Fully self-driving vehicles

Investment Analysis Clubs Macro Economic Trends and Risks
21

I cannot, under any circumstances, envision anyone offering shared rides with autonomous vehicles, due to safety concerns.

Wase and Google already have a very active car-pooling app. in Mexico, Israel, Brazil, and to some extent in the US. With a CoOp you would be able to have a background vetting process. Lyft and Uber already do collective ride sharing in San Francisco especially from large events. Collectivos have been working in Mexico for decades. When I get a ride to pick up my vehicle from the Service Center that I use, there are often 4 or 5 other people sharing the ride. CoOps could have selected “pods” of people that know each other. I’ve belonged to farming CoOps where I’ve known almost everyone else involved.

I can’t see that happening.

I agree, by the time it might be available, you and I will be gone. I’m sure if the concept of Uber was presented to you thirty years ago, you would say the same thing.

22

FSD cars are here…so are the accidents.

Who is to blame for any such accident? When the laws are set up we will see FSD cars all over and no one steering them.

But the laws can not be set up. Do you make the car company liable? Or do you make the owner who is not driving liable? There is no solution.

23

I’m sure if the concept of Uber was presented to you thirty years ago, you would say the same thing.

Not at all.

The issue isn’t sharing an Uber (or a taxi) with a stranger. It’s sharing a car with an unvetted stranger and no one else.

That’s the problem with ride-sharing an autonomous vehicle. You’re by yourself with whoever else is in the car - and you’re literally miles and miles away from any employee of the company providing you transport. You can’t move to another train car, or a seat in the bus closer to the driver. You’re not a station away from a security guard. There’s no Uber driver in the front seat who would be a witness to any untoward behavior.

If there’s a CoOp that you personally know everyone else involved, you can solve that problem - but I can’t imagine that economically or environmentally, there’s enough gain from the occasional carpooling from such a small group compared to just letting the AV out in the world to ferry passengers for single rides.

Albaby

24

FSD cars are here…so are the accidents.

Who is to blame for any such accident?

There are no fully self-driving cars available. Tesla sells a version of driver-assistance technology that it calls “Full Self Driving,” but the car is not capable of driving without continual driver participation and attention.

Any accident that occurs today that involved a car using any kind of driver assist (smart cruise control, lane stability, etc.) would likely get blamed on both the driver and the car manufacturer.

Albaby

25

can’t imagine that economically or environmentally, there’s enough gain from the occasional carpooling from such a small group compared to just letting the AV out in the world to ferry passengers for single rides.

True.

You’re not a station away from a security guard.

BTW young people move into shared housing in San Francisco with people they’ve never met in person. There are networking on-line groups for people in the same profession moving into the city with no other contacts.
No security guards

1 Like
26

BTW young people move into shared housing in San Francisco with people they’ve never met in person. There are networking on-line groups for people in the same profession moving into the city with no other contacts.
No security guards

Yeah, and no complete strangers, either. They’re moving into shared housing with people who have been vetted by whatever service they’re using (for co-housing that’s arranged by a company) or that they’ve gotten themselves comfortable with as a potential roommate through their on-line network groups. They’ve never met in person, but they’ve met enough to get some degree of comfort before agreeing to share a room.

Not a total random person who calls for a ride.

Certainly there will be some adventurous people who would be willing to get into an unmanned private automobile with a complete stranger. I suspect that enough would not, sufficient to keep the idea of autonomous ride-sharing from ever becoming a thing.

1 Like
27

Hawkwin writes,

aviation kills people and it is not yet illegal

Heh, and we still pay human pilots six figures to move cargo instead of using AI. It would seem that automating planes should be easier than cars but that has not happened yet either.

The FAA recently authorized “autonomous flight” for a commercial drone. Jeff Bezos surely sees the profit in removing the pilots from his air cargo fleet.

https://dronedj.com/2022/08/24/american-robotics-drone-auton…

intercst

1 Like
28

Driverless cars have staggeringly huge advantages when you gave near 100% mutually communicating system controlled vehicles. Far faster, far more economic with drastically minimized changes of speed let alone stopping requiring far less space. A few corridors like that in an urban area or on interstate highways would blow peoples minds.

But we will probably have to wait for us crippled types who confuse car driving with sex drives before the necessary steps will be taken.

The mix kills the benefit. Multi-tasking while driving is insane, and guessing what the car driver ahead of you, instead of the utterly predictable autonomous car, will do will soon be seen as insane.

david fb

1 Like
29

Al,

Self driving cars are definitely here. The accidents come with them.

Making this clearer, no car manufacturer can sell or declare the company sells a fully self driving car without incurring all the liability. Or minimally risking a full bankruptcy.

30

Self driving cars are definitely here.

No, they are not. There is no car that is capable of driving itself commercially available in the U.S., even for short periods of time. The only products that are available are “driver assist” systems, which are capable of performing certain specific driving operations (like adaptive cruise control or lane changing), but are not capable of performing all operations necessary for safe driving at any given time.

This is not just semantics, or manufacturers having to limit their claims about their vehicles because they don’t have regulatory authority to make broader claims. All of the vehicles that have advanced driver assist (GM, Tesla, Ford) that allow for limited “hands-free” driving are only Level 2 systems, and require the driver to be looking ahead, paying attention to the vehicle, and be ready to intervene at any time. They are not self-driving, and at all times the driver is both presumed to be and actually is in charge of driving the vehicle.

Albaby

2 Likes
31

Self driving cars are definitely here.

No, they are not.


Al

No offense....LOL....but I am taking my nephew's word for it over a lawyer's word for it.

He graduated MIT with a computer science degree three years ago.

The law is a a$$. There is no law for who is liable. If you have that we can start to perhaps call self driving cars by their name accidents and all as "self driving cars".
32

No offense…LOL…but I am taking my nephew’s word for it over a lawyer’s word for it.

He graduated MIT with a computer science degree three years ago.

I’m curious what a computer science degree from the trade school down the river has to do with knowledge about the current state of self-driving technology would be…but no matter.

Would you mind asking your nephew which company has manufactured a fully self-driving car? Ordinarily, that type of milestone breakthrough would be all over the news. None of the major automakers has done it, nor any of the companies that are specifically working on autonomy (like Waymo or Cruise) - the best they’ve gotten to is Level 4.

If you could find out who actually had achieved that milestone, it would be really useful to know. Do you mind finding out?

Albaby

5 Likes
33

If you could find out who actually had achieved that milestone, it would be really useful to know. Do you mind finding out?

Albaby

They are all side stepping or refusing to make the claim.

Trade school…LOL…sure

34

They are all side stepping or refusing to make the claim.

No, they’re not. None of their systems are self-driving yet. The most advanced available for purchase is Mercedes’ Level 3 system, which is not self driving. There are numerous pilot programs for Level 4 systems, but those too are not yet self-driving (nor are they available for purchase).

Trade school…LOL…sure

It’s a joke. I went to Harvard, back in the day. We used to (affectionately) refer to MIT as the trade school down the river.

Albaby

1 Like
35

Can you substantiate this assertion? The reason I ask is because self driving cars are made up of two systems, the vehicle itself and the AI hardware and software that replaces the driver. The AI hardware and software are not that hard to update. What is the insurmountable problem?

There are a few answers to this:

  1. My “gut feeling” based on 40+ years of relevant engineering experience. Including a year plus of experience in the car itself.
  2. Empirically, even the current version of software can’t run on the older hardware. The same is likely true of the new AI software that will come out in a few years (it won’t run on the current hardware).

A vehicle with upgraded hardware is not “currently on the road”. Once I see the new hardware, I may change my opinion. But right now, with the current vehicles, with the current MCU2, I don’t think any of those will ever be true full self driving.

I’m still not sure about sensors. Some argue that the current set of sensors (all visual) are not sufficient. But I am not sure if I buy that argument … after all, humans use those same sensors (visual) to drive.

2 Likes
36

Certainly there will be some adventurous people who would be willing to get into an unmanned private automobile with a complete stranger. I suspect that enough would not, sufficient to keep the idea of autonomous ride-sharing from ever becoming a thing.

I cannot, under any circumstances, envision anyone offering shared rides with autonomous vehicles, due to safety concerns. Shared transportation spaces can sometimes get dicey even when supervised by a bus driver,
>>>If there’s a CoOp that you personally know everyone else involved, you can solve that problem - but I can’t imagine that economically or environmentally, there’s enough gain from the occasional carpooling from such a small group compared to just letting the AV out in the world to ferry passengers for single rides.

Hi Albaby

I first saw your last reply this morning. Thanks for your input. I must say I’m somewhat surprised that you don’t see a future with AV “no-owner” transit. This, especially true, in that you were able to accept the concept of Uber in 1990 two decades before its reality.
1990 Uber? Cell phones? In 1990 cell phones were only available to the wealthy and usually installed in a car (but not always). They were not that reliable, and the cost of a Motorola cell phone in 1990 was around $4500.00, or $10,000 in today’s money. Not really available to the average Uber driver. But that’s the least of the problem imagining Uber in 1990:
Cell with data
Cell phones with internet
Cell phones with GPS
Heck, in 1990 the internet was run off modems, even a fast processing graphics computer took five to ten min to load on internet page. As far as I know there was not way for a individual contracting Uber driver to even get a PayPal type uber payment. But somehow you understood that all this was going to happen, that Google would map every inch of the country and cell phones would allow Uber drivers to get text messages and drive to a GPS location and take a faire to another GPS location, years before the Google boys were even in Grad School. But somehow “ownerless” AV transit escapes you.

Here are some thoughts on “ownerless” AV transit.
many players in the AV space seem to be proposing that the best way forward is individual adoption. In other words, AVs would incrementally replace conventional single-occupancy vehicles (SOVs), both in the garages of private owners and the fleets of ride hailing services. This vision seems far too limited in scope and creativity for such an innovative technology. Instead, the broader promise of AVs lies in their capacity to transform mass transit. replacing each single-occupancy taxi or ride-hail vehicle on the road with an AV wouldn’t do much to address urban sustainability goals, like decreasing road traffic, congestion, and reducing land used for parking, or rural access concerns, particularly around transportation costs. Not to mention that, in a single-use model, this cutting-edge, costly technology would be limited to users who can afford private rides. On the other hand, AVs in mass transit can serve a public good, and data drawn from a large rider base can help simultaneously mature AV tech and improve the overall health of public transit networks.
>>>Think of a fleet of shared, on-demand AV shuttles continuously moving people to and from traditional fixed route mass transit systems
https://ridewithvia.com/resources/articles/the-future-of-aut…

And, I don’t think that ownerless AV transportation would be limited to Via’s approach. There are a multitude of ways to create AV “ownerless” carpooling. Of course, the technology has a long long way to go.
The future of cars in thirty years will not look anything like the cars you have seen your whole life. And the future of cities is moving fast toward very high density living. Having a place to garage your vehicle will become too expensive. Even now, if you live in San Francisco, having a place to park you car every night runs about $2,500 a year. If you drive your car to work, it’s about $1,000 a month for parking.

There’s a company called “Easy Mile” that has already put the Idea of AV public transit to the test.
https://www.wired.com/story/driverless-mass-transit-hits-spe…

The ultimate benefits of AV technology to public transportation remain unknown; there’s only so much you can learn from a demonstration project. EasyMile, for example, has provided shuttles for pilots in 19 states; the shuttles can carry six people and travel up to 12 mph.

Driverless public transportation is not here yet, agreed. But in 1990 cell data and Google maps were years and years away and that didn’t seem to get in the way of your acceptance of the Uber Idea in 1990.

37

I must say I’m somewhat surprised that you don’t see a future with AV “no-owner” transit. This, especially true, in that you were able to accept the concept of Uber in 1990 two decades before its reality.

Ummmm…I’m not sure how you got that from my prior post, but I think you’re seriously misreading it.

What I said was that I didn’t think that companies would offer shared rides with autonomous vehicles, due to safety concerns, but that I wouldn’t have raised the same objection to Uber if someone had told me about it 30 years ago. That’s not because I had some prescience about science and technology 30 years ago. It’s because if someone had described the concept to me 30 years ago, it wouldn’t pose the same issue. With a shared Uber, you’re not alone in the car with another passenger. There’s a driver.

People are reticent to be alone with strangers, with no opportunity to get away quickly if the stranger turns out to be weird or dangerous. That is mitigated if there’s another figure there or nearby that would be a witness and can intervene if something bad happens - a bus driver, a subway conductor, or even a cab/Uber driver.

That’s a problem for shared autonomous vehicles. You don’t need a driver. So passengers are alone with other passengers, without supervision - or even someone who can pull the car over and make one of them get out. That’s the problem discussed in your Wired article. Passenger safety and harassment becomes a serious risk factor, one that will pose a huge obstacle for that model. It’s not a question of predicting (or accepting) technological advancements. It’s an issue posed by human behavior and psychology.

So, no. While I think Level 5 autonomy will certainly feature robotaxis galore, I don’t think there will be much interest in the AV equivalent of UberPool. Actually, the opposite. I think that Level 5 will put some serious pressure on the viability of bus public transit services, especially for coverage routes. Public busses will still need to have an employee on them for safety, which dramatically reduces the cost savings they can derive from autonomy.

Autonomy will make cars much more useful than they are now, and possibly more accessible to people who can’t afford them now. That’s an issue for people who are concerned about the problems large numbers of cars cause for the urban form. The idea of shared rides is a bit of whistling past the graveyard for them, I think. It’s the comforting fiction that maybe we won’t see people shifting out of communal transit modes (like buses and trains) into single-user private passenger cars, that the AV’s will still collect multiple travelling parties together. I don’t think that’s going to happen, because of safety concerns.

Albaby

2 Likes
38

Can you substantiate this assertion? The reason I ask is because self driving cars are made up of two systems, the vehicle itself and the AI hardware and software that replaces the driver. The AI hardware and software are not that hard to update. What is the insurmountable problem?

The problem, perhaps not always insurmountable, is data. It’s clear that Level 5 will require vast amounts more of data and ways to process all of it much more quickly in real time, possibly including communicating with nearby vehicles to ascertain what is happening with each other member of the local hive (that last part may be unnecessary, but it seems possible.)

It’s also clear that our current chipsets and logic boards aren’t going to be capable of that, not to mention the software.

Have you ever tried to load iOS14 on an iPhone 3? And good luck with that, because you can’t do it. While that’s not a perfect example (you throw away the iPhone 4) it’s illustrative of the issue. Not only can’t you store the reams of data needed for Level 5, you can’t process it. And as Level 5 becomes possible you will likely need newer, more sensitive, and perhaps more numerous sensors, cameras, LIDARs, and who knows what else?

Could you just throw out the motherboards and start over? Well, possibly, but where are the cables for the new inputs? How do we even know the old cables will carry the new signals efficiently? Answer: we don’t. Could they? Maybe. But I have a closet full of old cables and connectors that will never be used again, and I’ll bet you do too. (*Note to self: clean them out, really, they will never be used again.)

So if you’re not sure the hardware can be adapted, and you are absolutely sure the software needs to be improved, what are the chances that you can continually upgrade todays Tesla (for instance) into tomorrow’s Level 5 automobile? I’d guess slim.

That doesn’t mean Level 5 is unattainable, just that the improvements will be incremental, just like the improvements in Word, Excel, TurboTax and all the other softwares that walked hand in hand with new hardwares as well. Sometimes they were backward compatible, but not forever.

That’s why.

39

It’s a joke. I went to Harvard, back in the day. We used to (affectionately) refer to MIT as the trade school down the river.

Albaby

The theocratic history of Harvard has never really changed.

…(of course no one can change history. Even if they write another history)…

40

There is no car that is capable of driving itself commercially available in the U.S., even for short periods of time.

Nevertheless, driverless cars will be in use on city streets in a couple of months. As I noted in an earlier post, Uber Eats will use autonomous cars from Nuro to deliver food in parts of Houston and the Silicon Valley this fall. No one will be in the car and they are limited to a top speed of 25mph. https://www.engadget.com/uber-eats-nuro-autonomous-food-deli…

Apparently existing technology can self-drive safely at lower speeds.

Seems inevitable that once food deliveries are demonstrated to be safe, cars will be introduced using the same technology to carry passengers within those districts. This system will then be adopted by crowded downtowns, residential suburbs, and retirement villages where lower speed travel is the norm.

I think this will particularly impact families, who as their kids grow up typically end up having to own 3-4 cars. Don’t need those extra vehicles if the kids can simply Uber a robo-taxi to take them to dance class or soccer practice. The business model will probably be similar to Zipcar subscriptions that are popular on college campuses. Subscriptions allow students to use a Zipcar when needed for an hourly or daily fee, with Zipcar lots distributed across the campus. Self-driving improves this model by allowing the car to come to the customer. My daughter in college uses Zipcar and it is much cheaper than owning a vehicle. Self-driving zipcars will be even better since you wouldn’t have to park.

I think autonomous driving is just about to take off