Self-Driving Cars

Counts as doomsaying, I'm guessing?

Maybe. Or it could become another entrenched interest with political support, and in the entire interim it's a net negative for everyone outside of those lending their cars, since it's an increase in prices while being enough of a source of extra vehicles that it would stall efficiency (although efficiency in an inherently inefficient approach to 'people getting places' is, I guess, moot after a certain point) in growing to capacity. Plus whatever extra regulatory headaches it can contribute to everything. Either way, 'people could buy cars just to lease them to companies to rent them to other people' is an unusual downside to autonomous cars.

Ah, okay. Didn't realise it'd fall under that, either.

Edited by RainehDaze on Apr 25th 2024 at 6:24:53 PM

The cap on rent extraction is when the price of ridesharing becomes equivalent to the price of direct ownership. Prices higher than that would not be tolerated by the markets. The absolute worst case is that consumers break even.

While it might be possible that small fleet operators could attempt some kind of cartel price-fixing, that is not a reason to stop autonomy; it's a reason to prevent cartel price-fixing. I feel like this should be obvious.

Anyway, if we use Tesla as an example, its declared plans are to reserve a portion of its robo-taxis for its own fleet operation, which will price itself to compete with both auto rental and human-driven taxi/ridesharing services. Owners will be able to second their cars to said fleet for a share of the income.

Since Tesla will establish the pricing for this model (and likely dominate it for some time to come, until other manufacturers catch up), the rent extraction will be bounded by these factors. It can't charge below cost, for obvious reasons, and it can't charge more than competing services, ditto.

Or, if we're determined to believe that Tesla's autonomy efforts will fail, [insert company here] will do it instead.

ETA: "What if bad actors?" is a really obnoxious 'gotcha' question because it can be used against anything and is totally unconstructive.

Edited by Fighteer on Apr 25th 2024 at 2:15:54 PM

I still feel like the most likely scenario for full-autonomy (outside of car companies going full monopoly and all refusing to sell cars instead of rent them) is a dramatic reduction in household/family vehicle numbers. Who needs a two-car household if the car can take itself from one household member to another? Even a geographically dispersed family could share an occasional use vehicle (like a people carrier, small van or camper van) on the basis that you can make it take itself to another family member.

This is especially true if we spend any substantive amount of time with fully autonomous vehicles that are condition limited. Geofenced vehicles could be fully autonomous in a one geographic area but need a driver in another. The other big geographic limiter would be regulatory differences. If a car gets certified to drive itself in Arizona, Florida and California then that’s great, but nobody in any of thosue states is going to give up their car fully if the self-driving rental is limited to the state borders, they’ll just use it autonomously in that state and drive it manually when out of state.

I suspect when it comes to people renting out their vehicle to a robotaxi service, people are going to quickly realize that the only way to make enough money from it to be economically viable is to treat it like a part time job.

If I'm renting out my vehicle as a robotaxi (whether independently or more likely via a network), the car is going to need to be charged and repaired more frequently, cleaned much more often, and it will probably have to be replaced entirely sooner.

In order to maximize my profit, I would need to control these costs as much as possible. Can't do anything about increased charging costs or sooner replacement, but basic repairs and cleaning is something I could do. And both of those take time doing work that maybe I don't want to do (ie cleaning out a drunk person's vomit), which makes it a job. And maybe that's fine, but it does mean that this isn't really free money.

Sure these tasks could be outsourced in principle, but I suspect that the costs of outsourcing + increased charging costs + cost of sooner replacement are likely to exceed the expected income from renting out the vehicle in the first place.

There's nothing new about "private" rental/car-sharing apps; see Turo as an example. That's not what we're talking about here, though.

Let's say you own a Tesla and want to get in on that sweet robo-taxi cheese. You would sign a bunch of waivers, then you would second your car to Tesla's fleet during hours when you aren't using it yourself. At that point it becomes, effectively, a Tesla robo-taxi. The company takes care of routing, dispatching, charging, cleaning, maintenance, etc., while it is in their custody. Once it comes back to you, it's either as good as it was before or they promise to fix it.

So all of those fleet costs would be borne by the fleet operator, who would then share the revenue with the car's owner.

Again, insert [other company here] if the name Tesla causes you anxiety.

Edit: Somehow I lost a bunch of stuff I added. Here goes again.

We're probably going to go through a period where people abuse robo-taxis by vandalizing them, interfering with their operations, vomiting in them, etc., and operators will have to bear those costs until we get it out of our systems. People be assholes, film at eleven.

Of course, since you'll need a payment account to use the services, most people will probably learn their lesson when they get a $5,000 bill.

Edited by Fighteer on Apr 26th 2024 at 10:14:56 AM

Effectively you’re talking about a combined taxi-company/automaker borrowing sold cars back from consumers to use as their taxi fleet.

While self-driving could help with that business model I don’t think it solves the fundamental reasons we’ve not seen that business model emerge already, which are that it requires a huge amount of trust in a taxi-company/automaker for people to loan their property to said company for what in the end won’t be a ton of money, with the other main reason being that it’s probably not a viable business model for the taxi-company/automaker.

If Tesla is covering all the costs (including wear and tear) on such cars, paying extra money to utilise cars which are out of position and giving the owner money where is the profit for Tesla?

That’s without getting into how regulations in many countries means that a car itself needs to be licensed to operate as a taxi, self-driving removes the licensing for the driver but not for the car itself, who pays for that?

I suspect the idea of loaning your vehicle as a robotaxi will eventually work in some capacity, but Tesla's plans feel more than a little "too good to be true" on the part of the owners.

Optimistically this just means the payouts might be lower than anticipated. Pessimistically Tesla will try to sneak as much of the burden back to the owners as possible (ie claiming damages were caused when the owner was in control when they weren't, and then demanding an unreasonable standard of proof).

The distinction of "robo-taxi" is much simpler than some people here seem to be making it out to be. There is no driver. All of the other aspects of vehicle sharing apply. You need hailing, dispatching, routing, payment, insurance, maintenance, fueling/charging, cleaning, garaging, etc. Lawyers to sue the bad apples, accountants to balance the books, CEOs to get golden parachutes, and so on.

Nothing about this is novel except the autonomy, and taxi/rental services have the economic model already figured out. Cutting out the driver removes one of the major costs, and so, all other things being equal, robo-taxi service should be cheaper than taking a human-driven taxi or rideshare.

There's no need to overcomplicate it. The complication is in the autonomy, and that's what companies are working to solve right now. Human factors like abuse and vandalism figure into that, of course, but we're acting like this is some kind of novel objection that's surely going to cause the business model to fail because nobody's thought of it.

Yes, yes, corporations are evil, fuck the corporations. I get it. But we already live in that world, so we might as well learn to deal. It's only going to get more pervasive. I'm sure there are other places where The Revolution can be planned.

Edited by Fighteer on Apr 28th 2024 at 4:51:56 AM

Agreed, which is exactly what this whole “loan you car back to a combined automaker/taxi-company so they can use it as part of a taxi fleet in exchange for payments to you alongside doing maintenance of the vehicle while it’s operating as part of the fleet” thing is, a massive over-complication.

Tesla or whatever company just need to sell or run robo-taxis, it’ll be huge and help a lot of people.

Edited by Silasw on Apr 28th 2024 at 10:50:52 AM

I feel like the problem with this is inherent in the description you’ve given - a market dominated by a few operators is no longer a competetive market, and those companies can then drive up prices (and their profits) to arbitrarily high levels.

You say that the top limit for prices will be the point where owning your own car is preferable - but car sales to individuals also operate on economies of scale - in the future you’re envisioning there will be far fewer companies, and far fewer car lots, selling to individuals. And there will be less competition in that market as well.

It feels like you’re describing a world where the car/ride industry is similar to the airline industry - and the prices and customer service in the airline industry are horrible. (Ditto in Britain’s privatized rail system.) That doesn’t strike me as desirable, on the whole.

The existence of robo-taxis would be very handy for me personally (I don’t have a car), but the consequences of moving many things (movies, music, software) from a purchase model where you pay for something once to a subscription model where you pay monthly or yearly have been sufficiently dubious that we should be wary of the idea that moving to that model for cars would be a positive good.

Automated public transit, on the other hand, could be far more flexible than cirrent public transit, with lower expenses, so you could increase the number of buses and thus the variety of routes. That would be great. More transit options for people who drive rarely would be great. Getting down from two cars per family to one would be great.

Edited by Galadriel on Apr 28th 2024 at 3:24:08 AM

I don’t see how the economic situations post-transition between companies will be much different than it is already, and, at least in the U.K. we don’t have a problem with national monopolies in the taxi industry.

God, imagine if taxi services had the sort of incoherent pricing that the British rail service does based on who knows what inscrutable logic.

Journey time for a concert I went to a few weeks back: about 35 minutes, just under 20 miles. Less than £6 even before railcard discounts because it wasn't some early-morning period.
Journey time for me to get to work: 20-30 minutes in the other direction, very slightly over 20 miles. Costs £12 and can't be discounted below that because it's a peak train.

Both are returns, fwiw.

I don't think the loaned robotaxi concept is completely unworkable, just that the economics don't seem to make sense for most owners.

I suspect the use case will be long term loans where someone knows they won't be using their vehicle for an extended period of time (ie several days to weeks) rather than short term loans where people expect their car to make money overnight and be consistently back safe, sound, and fully charged in the morning.

So someone might loan out their vehicle for a week when they fly somewhere on vacation or for a day when they know demand will be extremely high (ie the day of a big sports game in town) and collect a small payday. But the people expecting the money it can earn overnight to significantly impact their quality of life will likely be disappointed.

I also think there's a universe where "robotaxi owner-manager" is essentially a job people have. It doesn't work with Tesla's current scheme, but I could easily imagine someone knowing how to take care of a vehicle amassing a small fleet of a half dozen or so vehicles that they do all the maintenance work for.

Some taxi companies can be that way, for my local taxi company if I want to go from my village (next to my town) to the nearby city I have to make sure I call the office in my town rather than the one in the nearby city or the price will be higher, even for the return journey later in the night.

I think the biggest argument against loaned robotaxis is that there’s no technology barrier today that stops people loaning their person-operated car to a taxi company, yet they don’t.

I'm not sure if there's more logic to that price differential or less. I'm going to go with more, because the journey in either direction I'm talking about is the exact same damn train and line. <_>

Two words: Uber and Lyft. The business models aren't perfectly comparable, but it's also the case that taxi drivers often use their vehicles for their own personal errands and such, so there's semantic overlap.

In essence, private owners will be like taxi drivers except that the car will be driving itself, saving their time and effort for something else. Oh, and the answer for who bears the costs of maintaining the vehicle may vary depending on what sort of arrangement they have with the robo-taxi operator. I imagine we'll see a number of different options before the industry converges on the best one.

The journey I was booking was the same, the only variable factor was which office I booked it via. Imagine if your train price varied depending on which station you bought your ticket from, for the exact same journey.

I know the US is different, but in the U.K. both such companies are just regular price hire companies. The drivers and vehicles are registered with the local authority and the drivers are actually paid employees of Uber/Lyft with legally mandated paid annual leave, pension contributions etc… Legally there’s no difference between an Uber and a limo service.

That is very different from what you’ve been suggestion until now, I can absolutely see independent taxi drivers (which do exist in the U.K., Uber is rare in being a company with directly employed drivers) getting self-driving cars and letting the car do the work, but that’ll be an adaptation by existing people in the market, not some great flood of new people into the market.

It really isn't. Either I'm expressing myself badly or being misunderstood.

As I said before, taxis and ridesharing exist today. As far as the passenger is concerned, autonomy would be like that only without a [human] driver.

As far as the vehicle owner is concerned, they contract with a fleet agency — it could be a municipal taxi service, a private company, whatever — and set hours during which their vehicle will operate as part of that fleet. The promise is that they get it back clean and well maintained, and earn a cut of whatever revenue it brings in.

While one might anticipate this creating an additional have/have-not class divide, I firmly believe that competition will drive marginal profitability down to a stable level that is beneficial on net for the consumer of these services.

If the total cost to own and drive a vehicle is X, and the cost to take robo-taxis to work and shopping and whatever is Y, as long as Y < X for most people, it's a win. Plus we'll have dramatically increased road safety, which has both direct and indirect value to society.

If a person is in a financial (or legal) position where they cannot drive their own car, their transportation options will be expanded by having robo-taxis. It's a win for them as well.

I suppose we could end up with an underclass of people who get blacklisted from taking rides because they abuse the vehicles or something, but this is not a problem technology can solve. Also, it already exists today. Trash your Uber a few times and see if they let you keep using it. Lose your driver's license for too many violations and good luck getting to work.

ETA: If we want to express things in pure economic terms, autonomy creates a new form of arbitrage in the price of transportation: between the marginal cost (and associated risks) of a human driver and the amortized cost of self-driving hardware and software. Sooner or later all arbitrage hits diminishing returns, after which it becomes more or less equal to any other investment.

The other form of arbitrage is in vehicle utilization. Most cars spend most of their time not being driven, so the fixed costs of ownership contribute more to the total cost. Robo-taxi service allows the cars to be utilized more efficiently, eliminating a portion of the waste associated with them.

Edited by Fighteer on Apr 29th 2024 at 11:18:00 AM

What you seem to have been saying (and correct me if I’m wrong) is that under this new definition of a taxi operator (any individual with a self-driving car that choose to have it take people places for money) we will see a huge expansion in the number of people who decide to become a taxi-operator compared to today (where it’s just people who do it as either a primary or secondary job), which is what I’m disagreeing with.

Even if it makes technical economic sense to loan out your self-driving car as a taxi I expect it to sit in the realm of people who do it professionally and that guy you know who has a side gig selling limited edition Lego, or that person who gets free stuff online then sells it on, or the person at your work that sells bath stuff as as side thing. Most people aren’t going to do it not because it doesn’t make economic sense, but because people are possessive of their stuff and value their time and comfort at a pretty high level.

Let’s say that my area currently has 100 taxis operated by around 100 people, if we convert all cars in my area to self-driving ones I don’t think we’ll see a dramatic change in the number of people involved in the business. Taxi companies will fire their directly employer drivers bringing the number of operators down, some independent drivers will start ‘operating’ multiple cars, some people will start new taxi businesses due to the lower barrier to entry and a handful of people may drop out of the industry due to enjoying driving more than fleet operations.

Taxis will be cheaper and this will see a moderate increase in demand, but it’ll be akin to the demand increase the cheapness of Uber bought to many cities, not a massive society wide realignment of how we do personal transport.

If anything we might see a bigger realignment from automatons public transport bringing down the price of buses/trams/trains such that local government are willing to run more routes and that sees people reduce their usesue of and ownership of personal vehicles.

Oh, I see what you're saying, but we aren't contradicting each other. I believe there will be an initial surge in vehicle owners attempting to use their cars to generate passive income by seconding them as robo-taxis, but this will hit a law of diminishing returns and eventually become only marginally profitable. (Unfortunately, Elon Musk is kinda-sorta promising this as a reason to invest in Full Self-Driving, something that will likely rebound negatively on the company, but that's a separate topic.)

Meanwhile, people who drive professionally (taxi drivers, Ubers, etc.) will be outcompeted by autonomy and will find themselves out of a job.

This won't happen overnight, of course. Even assuming every single new vehicle produced today were autonomous, it'd take at least 20 years to get all of the manually driven ones off the road. There will be time for society to adjust.

ETA: As a matter of fact, I would not be at all surprised if companies like Tesla initiate vehicle buy-backs to supply their robo-taxi fleets at some point in the future.

Edited by Fighteer on Apr 29th 2024 at 1:43:42 PM

People who drive professionally won’t be out-competed by the side-gig autonomous-taxi lot, they’ll be the first adopters.

People who drive professionally tend to have more modern vehicles and replace them more frequently, I expect that the taxi drivers round my area will have switched to self-driving taxis well before someone like me is in a position to buy such a car.

Sure Tesla plan to do a mass rollout using existing vehicle hardware, but even then I’d blind guess that Tesla ownership rates and driver-assistance-software utilisation are higher amongst the professional driving community then they are amongst the general population.

Okay, let's say a professional taxi driver buys a robo-taxi and puts it into this type of service. Because of the reduced marginal cost, they would earn less from it in fares, since that's the core rationale. So they either have to buy more cars or do something else on the side. They could make up for reduced fares with more time in service, I suppose, since they wouldn't be limited by the hours they are able to work, but that's just an extension of the logic.

In the end, they aren't a "taxi driver" any more, since there are no drivers. They're just an owner of a robo-taxi, something theoretically available to everyone. There would no longer be a bar in the form of insurance, licensing, training, union membership, medallions, or other factors.

We could have an entirely different conversation about the prices of taxi medallions and all the other nastiness around that particular enterprise, but that's not germane to this topic. A lot of taxi drivers don't even own their vehicle.

Edited by Fighteer on Apr 29th 2024 at 3:52:05 PM

If they’re still doing any actual work around it (like picking which jobs to send the robo-taxi on) then they’ve become an operator, a one-person version of a taxi-company office.

Insurance and licensing will still be a thing, in the U.K. you even need specific insurance to use your personal car for work purposes. In the end you’re still running a business where you have responsibility for the safety of paying member of the public, that’s always going to need insurance and licensing.

As for the rest, taxi medallions aren’t a thing here, I’m pretty sure it’d be illegal to require someone to be part of a union and you only need specific training if you’re doing street pickups without any advance booking.

It’s also a US specific problem to the best of my knowledge.

Again, I think I see at least part of the issue. You would not dispatch your own vehicle in this situation. There would be no "individual operators". As I said, you would second your vehicle to a fleet that would do all the administrative work.