Autonomous Car Progress

[FloridaJohn]

I am very excited about the entrepreneurial aspect of people/ex Uber and Lyft drivers now owning Robotaxi making almost passive income. A slight change in their life where they now don’t need to drive to make money.

In fact, imagine they could double their income working on their day job while your Robotaxi is making money for them self driving all day.

Take it even further, I can share out my Optimus to roboshare requests. People paying to use my Optimus for their task work. Now I have a day job and two passive incomes.

Wow

If/when the barriers to entry get that low, the market will be flooded with Tesla robotaxis and the competition will drive the price down so low that it will no longer be profitable.

People will use it to get a "free car" by having the fares pay for it, but will end up with a worn-out abused car by the time it is paid for. This will drive down the price to a "break even" point.

People who have a change in their financial situation will use it to subsidize their car payments. Since they don't care to break even, this will drive down the price even further.

The rich people, on the other hand, will buy their own robotaxis to use as a personal chauffeur to drive them to work and to pick up the kids after school. This will lower the demand for robotaxis even further.

It may even get to the point where people can charge a premium to customers if they drive their own cars (traditional taxi or Uber, etc.) because then they can offer personal services like information about the city, or putting and taking out luggage from the trunk, etc. Of course, then we end up right where we started, but with a lot less jobs for people.

 

[diplomat33]

Cruise CEO on Bloomberg Tech to talk about Cruise's expansion plans.

 

[diplomat33]

I think he makes a good point about perceived safety vs engineering safety.

https://twitter.com/x/status/1690130997252124673

 

[ThomasD]

How many AVs would be needed to displace Uber and Lyft in the San Fransisco Metro area?

 

[JB47394]

How many AVs would be needed to displace Uber and Lyft in the San Fransisco Metro area?

I found a report stating that in 2017 there were 6,000 ride hailing vehicles providing 170,000 trips per day. That's 28 trips per driver per day. Busy. Between 2017 and 2022, Uber's global trip count has doubled. For chuckles, if we say that San Francisco also doubled in that time, then we keep twice as many drivers at that busy level. That's 12,000 vehicles.

 

[stopcrazypp]

How many AVs would be needed to displace Uber and Lyft in the San Fransisco Metro area?

This article claims there are 50k Uber drivers in SF.

https://www.sfgate.com/tech/article/Uber-Lyft-San-Francisco-pros-cons-ride-hailing-13841277.php

This report from 2017 based on 2016 data estimates there were 45k drivers total in SF and on an average weekday, more than 5,700 ride-hail vehicles operate on San Francisco streets during the peak period. On Fridays, over 6,500 ride-hail vehicles are on the street at the peak.

TNCs Today (2017)

This report provides the first comprehensive estimates of Uber and Lyft activity in the city.

www.sfcta.org

 

[rbt123]

I understand Waymos model , I think. I don't understand how cruise gets to profit. Waymo gets a rider and can make profits off of services tied to google while they have the rider. Brilliant, I get that. It lets them leverage a near monopoly and thus above normal profits. How does cruise get there

Cruise doesn't need a profit: GM does. GM and Honda will show a solid ROI on Cruise Origin vehicles within days of them rolling out of the factory. Honda will build out Japan. If the service fails financially, Cruise goes bankrupt and the GM/Honda keeps the pieces, the software and hardware parents, as the primary creditors.

IMO, the 20% of non-GM and non-Honda Cruise shareholders (Microsoft) are in a rough place because all positive cash-flows will be going toward GM vehicle leases for a very long time, until their fleet has millions of vehicles and cannot expand service anymore.

I'm not sure why Microsoft is a shareholder. Perhaps they intend to put something like an X-Box in the vehicles with a bunch of resumable games? Tie your home X-Box account to the Cruise account and the most recent game you were playing could be continued during the trip. So basically a marketing play?

 

[diplomat33]

Popular physicist and futurist Michio Kaku was on CNBC to discuss the recent news of CPUC approving Waymo and Cruise for commercial driverless robotaxis. He argues that this is just the beginning and more cities will follow.

Famed Futurist Michio Kaku: Cities will start to follow San Francisco's lead with self-driving taxis

Michio Kaku, famed futurist, joins 'Last Call' to talk the future of self-driving cars after San Francisco approves robotaxi expansion.

www.cnbc.com

 

[Doggydogworld]

To deploy a driverless fleet, Tesla will need to get the same permits that Waymo and Cruise just got. It was hard for Waymo and Cruise to get those permits.

It's hard to get permits for anything in CA. Neither company had any issue getting permits in AZ, TX, etc. Tesla could get permits in those states tomorrow if they wanted. They don't want. Robos are a cash furnace vs. the insanely profitable business of selling FSD hopes and dreams at $15k a pop.

How many AVs would be needed to displace Uber and Lyft in the San Fransisco Metro area?

The study @JB47394 linked says 6k ride hail cars operated in the city during peak hours. The total number of Uber/Lyft drivers is higher, because not all drive at the same time. But a 6k car Robo fleet could cover the peak and have surplus cars sitting around (or charging up, doing deliveries, etc.) the rest of the time.

I found a report stating that in 2017 there were 6,000 ride hailing vehicles providing 170,000 trips per day. That's 28 trips per driver per day. Busy.

It'd be 28 rides per day if every driver worked 20+ hours per day. They don't, of course. But a Robo can.
A lot of those 170k rides are into or out of the city itself. So it's not an exact match to a Robo fleet that's (currently) geofenced to city limits.
I don't think Uber/Lyft grew 100% in SF since 2016-17. Maybe 25-50%. Most of their expansion came by adding and expanding new cities.

This article claims there are 50k Uber drivers in SF.

I suspect that's SF metro area, not just the city. Most metro area drivers go into and out of the city at some point, of course, but only a fraction are in the city at any given time (e.g. the 6k during peak hours).

 

[diplomat33]

Multiple Cruise AVs stalled. Cruise says that a local festival caused slow wireless bandwidth which resulted in the vehicles having trouble connecting to remote assistance.

https://twitter.com/x/status/1690423649134854145

This dependence on a wireless signal for remote assistance is a major weakness IMO. I feel like Cruise should have thought of this issue before now. They knew remote assistance needs a good wireless connection. So surely, they should have thought about what would happen if the cars were not able to connect to remote assistance. The fact that they are just now saying that they are working on making sure this does not happen again feels like poor planning on Cruise's part, especially since they claim to want to scale so big. How are they going to scale to 1000s of robotaxis if the cars are dependent on a wireless signal for remote assistance?

 

[stopcrazypp]

Multiple Cruise AVs stalled. Cruise says that a local festival caused slow wireless bandwidth which resulted in the vehicles having trouble connecting to remote assistance.

https://twitter.com/x/status/1690423649134854145

This dependence on a wireless signal for remote assistance is a major weakness IMO. I feel like Cruise should have thought of this issue before now. They knew remote assistance needs a good wireless connection. So surely, they should have thought about what would happen if the cars were not able to connect to remote assistance. The fact that they are just now saying that they are working on making sure this does not happen again feels like poor planning on Cruise's part, especially since they claim to want to scale so big. How are they going to scale to 1000s of robotaxis if the cars are dependent on a wireless signal for remote assistance?

Reminds me of this from more than a year ago. So they still have not come up with a contingency plan for network outages.
Autonomous Car Progress

 

[diplomat33]

Reminds me of this from more than a year ago. So they still have not come up with a contingency plan for network outages.
Autonomous Car Progress

Apparently not which is why it is particularly disappointing. And their PR response of "we are working to make sure this never happens again" rings hollow to me since they've known of this serious issue for a year now and did not fix it. It indicates to me that their validation process is not strong enough since they let major issues continue for long periods of time before addressing them.

 

[cwerdna]

I swear I listened to this before Aug 11th: https://www.npr.org/2023/08/11/1193...nts-fret-about-push-for-more-self-driving-car. Title is "San Francisco's police, fire departments fret about push for more self-driving cars".

Many many people that are disabled and unable to drive that will find robotaxi helpful. Now that we have Uber in cities it is not so critical but in the USA our taxi network- outside of NYC was terrible.

Elderly, blind, alcholics and other substance abusers, people with severe neuroligical or muscle diseases (MS, early Lou Gehrigs, Parkinsons, and on and on). A huge number of people simply can't drive and are stuck at home. Often poverty goes hand in hand and Ubers in cities can be high $. My great hope for Tesla is that they can get to a ubiquitous RT solution that would enable Unable to Drive (UtD- we need a good acronym) to have a higher quality life. I even think it would be a positive for the GDP as they could participate more fully in society. This is why I care about autonomous driving.

One last point, most elderly begin to have vision issues while they are still quite mobile. Driving at dawn, dusk, and at night is very difficult and thus they are forced to artificially restrict activities.

Yep. My dad is getting pretty old and having vision problems. Driving at night has pretty much been a no go for awhile. He has some issue now where (from what I understand) where he can't see out one eye, something like that. My mom's vision isn't great but not as bad. Public transportation in my city (where they also live) is awful. If they need to go out shopping for even the essentials/basics, they need to drive. Taking the bus is inconvenient and slow.

I have an aunt who's never had a driver's license (AFAIK) but semi-recently became blind due to use of some eye drops that later got recalled.

 

[diplomat33]

Some more info on the Cruise mass stall event:

At about 11 p.m. Friday, as many as 10 Cruise driverless taxis blocked two narrow streets in the center of the city’s lively North Beach bar and restaurant district. All traffic came to a standstill up Vallejo Street and around two corners on Grant. Human-driven cars sat stuck behind and in between the robotaxis, which might as well have been boulders: no one knew how to move them.

The cars sat motionless with parking lights flashing for 15 minutes, then woke up and moved on, witnesses said.

Cruise blamed cell phone carriers for the problem. At 11:01 p.m. Friday, Peskin sent a text message to Cruise government affairs manager Lauren Wilson. At 8:25 a.m. Saturday, she texted back: “As I understand it, outside lands impacted LTE cell connectivity and ability for RA advisors to route cars.” Outside Lands is a three-day music festival held in Golden Gate Park, four miles from North Beach.

And Peskin says they will pursue "every means", including a court injunction to reverse the CPUC permit decision:

Peskin said city officials are pursuing “every means” to have the CPUC decision reversed, and are discussing whether to seek a court injunction. Another option: fining Cruise and Waymo thousands of dollars for each robotaxi road blockage.

San Francisco's North Beach streets clogged as long line of Cruise robotaxis come to a standstill

Just one day after state officials approved massive robotaxi expansion in San Francisco, a long line of the driverless cars come to a standstill and clog traffic in North Beach neighborhood.

www.latimes.com

I do hope that they separate Cruise and Waymo. I don't think it would be fair to also reverse Waymo's permit because of Cruise's incidents.

 

[stopcrazypp]

Some more info on the Cruise mass stall event:



And Peskin says they will pursue "every means", including a court injunction to reverse the CPUC permit decision:

San Francisco's North Beach streets clogged as long line of Cruise robotaxis come to a standstill

Just one day after state officials approved massive robotaxi expansion in San Francisco, a long line of the driverless cars come to a standstill and clog traffic in North Beach neighborhood.

www.latimes.com

I do hope that they separate Cruise and Waymo. I don't think it would be fair to also reverse Waymo's permit because of Cruise's incidents.

If it's a fine, it should apply equally to both based on actual incidents. If Waymo has less incidents then they would naturally be fined less.

While it's a bit unfair to Waymo (they originally had more restrictions because they came later, Cruise was already able to do paid driverless rides at night), I suspect if they pursue the legal route, it'll be an injunction to hold both to the same status as before the 8/10 meeting.
CPUC Approves Permits for Cruise and Waymo To Charge Fares for Passenger Service in San Francisco

It's rather unfortunate for Cruise to have such a major snafu immediately after their approval, which to the public screams that it was not justified. The logic CPUC used however is the same as in previous approvals, which is their criteria is based on what is established years ago, and while they acknowledge the concerns of halting traffic, that is up to a different process to evaluate (TNC proceeding, R.12-12-011).

 

[nativewolf]

I swear I listened to this before Aug 11th: https://www.npr.org/2023/08/11/1193...nts-fret-about-push-for-more-self-driving-car. Title is "San Francisco's police, fire departments fret about push for more self-driving cars".

Yep. My dad is getting pretty old and having vision problems. Driving at night has pretty much been a no go for awhile. He has some issue now where (from what I understand) where he can't see out one eye, something like that. My mom's vision isn't great but not as bad. Public transportation in my city (where they also live) is awful. If they need to go out shopping for even the essentials/basics, they need to drive. Taking the bus is inconvenient and slow.

I have an aunt who's never had a driver's license (AFAIK) but semi-recently became blind due to use of some eye drops that later got recalled.

Sorry to hear but yes, it happens...to almost everyone. I think RT will be great for elderly.

 

[KArnold]

WARNING - THIS IS SIMPLY A SWAG AND THEORETICAL.

There must be more to this story.

In one of the videos there is a car stopped with a very clear view forward. Did someone stand in front of the car until it stopped, put its flashers on, and phoned home? Then the video starts? And perhaps a "stuck" vehicle does not currently autonomously retry even if the way forward becomes clear before support intervenes? Perhaps. If true, this isn't really Cruise' fault but they do need to learn how to mitigate it the next time an activist tries it.

I'll bet that policy about retries gets revisited too.

I'm also assuming the AV does not need Internet connectivity for normal autonomous operations? I think that's a safe bet.

 

[stopcrazypp]

WARNING - THIS IS SIMPLY A SWAG AND THEORETICAL.

There must be more to this story.

In one of the videos there is a car stopped with a very clear view forward. Did someone stand in front of the car until it stopped, put its flashers on, and phoned home? Then the video starts? And perhaps a "stuck" vehicle does not currently autonomously retry even if the way forward becomes clear before support intervenes? Perhaps. If true, this isn't really Cruise' fault but they do need to learn how to mitigate it the next time an activist tries it.

I'll bet that policy about retries gets revisited too.

Cruise already came out and said it was because of congested mobile networks and similar things have happened previously last year. From other incidents, whenever things can be spun, Cruise tends to put the most positive spin as possible, so if there was deliberate sabotage, Cruise would have said that.

I'm also assuming the AV does not need Internet connectivity for normal autonomous operations? I think that's a safe bet.

Nope, they actually do need internet connectivity, look at the previous discussion linked upthread. Cruise vehicles regularly phone home and if there is no response, it halts. Waymo doesn't appear to have the same problems (although it's possible Google has a more robust network with backups, especially given they have their own Google Fi network). To be clear, it won't just immediately lose all control, it'll come to a safe stop, but it won't continue driving.

 

[cwerdna]

The incident also made local news: Cruise cars cause traffic jam in San Francisco's North Beach neighborhood.

 

[diplomat33]

So where are AVs in this cycle?

I feel like we are in the trough of disillusionment right now. There was an expectation that robotaxis would scale faster than they have and now with the recent incidents in SF, there seems to be growing mistrust. I see people say that AVs are clearly not ready yet for deployment because of the stalls and incidents.

IMO, the slope of enlightenment is likely to come when the tech further improves in reliability (much less stalls) but also when society better understands the advantages and limitations of AVs. I believe we will see faster scaling when we enter this stage.

The plateau of productivity will be when the tech is super reliable and scalable. In this stage, we will have lots of L4 trucks delivering goods across the country, robotaxis in almost every major city and L4 consumer cars are common from multiple carmakers.