Converting stupid streets to high capacity cooperating communicating intelligent streets, requiring cooperating communicsting vehicles for entry, would be utterly transformative.
Probably impossible until my non-cyber generation is out of power.
d fb
The only insight, the governor is trying to avoid certain looks instead of taking the bull by the horns. He may have little choice.
The mountain tunnel is one more of those engineering overreaches. I have a feeling behind it all the MBAs stare into mirrors loving themselves.
There’s plenty of reasons to think that won’t be the case. As DrBob pointed out, preliminary analysis of conventional TaaS ride-share (like Uber) is that it increases congestion. TaaS vehicles increase total miles travelled because of significant deadheading (personal cars that park at their owner’s destination don’t spend any more time on the road looking for their next fare). They also pull trips from other modes, like transit and biking/walking. IOW, TaaS leads to more car-miles - not fewer.
AV TaaS will also have deadheading and compete with other modes. Advocates also claim it will be significantly cheaper than conventional cars, in addition to being more enjoyable (because you can do other things) - and generally, when something gets both cheaper and better, people will consume more of it. More usage = more congestion, ceteris paribus.
Far more important than getting autos automated is making them other than stupid, deaf, and mute with each other — making them cooperative. Robo-taxis provide two great goods,
making certain that expensive mechanisms that are cars are put to intensive use, and
mostly do not steal valuable public space being parked curbside.
But that is nowhere near enough to make them a useful goal. Mutually cooperating communicating vehicles would transform how traffic flows. For beginners, they mostly end the need for traffic signals. Far better, they allow traffic to safely travel at far greater speeds while lowering the accident rates. Even further, they allow system wide knowledge to recognize the probability of a clog somewhere and to signal to traffic to avoid that intersection.
d fb
Neither of those things is much of a great good.
There’s really not much of a benefit to putting cars to more “intensive use.” Sure, cars go unused most of the day. So does my oven. So does my shower. So does my sprinkler system, my pool, my television, every article of clothing, etc. Cars get used up primarily as they are used, not over time. There are some time value of money savings from the accelerated usage, but they are modest.
As for public space - remember, AV’s don’t disappear after they drop off a passenger. They’re still going to take up space. Some of that space is on public roads (while deadheading) - and a lot of AV’s will end up parking anyway. And some (if not all) of that curbside space will probably still need to be there and dedicated to vehicles, because AV’s still need to leave the active road lane to drop off and pick up passengers, else they become impediments to the flow of traffic.
The problem with the Uber business model is that because the driver bears the cost of vehicle maintenance there is little incentive for Uber to reduce the number of vehicles on the road.
IMO, the future will be with dedicated robotaxi companies that will use AI to optimize the number and distribution of their vehicles. The AI will learn with experience, creating an increasingly more efficient transport system. I can see cities exclusively using a robotaxi service in its downtown areas thereby replacing much of its public transportation. Perhaps even taxing robotaxis to make its remaining public transport free to use. I can easily see a Boston where the major subway stations are the focal points for a robotaxi network and where personal vehicle use is restricted.
Replace “cars” with vacation homes and then look up VRBO or AirBNB.
I’m not sure that’s correct. One of the features of the Uber business model is that it scales up and down in volume as demand waxes and wanes. Since Uber doesn’t own the cars, and since many of the cars are the private vehicles of drivers engaging in a “side hustle,” the cars aren’t always searching for a fare. Uber doesn’t lose out if one of its drivers looks at a period of low demand and just decides to go home and do something else during the late morning or early afternoon. So its set up to change the size of the fleet over the course of a day, so there aren’t too many cars driving around waiting for a fare.
Uber’s already got an AI algorithm - with tons of data about matching vehicles to passengers and how to use surge pricing to optimize their fleet size to match irregular demand across time and geography.
A dedicated robotaxi fleet can’t do that. They have the number of cars they have. It’s a disadvantage. Tesla’s original concept of a Tesla Network would have mirrored the Uber model of bringing vehicles in and out of service during peak times, without Tesla having to bear the cost of vehicles not in service. That’s probably why they’re trying to keep that in place, even as they are building dedicated Robotaxi vehicles - their optimal fleet is one that’s just enough Robotaxis to handle the “baseload” level of transportation demand, with Tesla Network privately-owned cars activated during peak demand times.
Except buildings aren’t like cars. They don’t depreciate the same way cars do. Cars depreciate very fast with mileage and use, and a big chunk of their cost is the operating cost. If the useful life of your car is 300K miles (for example), you don’t lose out much if you drive it 15K miles per year for 20 years or 75K per year for 4 years. It’s basically the same cost - the same fuel, the same maintenance, the same tires, the same depreciation. Real estate doesn’t work the same way.
What makes you think Uber changes its fleet size over the course of the day? The number of Uber cars on the road at any one time may change, but that is at the whim of all those Uber drivers, not the company.
Exactly! And a good AI can optimize this so that it ends up significantly reducing congestion. Now throw in some Boring tunnels for high demand routes and to bypass congestion points and you might have a revolutionary transport system.
You are still living in the gas powered 20th century when ICEs were a mature technology that didn’t change much.
Have you seen how quickly BEVs are evolving? Just the batteries alone is pretty astounding. I’ve said before, BEVs are becoming increasingly like computers with one way being how quickly they lose value with time. How much can you sell a five year old computer for? How about a 10 year old hardly every used cell phone? BEVs that can’t be easily updated through software are going to depreciate enormously within 5 years regardless of usage. May as well get the most use out of it early as in five years no one is going to want to ride in something so antiquated.
It’s not at a whim. Uber changes the prices - and the amount of money the drivers are paid - in order to get more drivers to log in as available during times of high demand, and to reduce them when demand is low. Surge pricing and all that. They’ve got it down to a science - the prices are all set algorithmically so that at different times (and different parts of the service area) they match the cars-in-service volume to passenger volume.
They’ve already developed their AI algorithm - using the enormous amount of data they have on passenger demand - to optimize matching riders and vehicles. They’ve got about a decade’s worth of experience doing that (one reason why some analysts think that a partnership with Uber, rather than a rivalry, is likely for Tesla).
If that ends up being true (and I don’t think it will), then all BEV’s will be easily updated through software so that there isn’t a segment that depreciates so much faster than the rest of the market.
The same analysis still holds. Unlike residential real estate, much of the cost of TaaS is the operating cost (fuel, tires, maintenance) and depreciation. For the most part, you don’t save all that much money by burning through miles more quickly than slowly. Unlike a vacation home, which is going to cost you pretty much the same amount whether you’re using it or not, if you start driving your car 200 miles per day as part of a robotaxi service you’re going to “use up” the car’s useful life very very fast. Which is why comparisons to VRBO/AirBnB don’t work - having someone use your vacation home while you’re not in it has a relatively low marginal cost, but having people driving your car all day while you’re not driving it has a very high marginal cost.
If no one else mentions I will, a big part of uber’s market power is that many of the drivers are ignorantly “spending down” the residual value of their own vehicles faster than they are making money….
d fb
Maybe not directly, but since Uber controls what the app shows, to the customers, and to the drivers, they effectively change the fleet size and location throughout the day. For example, if they show drivers that an airport ride into the nearby metro area is about $60, they will attract a lot of drivers to the airport (presumably they set that higher price because they know there are a lot of customers requesting rides at that time), but when they set the price to $30, they will attract a lot fewer drivers to that area.
Similarly, if local rides are priced under $10, they will attract a lot fewer drivers than when they set them to $15+. I know at least one uberlyft driver that told me that when the local prices drop below $10, he goes home and does something else. And this guy has an electric car, so when the prices go low, usually midday, he goes home and plugs the car in to charge for a while. Then, when the prices go up again, usually in the evening, he goes out again and drives for a few hours. He also says that he hates airport rides.
Of course, for a privately owned car, over the air update is a big plus since one basically doesn’t need to do anything to stay current. For a fleet car, which probably needs to go in for cleaning and such, even if it is just cheap parking in off hours, then more manual updates wouldn’t be that big a pain.
Key point not mentioned: IF there are enough willing drivers with vehicles with Uber/Lyft. Not an issue with TaaS.
Higher use of vehicles is not an issue. The vehicles used to earn money are able to be written off against the revenue earned. More use = higher tax deduction that year.
But with Taas there has to be enough available vehicles to meet peak demand, or close to it. The rest of the time those vehicles are idle.
YES! + at least 16 more characters
The system scales to meet demand. Because there ARE events needing a large number of vehicles (sports, etc), TaaS already knows of them and has the capacity to handle it. TaaS will use the same vehicles to make multiple trips because it can plan vehicle movement very efficiently. Uber/Lyft lacks the efficacy of TaaS when doing large jobs. Drivers CAN choose to drop out (for whatever reason). Not so with TaaS.
Slow periods in a TaaS system allows maintenance (including cleaning interior, etc) and charging for many vehicles. There is no requirement for all charging to happen at centralized hubs. Charging can easily be a large number of chargers all over the network. All vehicles needing a boost in charge do not have to go far to get recharged–they can be recharged while waiting on the street for the next ride close to the vehicle. So it is not really wasted time. And far less “dead-heading”.
Uber certainly sets prices based on demand, but I don’t think supply (the number of drivers) is considered at all for deciding the price point. In other words, Uber maximizes revenues by charging the highest price customers are willing to pay at a given time regardless of the supply of drivers. It is purely a demand driven price point driven by algorithms that are based solely on fluctuations in demand.
For dynamic pricing to impact driver supply, the change in price has to be maintained long enough for drivers on the sidelines to react. It generally isn’t. If Uber were serious about using surge pricing to increase supply, it would design the algorithm to take into account real time changes in the supply of drivers. That the algorithm changes too quickly to consistently impact the number of drivers strongly suggests that Uber only cares about demand when setting price.
…The data I collected suggest that surge pricing doesn’t seem to bring more drivers out on the roads, but rather pushes drivers already on the job toward neighborhoods with more demand–and higher surge pricing… https://www.washingtonpost.com/news/wonk/wp/2015/04/17/how-uber-surge-pricing-really-works/
In other words, surge pricing is Uber playing Whack-a-mole.
What you are missing is that for BEVs, the depreciation due to usage is minor compared to the depreciation due to age. That is because of rapid advances in BEV technology that improve the performance while decreasing the price. A $60K 2024 BEV will depreciate by 50% by 2027 even if never driven simply because by 2027, new BEVs with the same or better performance will cost $35K. It is the economics of computers and cell phones.
With homes, the depreciation due to usage is insignificant compared to the depreciation due to aging. The same is becoming true with BEVs since maintenance costs are minimized and the product is in the new technology phase when significant improvements occur quickly and frequently.
It generally is. Uber pushes the surge price notifications to drivers, not just passengers. It induces drivers both to enter the pool as available generally, and to relocate to the specific geographic area that demand is surging. Your article from 2015 (nearly a decade ago!) notwithstanding.
It’s fairly complex. Uber’s been doing this for almost a decade now, and they themselves invested in AI systems and some pretty serious refinements to their algorithms. There’s a reason they’ve been able to have such a big moat around their business, and it’s partially because there’s no one that has more experience - and more data at matching passengers to vehicles, and managing both pools.
I mean, if you think that’s what’s going to happen, then sure. I completely disagree. I mean, we haven’t seen anything like that EV’s so far. A base Model Y RWD retailed for $46K in 2020, and retails today at $44K; the Long Range version went from $51K to $49K. Prices spiked in 2022 with the supply chain constriction, so you had a sharp rise and fall (which killed Hertz). Low mileage 2020 Model Y’s haven’t fallen 50% in value:
You have long predicted that there will be a “step change” sharp reduction in BEV prices with some revolutionary cut in battery production costs. And that may happen, some day - but it’s not likely to happen over and over and over again. After one or two major reductions in battery costs, then much of the cost/value of the car will be in…the car. The frame and motors and interiors and tires and everything else - which carry a lot of value forward in time, particularly if not driven very much. Combine that with the OTA updates you project will be ubiquitous in the industry soon, and I really don’t see the type of depreciation you describe being realistic.
Of course they do!!! Your sentence wouldn’t make any sense otherwise. If they set prices “based on demand”, WHAT EXACTLY DO THEY COMPARE DEMAND TO to set the various prices? Yep, it’s supply!!! If there are 100 people requesting a ride (“demand”) in area A, and there are only 10 drivers (“supply”) in that area, or heading to that area, then they put in “surge pricing” and suddenly the price rises, and suddenly more drivers are induced to head to that area. But if there are 100 people (“demand”) requesting a ride, and 90 drivers (“supply”) in the area, then there is no surge declared, and no sudden price increase in that area. And since their system is pretty smart at this point, they also take into account historical supply and demand in the various areas. Like their system has a parameter of “what happens when the game ends ad everyone heads out?”, or “rush hour”, etc.