, a fully autonomous car would be somewhat boring. No different than taking a Uber today.

Driving is boring. Not having to drive is less boring. And the difference between an autonomous car of my own and taking an Uber is that I can get in my car whenever I want, and I can park it so it's waiting for me when I'm ready to return home. No need to wait for the Uber. And no need to worry about getting a cell signal to call for the Uber to take me home again.

An autonomous robotaxi is hardly any different than taking an Uber, but a car of my own is a whole different thing, which is why nobody is selling their personal cars and taking Ubers instead.

 

[EVNow]

An autonomous robotaxi is hardly any different than taking an Uber, but a car of my own is a whole different thing, which is why nobody is selling their personal cars and taking Ubers instead.

Yes - autonomous robotaxi isn't very different from Uber. But - from a technical POV - a fully autonomous car is boring (atleast once the novelty wears off), unlike something that improves every release. May be thats just for us nerds ...

BTW, a lot of people, mostly young have either sold off their cars or never bought a car because of Uber/Lyft. Infact a lot of them don't even have driving license.

Half of young people turn their backs on their driver's license - Politically Now

Madrid

politicallynow.com

Why Aren't Young People Interested in Getting Licences or Buying Cars?

On this Earth Day and in the middle of this Extinction Rebellion, I like to think that it's because they see what is coming down the road.

www.treehugger.com

 

[diplomat33]

Yes - autonomous robotaxi isn't very different from Uber. But - from a technical POV - a fully autonomous car is boring (atleast once the novelty wears off), unlike something that improves every release. May be thats just for us nerds ...

BTW, a lot of people, mostly young have either sold off their cars or never bought a car because of Uber/Lyft. Infact a lot of them don't even have driving license.

Half of young people turn their backs on their driver's license - Politically Now

Madrid

politicallynow.com

Why Aren't Young People Interested in Getting Licences or Buying Cars?

On this Earth Day and in the middle of this Extinction Rebellion, I like to think that it's because they see what is coming down the road.

www.treehugger.com

By the way, I think those stats is one reason why so many companies are betting on ride-hailing. They see a long term trend in large cities that favors the ride-hailing model. And if they can "solve FSD" where they don't need safety drivers and don't need remote assistance either, then there is a lot of potential in having autonomous EVs, with the low maintenance of EVs and the low cost of not paying a driver, and the advantage of autonomous cars being able to operate 24/7 unlike human drivers.

 

[diplomat33]

I don’t think Waymo tech will ever make it to consumer cars in my lifetime - that’s why I only find it mildly interesting.

I don't know how old you are but if we are talking decades, then I think Waymo tech will definitely make it into consumer cars by then. I'm 43. I definitely think Waymo tech will make it into consumer cars in my lifetime.

 

[StealthLSU]

By the way, I think those stats is one reason why so many companies are betting on ride-hailing. They see a long term trend in large cities that favors the ride-hailing model. And if they can "solve FSD" where they don't need safety drivers and don't need remote assistance either, then there is a lot of potential in having autonomous EVs, with the low maintenance of EVs and the low cost of not paying a driver, and the advantage of autonomous cars being able to operate 24/7 unlike human drivers.

Yep, and this is why waymo probably doesn't care all that much about being level 5. They could be insanely profitable by being level 4 in many cities and just forever being geofences to certain areas.

Telsa on the other hand kind of needs level 5 since car owners drive everywhere.

 

[EVNow]

By the way, I think those stats is one reason why so many companies are betting on ride-hailing. They see a long term trend in large cities that favors the ride-hailing model. And if they can "solve FSD" where they don't need safety drivers and don't need remote assistance either, then there is a lot of potential in having autonomous EVs, with the low maintenance of EVs and the low cost of not paying a driver, and the advantage of autonomous cars being able to operate 24/7 unlike human drivers.

Yes - if you see my old posts I write a lot about it (and why if Tesla ever gets FSD, their cars will be too expensive to buy).

But at this time, I'm not that interested or excited about companies making more money by employing less people. The world doesn't need a richer Uber or Google.

I don't know how old you are but if we are talking decades, then I think Waymo tech will definitely make it into consumer cars by then. I'm 43. I definitely think Waymo tech will make it into consumer cars in my lifetime.

Or may be someone else will figure out how to do FSD without lidars. So, waymo "tech" becomes obsolete. Remember a lot of "sure bets" in the past have gone bust. Anyway, Waymo definitely doesn't think about consumer cars at present. Not even in their long term plans.

 

[daniel]

Yes - autonomous robotaxi isn't very different from Uber. But - from a technical POV - a fully autonomous car is boring (atleast once the novelty wears off), unlike something that improves every release. May be thats just for us nerds ...

I don't own a car for excitement. Excitement in a car is bad. Safe and boring are the highest ideals for a car. I own a car for transportation, and if I can sleep instead of drive, that's a good thing. And once we have autonomous cars they will still continue to improve. Nothing is perfect so there's always room for improvement.

 

[diplomat33]

Yep, and this is why waymo probably doesn't care all that much about being level 5. They could be insanely profitable by being level 4 in many cities and just forever being geofences to certain areas.

Agree 100%

Telsa on the other hand kind of needs level 5 since car owners drive everywhere.

In terms of autonomous level, yes, L5 makes more sense than L4 for Tesla since owners would not be happy with geofenced L4. But I don't think Tesla necessarily *needs* L5. They could certainly do great with just "L2+". After all, there will be a driver in the car anyway, so requiring driver supervision would not be a big deal. And I think most Tesla owners would be fine with requiring driver supervision. And Tesla could certainly make a lot of money selling "L2+" that is really good and works everywhere.

Having said, there are advantages to L5. If it were safer than manual driving, then it would be a major selling point. Additionally, L5 would offer some useful applications like being able to send your car across town to pick up your kid from school and bring them back or Tesla could deploy a ride-hailing service everywhere in the US. And from a marketing point of view, there would be a real cool factor too. I think people would buy Teslas just for the cool factor of having a fully autonomous car.

 

[EVNow]

But I don't think Tesla necessarily *needs* L5. They could certainly do great with just "L2+". After all, there will be a driver in the car anyway, so requiring driver supervision would not be a big deal. And I think most Tesla owners would be fine with requiring driver supervision. And Tesla could certainly make a lot of money selling "L2+" that is really good and works everywhere.

Musk won’t stop at L2+ anyway. They will actively develop FSD for several years going forward. Continuous improvement is what I’d expect. Whether and when they get to L4/L5 is the question.

 

[EVNow]

I don't own a car for excitement. Excitement in a car is bad. Safe and boring are the highest ideals for a car. I own a car for transportation, and if I can sleep instead of drive, that's a good thing. And once we have autonomous cars they will still continue to improve. Nothing is perfect so there's always room for improvement.

Excitement is not when you drive - per se - but the fact that AP would continue to improve.

Hey - better to see improvements in the car you have than the car you will likely never have.

 

[daniel]

Agree 100%



In terms of autonomous level, yes, L5 makes more sense than L4 for Tesla since owners would not be happy with geofenced L4. But I don't think Tesla necessarily *needs* L5. They could certainly do great with just "L2+". After all, there will be a driver in the car anyway, so requiring driver supervision would not be a big deal. And I think most Tesla owners would be fine with requiring driver supervision. And Tesla could certainly make a lot of money selling "L2+" that is really good and works everywhere.

Musk promised that we'd be able to use the car as a robotaxi or to haul the kids around. And I'm looking forward with trepidation to the time, not all that far away, when I'm too old to drive safely.

However, geofencing is not the only difference between L4 and L5. I would happily accept certain restrictions. For example, if the car cannot drive in snow that would not affect me at all.

I love the Level 2 features I have right now with EAP. But I will buy the first sleep-in-the-back car I can get. I hope that's a Tesla. But if another company offers to sell me that car first, it's bye-bye Tesla.

 

[Bladerskb]

I'll bet 10 likes Tesla will have a no disengagement coast-to-coast drive (excluding supercharger parking lot) before Mobileye launches any L4 (driverless duh) service in an area larger than 20 sq. miles.

For a company that had supposedly won and it was ‘game over’. This is such a disappointing bet. A 99% highway drive?
I thought Telsa was 5-10 years ahead?

Looks like Tesla is worse than you thought so I will suggest a better bet. Mobileye, Huawei, Tesla. Who will be first to release door to door L2 to all qualified customers that works virtually anywhere in their respective country.

Tesla has the upper hand because China is exponentially harder to drive in than the US. Qualified customer is everyone who has paid for the option.

So what’s your pick and order?

 

[daniel]

For a company that had supposedly won and it was ‘game over’. This is such a disappointing bet. A 99% highway drive?
I thought Telsa was 5-10 years ahead?

Looks like Tesla is worse than you thought so I will suggest a better bet. Mobileye, Huawei, Tesla. Who will be first to release door to door L2 to all qualified customers that works virtually anywhere in their respective country.

Tesla has the upper hand because China is exponentially harder to drive in than the US. Qualified customer is everyone who has paid for the option.

So what’s your pick and order?

Speaking for myself, I would not pay for door-to-door L2. I already have excellent highway L2. In the city, situations arise so suddenly that L2 involves so many disengagements it's not worth it. And if it's good enough that disengagements become rare, it could be hard to react in time when one does arise. I don't feel that L2 in the city is worth much.

The thing I would pay for is L3: No need to intervene unless the car gives at least 5, but preferable 10 seconds warning. I would pay $1,000 for highway L3, or $20,000 for door-to-door L3 provided that disengagement warnings were relatively infrequent, including on South Kihei Road, which is a major high-traffic road here, and the only north-south artery through Kihei other than the highway, but has just two lanes and in places no shoulders, so that pedestrians and bicycles present a very serious problem for FSD.

 

[EVNow]

In the city, situations arise so suddenly that L2 involves so many disengagements it's not worth it. And if it's good enough that disengagements become rare, it could be hard to react in time when one does arise. I don't feel that L2 in the city is worth much.

City is a nebulous term. I use AP all the time in my suburban life - and it’s very useful. If it can handle stop lights, traffic lights and turns, even better. Won’t mind monitoring and taking over when needed. I think AP + me is better tan just me.

There won’t be a sudden L3 availability - I expect L2 to get progressively better and hit L3 some day (2 or 3 more years?).

 

[diplomat33]

At IAA Mobility 2021, BMW demo'ed their automated valet parking project. There is also automated charging and car wash.

 

[scottf200]

City is a nebulous term. I use AP all the time in my suburban life - and it’s very useful. If it can handle stop lights, traffic lights and turns, even better. Won’t mind monitoring and taking over when needed. I think AP + me is better than just me.

100% agree and I use AP all the time as well in our Chicagoland suburbs. I like having eyes all around especially when I do things like lane change. I typically let it do the lane changes on 35-50 MPH suburb roads, for example, because it 'sees' better.

 

[Dutchie]

100% agree and I use AP all the time as well in our Chicagoland suburbs. I like having eyes all around especially when I do things like lane change. I typically let it do the lane changes on 35-50 MPH suburb roads, for example, because it 'sees' better.

Same here. I feel I am a better driver for it as I am much more aware what is going on around me.

 

[daniel]

City is a nebulous term. I use AP all the time in my suburban life - and it’s very useful. If it can handle stop lights, traffic lights and turns, even better. Won’t mind monitoring and taking over when needed. I think AP + me is better tan just me.

There won’t be a sudden L3 availability - I expect L2 to get progressively better and hit L3 some day (2 or 3 more years?).

In Spokane I used autosteer on Sprague Ave, the major east-west thoroughfare. There was enough traffic that there were nearly always cars in front of me at stop lights. But on the residential streets, with a stop sign at nearly every intersection, or every other, autopilot wasn't worth bothering with. And here, there's no place it would be safe other than the highways. Maybe you could risk it on the major streets in Kahului, but I wouldn't.

Of course I don't know how the latest FSD beta would do, but I would not try it even if my car had it.

 

[daniel]

Same here. I feel I am a better driver for it as I am much more aware what is going on around me.

I feel the same way about autosteer on the highways.

The city streets where I live just aren't suitable for it.

 

[powertoold]

At IAA Mobility 2021, BMW demo'ed their automated valet parking project. There is also automated charging and car wash.