[kjksf]

Here's what this research tells me: demand for transportation is much higher than the current use of transportation.

Current use is limited by cost and convenience.

Self-driving car is likely to fix both so it'll lead to more use of transportation.

This is a good thing. We're currently deprived and if things go well, we'll get more of what we want.

The congestion problem is mostly overblown.

First, maybe with rare exceptions like L.A., the traffic is only bad during rush hour, when people are going to work and getting back home. Other times there's plenty.

Second, the way we drive currently is very inefficient. Just last weak I was walking in San Diego along a street at ~5:30 PM i.e. rush hour.

I just eyeballed but ~80% cars were single person.

Not to mention that ~40% cars were gigantic, because it looks that if people can afford gigantic cars, they'll buy them. And in US they can afford it.

Robotaxis would fix those 2 issues.

The cars would no longer be an expression of personality and a status symbol but a utility operated by an organization focused on practicality and cost, like buses and trains.

It's also very easy to use pricing to force people to use the available resources efficiently during congestion.

Let's say a ride is $10 if you drive in a car alone. $5 if you share with another person and $3 if you share it with 2+ people.

If that pricing delta is not enough, increase the price of to $20 (vs $3) or to $50.

Or provide commuting passes tho employers (kind of like Google buses) where a company pays a $100 to robotaxi company per month and the employee gets to use it for free for commute, but only in shared mode.

The future with robotaxis is much brighter than those doomsday prediction of traffic.

[CydeWeys]

> The congestion problem is mostly overblown.

It's really not. Congestion is a huge problem already, and increasing the current road utilization by 83% will bring entire cities to a standstill for the entirety of waking hours.

> I just eyeballed but ~80% cars were single person.

Yes, and in the AV future the majority of cars on the road will have zero occupants, so things will get much, much worse as far as how many people are actually being transported vs road utilization.

You then go on to talk about robotaxis, but that supposes that most people will give up the idea of personal car ownership entirely instead of just buying their own AVs; this doesn't seem likely.

[rrrrrrrrrrrryan]

We're going to have to slowly shift to more usage-based pricing if we want autonomous cars to work. Start off charging tolls per mile for zero occupancy vehicles when they drive in congested areas, then expend it to single occupancy vehicles in congested areas, etc.

A bit more controversial, but I personally love the idea of a congestion-free fast lane (say, 100+ mph) exclusively for autonomous vehicles, with prices that fluctuate based on supply and demand. We already have this for a couple freeways in Southern California and it's lovely to have as an option.

[neaanopri]

Gas taxes are already effectively toll-per-mile, with a bonus incentive to get fuel efficient cars.

If we start seeing enough EV adoption to dent gas tax revenues, that's a win.

[maremp]

If you’re talking gas tax in the US, I’d say that’s laughable. Using eia.gov[1] as a reference, the prices are half of what we have in the EU. We’ve observed a steady increase in the prices (mostly due to taxes). But it didn’t have the effect, people just deal with the higher price and carry on. There’s even more cars than there used to be. And the average living standard in the east/SE EU is lower of that in the US.

When I was in Göteborg, Sweden, I saw this system where the toll increases based on the time, and it’s highest (~30 SEK = ~3€) in the rush hour. There was no price in the off hours. Maybe if it costs more to drive in the rush hour, one might think about not doing it or finding someone else to split the cost.

To rely on the “tax system” to take care of the congestion/traffic pollution problem, we’d have to bump the taxes and tolls by 10x to see any effect. EV is a step in the right direction, but only by a small amount. We need incentive to drive less, something that forces everyone to think before they drive. And by everyone I have in mind the person who is driving and also the person who is making them drive to the office to have their physical presence to do the same work they can do from anywhere. Make it more expensive for a company to have car riders and make them count commute as part of the workday. I believe remote work is a part of the solution to this and making it more expensive to get everyone to the HQ every day might make more managers reconsider if it’s necessary.

[1]: https://www.eia.gov/petroleum/gasdiesel/

[CydeWeys]

The linked article addresses how gas taxes are nowhere near good enough to solve the congestion problem, and especially not when more cars go EV.

[wutbrodo]

This is already the case, and has been so for a little while.

[erik_seaberg]

The number of people who need to commute during rush hour can't increase 83%. Maybe some transit commuters will switch to cars, but maybe some car commuters will finally be able to switch to transit if these make the first and last mile feasible.

And 83% was the first and only week. Maybe the number doesn't stay that high once indulging in the novelty wears off.

[Thriptic]

It probably could increase quite a lot because more people might be willing to live far away and enjoy cheaper / more housing if it means that commuting won't be dead time. With self driving cars you could actually work during your commute. As you point out, a number of existing commuters might also transfer from train to car travel.

[CydeWeys]

You're forgetting the unoccupied trips. There will be unoccupied trips running errands even at rush hour, because why not, you're not sitting in traffic, just your car.

Also you seem to be thinking more of the robotaxis model, but the article is talking about the more likely future of everyone continuing to own their own car, only now it can drive. If cars are driving back empty after a commute then rush hour will have roughly twice the traffic.

[xnzakg]

One way to prevent congestion would be to instead of paying for the distance traveled, you simply paid for the time the car is on the road, giving a financial incentive to send the car outside of rush hours. Would probably also cause people to use alternative routes rather than everyone trying to use the shortest one when is congested.

[CydeWeys]

Yes, this is one of the main solutions proposed by the linked article.

[BubRoss]

Wouldn't that be a trip that someone would normally be driving anyway? Also wouldn't it make sense to have a store deliver to lots of people on one circuit, this reducing traffic?

[CydeWeys]

No, once you don't need to drive it personally you can run a lot more errands, and thus will. See the linked article.

And look at how much more random stuff people buy when a click of a mouse gets it delivered to your house vs having to go to stores looking for it.

[BubRoss]

If one car is going in a loop and delivering to lots of people it would be less time on the road than all the cars of the individual errands.

[Aloha]

If the future is car as a service, then yes congestion does not automatically get worse, you call an automated car to pick you up, and another one to take you home, it might also mean that more ride sharing happens too

[Moru]

But we already have buses and bicycles. Car as a service might just as well make the congestion worse since it will go empty to pick you up.

[maccam94]

Think Uber carpooling, minus drivers, much cheaper, and short pickup times. Maybe 6-10 person shuttles instead of sedans/SUVs in denser areas. To a car service, moving an empty vehicle is inefficient and a waste of money.

[masonic]

Keeping an empty electric vehicle in motion all the time is cheaper than parking it, in urban contexts.

[jcranmer]

Because you're not pricing the externalities of having zero-occupant vehicles roaming around streets that are already at or near maximum capacity.

[Supermancho]

> Yes, and in the AV future the majority of cars on the road will have zero occupants

Where do you get that idea from? or maybe, how did you come to that conclusion? It's not evident that would be the case. Traffic mirrors demand. Demand goes both ways, unevenly. Prior to the end-of-the-work-day, you would see an increase in empty vehicles consolidating for the imminent exodus of a city, but that would hardly be the common case.

> that supposes that most people will give up the idea of personal car ownership entirely instead of just buying their own AVs

There are multiple ways to ensure this occurs. The same reason horse carriages aren't allowed on Freeways (legal/safety barriers), or almost nobody uses gliders or natural gas cars to commute (impractical), etc. US Society is subject to change/more malleable than you might imagine. Maybe not in our lifetime, but ubiquitous phone booths disappeared in mine, so I hold out hope.

[spc476]

So here's a hypothetical scenario. Car drives kids to school (in my experience, school tends to start before work). Drives back empty to pick up parent to drive to work. In the afternoon, car drives (empty) to pick up kids and drive them home, then drives (empty) back to work to pick up the parent. Half the time it's empty.

[Supermancho]

> Drives back empty to pick up parent to drive to work.

This implies there are no other people for which a rideshare would be suitable in either direction at any given time. This seems, like an unreasonable way to predict how mass AV would work. Again, demand is asymmetrical, which means there are opportunities for aggregation and overlap, when talking about mass adoption to the point that "the majority of cars on the road will have zero occupants".

There's the issue with demand for AV availability (ie traffic), which can never be reduced to zero. This would incentivize better judgement than "I'll stagger the school and work times"

Logically, why would AVs be 1-4 occupant vehicles? Once you have AV, you have AV vans and buses, similar to airport shuttling.

[CydeWeys]

We're not talking about rideshares, we're talking about autonomous personal vehicles. Most people are not likely to give up their personal vehicles; they'll just have self-driving ones.

[wutbrodo]

This seems like an odd argument to me: "enabling efficiency isn't useful if you assume people's habits won't adapt to the new status quo". Once you've moved personal vehicle use of the sort you describe from essential to luxury, I don't see why you couldn't make users bear the cost of their frivolous externalities.

[Supermancho]

> We're not talking about rideshares, we're talking about autonomous personal vehicles

> they'll just have self-driving ones.

That won't last long, because of market inefficiency. Why pay for maintenance and energy on a vehicle? It's just as unrealistic as saying we'll keep personal vehicles running on gasoline, to me.

[dsfyu404ed]

In your scenario the number of trips is the same but the car is more often empty because the parent doesn't have to chauffeur the kids around.

[CydeWeys]

Now the parent can be in a different car at the same time, going something else, thus doubling the number of cars on the road.

[dillonmckay]

Yes, that was the point.

[Moru]

But it still takes up space on the road

[darawk]

> It's really not. Congestion is a huge problem already, and increasing the current road utilization by 83% will bring entire cities to a standstill for the entirety of waking hours.

That's not how equilibria work though. Currently congestion is the primary bottleneck on personal consumption of transportation in major metro areas in the US. That will not change with AVs. That is to say, we will probably reach a similar congestion equilibrium as we have now, there will just be more transportation happening, which is a net good thing provided there are not also negative externalities (such as emissions) from that transportation bump.

[amarshall]

Driverless private cars (which the article is about) and driverless taxis aren’t the same thing. The former does nothing to solve the traffic density or vehicle size problem. Even focusing on driverless taxis, Uber and Lyft have offered for some time options for nicer cars, larger cars, and nicer larger cars, so presumably there’s a market for taxis to still be a status symbol, unlike you state. Finally, your economic incentives have nothing to do with driverless cars; it’s just a tax or toll on non-pooled taxi rides—that could be implemented right now.

And, in the end, even a full car is still less dense than a bus or train.

[Falkon1313]

The uber/lyft/limo large cars vs standard cabs/cars are usually used for a trip when there are more passengers or cargo. With private cars, the larger ones are bought when there might ever be a trip with more passengers or cargo, even if 99% of the time they just have a single occupant.

Also normal cars spend most of their time parked. Driverless taxis wouldn't need to.

[Someone]

”The cars would no longer be an expression of personality and a status symbol”

You’re not the only one claiming that, but I don’t think it automatically follows. People can currently afford to buy cars for themselves, so unless self-driving cars will be (much) more expensive, people still would be able to afford them. So, in this case, the question is what else, if anything, people would want to use their money on instead of on owning a car.

If, on the other hand, self-driving cars will be much more expensive, there’s a risk that they will be too expensive. That may be offset by having self-driving cars make more hours each day, but I don’t think that’s a given.

One reason people own cars is that it guarantees a car is available when they need it. If people value that highly, self-driving cars will have to cater for that. I think that would mean the number of self-driving cars would have to be fairly similar to the number of cars now on the road. If that’s the end-game, why wouldn’t many people own a self-driving car, or lease one for their personal use?

[kjksf]

When you use X and then a substitute Y arrives that is 10x cheaper and more convenient, you'll switch to it.

Using robotaxi network will be much cheaper than total cost of owning a car, mostly because cars are dramatically underutilized.

On average a car is used 1-2 hrs a day out of 12-16 hrs it could be used as a robotaxi. To start off we have 5-7x price difference.

Then most use is not shared. Assuming robotaxis will be used by 2 people on average, it's another 2x cost difference, 10-14x combined.

The cars will be much cheaper. Think $25k Toyota corolla, not $45k SUV or pick up truck. Fleet operators will ask for cars that are cheap and reliable, not for luxurious interior.

The cars will last much longer. 500k miles vs. 100k miles of a typical car for personal use. 500k is already a reality with Teslas (https://qz.com/1737145/the-economics-of-driving-seven-teslas...) and Tesla is working to make 1 million miles cars.

Not to mention that GM spends 3 billion dollars on advertising each year, which goes to increase the cost of your car.

It'll be a no-brainer to pay $200-$300 per month for (effectively) unlimited robotaxi service than $400 per month for a lease + $100 month for insurance + fuel + cost and inconvenience of refueling.

Furthermore, you won't be able to buy those cars.

Consider the following scenario to its logical conclusion:

Tesla (or whoever) introduces a robotaxi service.

A car that costs $30k to make can be sold for $40 k to you or put to use as a robotaxi and make $100k - $300k over its lifetime (big spread because it's hard to predict how the pricing will pan out).

Why would Tesla sell a car to you for $10k profit if they can make $70k profit?

For the first 10 years they'll be busy making cars exclusively for robotaxis and they'll prioritize grabbing market share. There will be no $40k self-driving Teslas to buy.

But won't GM or Ford or Toyota sell me that car?

In 2008 car companies went bankrupt because of 20% drop in car sales.

If robotaxis take off, they'll shrink car market by way more than 20%, permanently.

I can't see how this won't end up in bankrupting pretty much every car maker as we know them.

Those that will have self-driving technology will also switch to making robotaxis exclusively.

Those that don't will either become OEMs to the kinds of Waymo (and eventually be bought by them when they vertically integrate) or go bankrupt.

Privately owned cars might become like yachts: sold in very low numbers to multi-millioners.

[Someone]

”On average a car is used 1-2 hrs a day out of 12-16 hrs it could be used as a robotaxi.”

…assuming demand for cars is flat 12-16 hours per day, and that you could somehow, make all their operating hours useful operating hours. For example, society would have to change for commuter cars to end up in places where they immediately can be used productively.

”Why would Tesla sell a car to you for $10k profit if they can make $70k profit?”

Because, if they don’t, somebody else (e.g, a Chinese car manufacturer) will sell me a car for that $10k profit, and Tesla will make zilch.

If cars become that much cheaper, and wouldn’t need a driver’s license to use them, I don’t see why, by necessity, fewer people would own cars.

Your scenario assumes some manufacturer will gain a near-perfect monopoly. That might happen, but it’s not a given.

Also, I don’t think it is a given that the government wouldn’t break up such a monopoly.

[rootusrootus]

> Why would Tesla sell a car to you for $10k profit if they can make $70k profit?

Because there is a niche that is profitable and if Tesla doesn't fill it, someone else will? After all, Tesla has had no motivational difficulties trying to fill that niche right now, so someone will be willing.

I also question the 70K profit part. Assuming robotaxis actually become a thing, why do you think the cars will make much profit at all? There will be a race to the bottom. I know a lot of Model 3 owners trying to justify their FSD purchase by suggesting they will eventually make back all of that money and more, but I can't figure out why they think this time they will be immune to market dynamics.

[jcranmer]

> It'll be a no-brainer to pay $200-$300 per month for (effectively) unlimited robotaxi service than $400 per month for a lease + $100 month for insurance + fuel + cost and inconvenience of refueling.

It's already possible: a monthly transit pass is often in the $100-150 range. The fact that people have been choosing to forego viable transit options in favor of a much more expensive personal vehicle option for a long time is strongly suggestive of the fact that people really don't want to give up their cars.

[SergeAx]

No, this is because transit pass cannot fulfill ALL their needs of transportation, but mostly commuting.

[jcranmer]

That sounds like someone who has never even attempted to make that work.

Transit is least effective in suburbs, where you have to try very hard and basically live in exactly the right location to make a transit-only experience work. But in both large urban conglomerations and even in many small rural cities, it is actually quite effective to do all of your transportation needs using local transit.

[OnlineGladiator]

> The cars will last much longer. 500k miles vs. 100k miles of a typical car for personal use. 500k is already a reality with Teslas (https://qz.com/1737145/the-economics-of-driving-seven-teslas...) and Tesla is working to make 1 million miles cars.

First of all, thank you for an interesting article! Looking at it quickly though, if I make some rough extrapolations from their data, it seems like most Teslas will have lost about 40% of their battery life by 500,000 miles. So I really think it's disingenuous to say that a Tesla can drive that long, when its range starts to make it significantly less useful, when replacing the batteries would cost more than buying a cheap car. Basically I think you are just exaggerating the benefit of longevity for electric cars.

That said though, the Tesla batteries are holding up better than I thought and I agree with you electric cars are better for higher mileage applications!

[skinnymuch]

Why wouldn’t there be different classes of cars? If I’m going to work every day in an autonomous car and the price is not that expensive, there’s a chance I’ll decide to pay 25% or 40% more of what you say will be a dirt cheap price, to be in a nicer car with better seating and amenities. Especially if it becomes standard to pop open an iPad or laptop during your commute.

Others have already covered how any potentially profitable niche will be covered.

[NicoJuicy]

> It'll be a no-brainer to pay $200-$300 per month for (effectively) unlimited robotaxi service than $400 per month for a lease + $100 month for insurance + fuel + cost and inconvenience of refueling.

My car costs 1.800 € ( I'm not the first owner of that Mercedes), I barely use it as I live in a big city.

I'm not paying 200€/month ;)

[mcguire]

Then, when Tesla owns the personal transportation market, they can raise prices to satisfy their stock valuation. It's brilliant!

:-)

[imtringued]

I bought my car at 2000miles for $8000. It now has 100000 miles after 5 years, presumably it will turn into dust in another 10 years which kills the underutilization argument. I see no reason why it would be cheaper to use a self driving taxi.

[pretzelstrash]

> On average a car is used 1-2 hrs a day

That amount of time spent with a scooter service would cost $400-$800 a month. A robotaxi service could be more expensive than you think.

[throwitawayfam]

People could send their idle vehicle to work for them, while they work, while they sleep.

[jcranmer]

Traffic demand is very bursty; most people work and sleep at the same time--so who is driving the demand during that time?

[cookie_monsta]

> the traffic is only bad during rush hour

This would be ok if the rush hour actually lasted 60 minutes. In my city (pop ~6M) it goes from somewhere between 7-10am and 4-7pm with a little spike around lunchtime.

I don't work downtown and have the luxury of scheduling most of my driving as I please, so I nearly always opt for around 8pm. Being that my city is playing catch up trying to ease rush hour congestion, I often find myself alone on 4 lane arterials.

I am literally King of the Road.

Robotaxis aren't going to fix this, nor, sadly is throwing any other type of monetisable product at it. What we need to do is decentralise where people work away from downtowns, stagger work and school starting times so that the entire city isn't running around trying to hit the same 30 minute window and set up incentives for companies to accommodate remote workers.

Anything else smells to me like the auto industry trying to sell more cars and the infrastructure industry trying to sell more roads. They're free to do that, but under the guise of easing congestion? No.

[ralphc]

I was thinking you were talking about Atlanta until you mentioned traffic being light at 8pm. A lot of times it takes later here, even out to the suburbs.

[viburnum]

So much transporting is just to mitigate land use failures. More transportation is solving the wrong problem.

[bsder]

Um, you could solve most congestion simply by staggering working hours.

The problem is that even the Bay Area can't seem to pull this off.

Single people are generally fine. But the moment you have kids, you are now on regimented schedules that force you into some semblance of 9-5 ... and now you're part of the problem.

[maxerickson]

Solving it is better than continuing to live with it.

[CydeWeys]

The point they're making is that the only possible solution is to solve the land use problem, so that the majority of people can live lives that don't require individual automobiles. Broadly speaking, this means building densely enough that walking, biking, and mass transit are suitable for most trips.

[maxerickson]

Yes, their point is clear enough.

My point is that the politics of landuse are hopeless enough that other mitigations will be more beneficial for quite some time to come.

[CydeWeys]

I disagree, I think the politics of land use are more easily solvable than the simple physics problem of too many cars, not enough space. Plenty of other countries in the world (and even some US cities) have solved the land use problem to the point that most people don't use space-inefficient private vehicles to get around. It's solvable and there are working models to emulate. This is not true of your alternative.

[bcrosby95]

It's not financially possible to fix this quickly. You're talking about tens of trillions of dollars in new housing, at a minimum. Then all the commercial, infrastructure, etc. on top of it.

[maxerickson]

How does Northern Wisconsin emulate the Randstad? Does everybody move to Green Bay?

[rootusrootus]

So we all just need to compress our cities until they achieve NYC density?

[optimiz3]

So much health care is just to mitigate population control failures. More health care is solving the wrong problem.

So much computing is just to mitigate math education failures. More computing is solving the wrong problem.

Etc.

Thank goodness some people are willing to innovate.

[rootsofallevil]

No OP but I think you are misunderstanding the position:

Solve the land misuse problem and everything else follows

[BubRoss]

Self driving cars seems more realistic

[6gvONxR4sf7o]

How would robotaxis fix the ~80% of cars being single person? It seems like it would make it even worse, replacing some of those with zero people in the car.

[kjksf]

Network of robotaxis has a wonderful quality of increasing utilization with density.

It's a self-healing system: the more people use it, the more efficiently we can use it. Efficiency, it this context, is number of people we can transport per hour on a given stretch of the road.

It's important, so i'll emphasize it: high utilization of roads is a good thing. Roads are a fixed resource that have already been paid for. High utilization means we're providing more transportation value for the same, fixed price.

Most of the time roads are empty. Case in point: it's 1 PM in a busy part of San Diego. I just looked out of the window and the road is empty.

Even if there were 10 more, single occupancy cars on that road it would not degrade transportation service for anyone. Plenty of capacity.

The bad part is congestion caused by going over capacity. That's the friday night scenario in my neighborhood or commute to/from work traffic during weekdays.

Robotaxis give us obvious tools to combat that.

When the need for transportation at a given time is so hight that it would cause bumper-to-bumper traffic, it implies that a lot of people are sharing their route with other people.

It is the case today but we have no practical way of arranging for those people to share the car.

With robotaxis, the fleet can see that 4 people on the same block want to go to roughly the same place so it can direct them to the same car, so that they can more efficiently utilize fixed resource of cars and roads.

What if they don't want to share a car with other people?

Make them with congestion pricing.

Make them pay $10 for a single person ride vs $3 for shared. If $10 doesn't work, then jack it up to $20, $50 or $100. There is a price that will make 90% choose shared ride during congestion.

Or if things are really that bad and the trip is non-essential, they'll choose to shift their trip. Maybe 5:30 PM is not the best time for trip to walmart.

We can certainly do better than what we do today during congestion: mostly empty, overly big cars.

[NicoJuicy]

I'm not sure if we are fixing the right thing with robotaxis.

If you need commute pricing, to make traffic less busy and robotaxis successful.

Then why don't we fix commute riding first, instead of waiting for robotaxis and waiting for this problem to popup.

[6gvONxR4sf7o]

That would be the case for uber pool too, but it isn't used nearly as much.

[yazaddaruvala]

Today we buy cars for presentations of status, personal expression, enjoyment and “the worst case” I.e. family and kids all need to go somewhere at the same time.

“I’m a truck person”, “I’m a hybrid person”, “I’m outdoors-y”

Then we use it 95% of the time for single person use, to commute (not enjoy) with zero recognition of personal expression.

Even if people don’t share the robotaxis, they will be optimized for pragmatic use. Taxi sizes will scale with number of people (think enclosed, self driving motorcycles, etc).

[6gvONxR4sf7o]

In practice, I don't expect to see it. Economically, there's value in having whatever cars are closest to you be ones that fit your group. I expect that value will exceed the savings of a smaller car. I bet we'll see something like uber vs uber-xl, rather than something as granular as you're suggesting.

[tux1968]

True, but if you're alone, you could pay less to share a ride with others being picked up and dropped off along the way to your destination.

[OnlineGladiator]

The same is true for Uber. It's much less popular than people getting their own rides.

[jfk13]

> ...it'll lead to more use of transportation.

> This is a good thing.

Why? Wouldn't it be better if we could reduce our dependence on (mechanised) transportation overall?

[kjksf]

Because in all of human history humans have shown zero interest in depriving themselves of things they want.

I want transportation. You want transportation. Everyone wants transportation.

You, me and everyone else wants to go to the beach, get food delivered, get 1 day shipping from Amazon, eat bananas grown in Chile, drink coffee harvested in Africa, drink wine grown in France, go to a baseball game or visit relatives for Thanksgiving.

All are optional activities you can do without. Are you doing without them?

I need transportation. You need transportation. Everyone needs transportation.

You, me and everyone else needs to go to work, go to a hospital etc.

Reducing transportation is not an option. You'll not give up on things enabled by transportation.

So the next best thing is making transportation cheaper and more efficient.

We've optimized food production to produce 10x more food using the same amount of land. It was a good thing because the option of starving or killing 9 out of 10 people is not very appealing.

Robotaxis will allow us to optimize transportation to better use the roads and cars. High utilization of the roads (as long as it doesn't become a crippling congestion) is a good thing. It means we're using a fixed resource to provide as much of what people want at the lowest possible price.

And since people bring buses as some alternative to robotaxis: a self-driving bus is also vastly superior to regular bus and when we have density of 100k robocars in a city the size of SF, it'll be a no brainer for fleet operators to have part of the fleet be minivans and buses, taking 16+ people, for the times of congestion.

[sooheon]

> Because in all of human history humans have shown zero interest in depriving themselves of things they want.

Any malignant cancer shows zero interest in moderating its growth, until its host dies.

I agree with your description of reality, but I think fundamentally we need to price in ecological and medical externalities before I can be quite as optimistic as you are.

[Spooky23]

The problem is those numbers are off.

With margin, the services are always more expensive. I drive a fancyish SUV, and my average per-trip cost is around $3. I can Uber for $8-10 minimum or ride a bus for $1.5.

Robot cars are going to be more like Uber, as you won’t have the driver losing money on depreciation.

[yazaddaruvala]

Robotaxis should roughly drop the cost of “Uber’s” to 1/3.

Currently Uber only collects 25% of the fair for itself. The rest is for the human driving.

Robotaxis can also be used 24/7 so better capital utilization on the initial investment for the car.

Maintenance costs reduce with scale (+in house mechanics to improve margins). Maintenance costs reduce with electric drive trains. Maintenance costs reduce with better initial manufacturing (to optimize overall car costs rather than optimizing for initial sale costs to sell to individuals).

Increased scale can mean reduced margins i.e. Amazon.

And this is all before the tech is democratized. Once there is competition, Uber’s 25% cut of current fairs (for software engineering and support) starts to drop too.

[Spooky23]

Remember that Uber doesn’t make money. And they pay their drivers less than the driver is losing to depreciation. Also, the robot isn’t free. Software isn’t free — it’s really expensive actually.

All of the cost savings that you describe also don’t require a robotaxi to realize. The cost per trip of my wife’s Nissan Leaf is about 60% of the SUV.

Even the durability can be addressed by buying a Toyota or Honda vs a Jeep or whatever.

What will be an issue for the Ubers of the world is accounting. Without some sucker wasting his car away, Uber needs to invest billions in assets and infrastructure. That capital structure is more like an airline than Google. Not something Wall St rewards.

[yazaddaruvala]

> That capital structure is more like an airline than Google. Not something Wall St rewards.

This is the part we are disconnected on. It is the hardest part to guess about.

If the accounting works out like an airline, I agree with you that costs will remain high and will not change transportation significantly.

However, unlike an airline, the variable cost per ride is low. The way I see it, the economics will work out more like a data center (relative to an airline). Which is something Wall St does reward.

[desas]

> Currently Uber only collects 25% of the fair for itself. The rest is for the human driving.

Who pays for the car currently? In the UK it's not uber, it's the human as far as I've understood it.

[yazaddaruvala]

If you believe Elon Musk the Model 3 currently sold for $50-60k USD will be worth $100k USD as a robotaxi. For us in 2019 it really is hard to imagine the worth of a car that is used 99% of the time rather than parked 95% of the time.

When robotaxis have been real for a decade, car manufacturers will build and run robotaxi fleets.

Unlike today, vertical integration of manufacturing, maintenance, recycling and operating the fleet of cars creates a lot of consistent demand and can bring significant economic advantages even more than Musk’s current estimate.

Even the car manufacturers goals change. Rather than optimizing to sell cars (have them degrade and sell more), cars will be optimized for durability, maximum materials reuse, lowest variable cost per ride, ergonomic rides.

[jcranmer]

> Rather than optimizing to sell cars (have them degrade and sell more)

People often talk about cars being afflicted by this style of planned obsolescence, but the facts are that cars last far longer now than they did 50 years ago. What planned obsolescence does exist is in the form of trying to shame you for running an "old" car as opposed to knowingly making the manufacturing quality of cars utter shit.

[yazaddaruvala]

I agree, as a non-professional driver the lifetime of the car is measured in years/aesthetics.

Meanwhile, a professional driver can put on 25k-50k miles per year[1]. Compared to 10k-15k miles per year on average[2]. Meaning professional drivers hit the car’s mileage limit, 200k-300k in 6-8 years. Already it would be better for them to have cars that are optimized for cleaning, maintainability, and longer lifespans. And this doesn’t even talk about lifespans of wheels or the expedited cost of maintenance.

Taking this math further: This is targeting 40 hour work weeks for professional drivers. If we targeted 75% (a lower bound) of the total hours in a week 168, we see robotaxis will drive 3-4x the miles of today’s Uber drivers. Setting lifespans of 2-4 years per taxi.

Large robotaxi fleet operators will likely become manufacturers but regardless they will change the mental model of how car manufacturing currently operates.

Robotaxis will not only absorb the profits of the taxi industry but also the profits from the car manufacturing industry, car maintenance industry, car rental industry, last-mile delivery (including food delivery) industry, rental housing industry, and more. The scale will eventually be unimaginable allowing the margins to be astonishingly low.

[1] https://www.quora.com/How-many-miles-does-a-full-time-driver...

[2] https://www.carinsurance.com/Articles/average-miles-driven-p...

[l3robot]

Not also to mention something often oversight. An autonomous car can park far away from it's user, meaning we could drastically reduce de number of space taken by parking in the city. First, giving back this spaces to the people as green parks and services places. Second, lower one of the biggest problem in our current city: urban heat island.

[amarshall]

But then you’ll likely have a bunch of extra traffic created from the cars going to and from their owner to parking, rather than just to and from their destination. Or from the car just circling the block because the owner would rather not have to wait long for its return.

[jellicle]

Right. You go to the opera, you order the car to drive around the block for two hours and then pick you up. Cost of gas is cheaper than parking and by definition, you're not affected by any traffic congestion caused as you're sitting watching your show. Sure, 1000 other people in the theater are also doing it, but that's not your problem.

30 minutes before the show is over, you order the car to wait in the "no stopping" zone in front of the theater for easy access. You exit the theater and find your car is tangled in a mess of traffic that takes 45 minutes to clear as most of the cars are double-parked and blocking traffic, waiting for their owners. For the duration of the show, no useful traffic has been able to move for a mile in any direction, as empty cars are just idling around and are content to wait, fully stopped.

Tragedy of the commons.

[alasdair_]

The “circling the block” is something that Uber etc. already exacerbate.

I think one solution would be to charge for each minute you are in a congested area to avoid having a whole bunch of empty cars clogging the streets.

[CydeWeys]

There's a big limit on how bad it can be with Uber though because each Uber is currently driven by an actual person. That person costs money and places a natural limit on how many unutilized Ubers there will be.

With AVs, this limit no longer exists, and people will start using AVs for runs that currently are prohibitively expensive with care that require drivers.

[ajuc]

> Let's say a ride is $10 if you drive in a car alone. $5 if you share with another person and $3 if you share it with 2+ people.

This is overly complicated, hard to implement and to prevent abuse. There's a much simpler way already in place - tax on fuel. And it's much harder to circumvent or abuse. You can hope they won't catch you driving alone and paying as if you drove with 3 people in car. You can't hope to drive without fuel.

Increase the fuel prices 10 times and people will be much more efficient with their use of transportation. You use almost the same amount of fuel no matter if you move 4 people or 1. And public transport uses slightly more to move 20 or 50 people. With high enough fuel prices we can reduce traffic jams to 0.

The problem is - it will suck and people will complain. Ultimately people spend time in traffic jams because that's what their least sucky option is.

[bbatha]

Another point on smaller cars. Robotaxis allow tuning the size of the car to the workload. Commuting? Two person car. Shopping with the family? Sedan. No need to make everyone pick a car that meets their general needs any longer.

[m463]

> use pricing to force ... resources efficiently

I think this is the answer. It just won't be a very popular one, because it will on the surface benefit wealthy people.

[whoisthemachine]

It's also highly possible that robotaxi companies would incentivize efficient use of their vehicles in order to drive costs down (and increase profits).

[masonic]

They'll optimize profit by being able to respond instantly to hails, capturing as much market share as possible. So, the entire idle fleet will orbit downtown all the time.

[whoisthemachine]

Is that necessary? Couldn't such companies also have many cars parked around downtowns, ready to be activated for a ride at any time? Remember, without a driver, there's no need for someone to "orbit", looking for riders, rather the riders call the cars themselves.

[masonic]

parked... where?

Urban parking is already expensive and often hard to find.

[runeks]

> Current use is limited by cost [..]

Use is always limited by cost. It’s not possible to separate demand from price, ie. demand only makes sense for a given price.

[em-bee]

demand for transportation is much higher than the current use of transportation.

this is an important observation. and the demand will only increase.

because as it is easier to get around we are more likely to meet people at further distances, leading to families that are connected across these distances and thus increasing the demand for travel even more.

[kps]

Single-person cars can be half width and share a lane, and run much closer together at speed.

[CydeWeys]

This is only really a wide savings once manually-driven cars are banned from the roads entirely.

How long, if ever, do you think it will be for that to happen?

[beamatronic]

Cars are a dead end. No matter who drives them. We need subways like Manhattan. Or at least like the DC metro. Eminent domain and political willpower is all it takes.

[justicezyx]

> First, maybe with rare exceptions like L.A., the traffic is only bad during rush hour,

If this is not the only problem, what is the problem...

[coding123]

I imagine the bigger problem for road maintenance is the off rush hour trucks driving from 1am to 5am that are tearing up the road at 70mph. Not as much damage when they are stuck at 35mph. So different problems.