This is a common pattern to electric and self-driving enthusiasts, but unfortunately it's the automotive equivalent of Wired articles about the internet circa 1996.

Problems include:

- Cost. All of the (limited) Tesla automation technologies cost more than my last car.

- Liability. Who is liable for mishaps? As Uber has demonstrated accidents and traffic violations happen with automated tech.

- Utility. The majority of users aren't high income people in SFO and NYC. We have parking, getting picked up is cool, but not high ROI.

- Pool vs own. As we've seen with transportation services as varied as stagecoaches, cabs, railways, and airlines, service based transportation models aren't cheap. Service price is always demand driven, it will cost more when you need it.

- Owning a car in the US is one of the greatest values available in any market. I can be at any point in the CONUS in <3 days for under $500 with most cars.

I think self driving cars may put the bullet in some cabs for good after Uber and Lyft implode, and may bring train-like scale to intercity transit. But the fantasy being sold today is just that.

> Cost: We're not there yet, but not paying registration/insurance/repayments/garage space will make this economical in time - at first at least for second car owners > Liability: Insurance - I think automation will be safer i.e. less accidents > Utility: I think door to door driving without looking for parking is a pretty good perk > Pool vs. own: Both of these models will be competitive with each other, so in the end it won't matter

plus there will be additional benefits Safety: Less drunk drivers on the road, less accidents from fatigue, etc. Cost: at a minimum, it will bring cab prices down Efficiency: Less traffic jams as autonomous systems won't slam on the brakes when they see a police car, etc. Smoother flowing traffic, higher legal road speeds if reaction time is shorter

And then there are other non-car transport which will benefit: logicistics, mail, deliveries etc. Minibuses that could pick you up door to door through automated route planning etc.

Thank you, I've wanted to make a point similar to the one you've made but already tied it all together.

Good point. Cars will still be valuable for road trips, but lots and lots of people will choose to skip car ownership and just settle for local, self-driving travel.

Maybe there will be a stratified market for rental self-driving cars; one for "just this trip" and another for "five days of exclusive use for a road trip."

I don't think the change will be quick, but I think you're overestimating the problems and underestimating the potential benefits here.

Cost is a short term problem. Costs will go down. Liabilities can be insured against.

Here is where I see it going:

Stage 1: Taxis start being replaced by self-driving vehicles at much lower cost. Car rentals too - liabilities potentially go down once they reach a certain level of safety as you don't face the risk of a poorer than average driver. Usage skyrockets as costs drop, and as services can cut pick-up times drastically by more optimally having a larger fleet parked around town and/or driving around town.

Stage 2: We start seeing pooling options from more and more rental providers to deal with high demand situations. E.g. Rental company crunches their numbers and see that my road => the local train station always maxes out capacity during rush hour and decides that rather than buying more vehicles, surge pricing coupled with offering a discount that brings the price back towards normal for each rider as long as it at most takes X minutes extra will be popular and more profitable.

Stage 3: They put in minibuses on some of the most congested streches and/or team up with the local bus companies to launch apps where you can tell them you're at the stop and get guaranteed pickup within Y minutes by either the regularly scheduled bus or a car. You pay a slight premium for the guarantee, which covers the car when the bus won't be there and a profit share with the bus company. (For me the only reason not to consistently use the bus is that if I need to be somewhere urgently, I can't always risk waiting for a bus that might be full; if I had a guarantee that if I press the button and walk to the bus stop, I will get picked up in 5 minutes, it'd make me use the bus more)

Stage 4: As self-drive increase in general, cities put the thumb on parking spots. E.g. in parts of London you already won't get planning consent for housing with more than 1.5 parking spaces per living unit as a means to cap car ownership. Expect to see that gradually driven down, with the expectation that people will buy parking space for their self-driven car elsewhere and/or forgo having one. Driving down the limits on parking will allow for denser developments, making ride share options etc. even more viable.

Stage 5: Youth grow up without depending on their parents to drive them anywhere from the moment they are trusted to go by themselves.

Basically, I see it as a process where the convenience of apps to get you somewhere will keep increasing to the point where people will find themselves increasingly opting to check these apps first and find themselves needing a car less and less. Some transport apps are already combining route-finding with then offering to order an Uber for you.

Expect to see more of that making it less attractive to get a car over time.

Especially as youth get used to a greater flexibility and level of freedom using these type of apps before they can buy a car. Car ownership many places represents freedom from parents driving you around, but more and more teenagers can expect to be in situations were parents opt to order them a car instead of driving themselves.

Before long, a whole generation will experience car ownership as irrelevant to the ability of liberating their transport options from parental control.

Sure, some people will still opt to own one, but many already forgo car ownership, and that number will certainly rise rapidly.

The funny thing, some of this already exists these days. It is called ZipCar, Car2Go, or DriveNow. All of them suffer from the very same problem: service.

It does not matter if a car is selfdriving or not. To be recognized as available by someone means a utilization of about 30% by the providing company. But a utilization of about 30% does not drive down costs. You still of costs for producing and servicing the vehicle, which is way higher than just the energy costs.

Tesla has no experience in free floating vehilce fleets. Uber has no experience in such thing. ZipCar has. Car2Go has. Car2Go as an subsidary of Daimler even has experience in car production.

Now think of that.

ZipCar does not pick you up and drop you off at location. I can tell you as a Zip Car user, there are many issues that caused me to discontinue, none of which was service.

1 - the zip cars available to me are a few blocks away to walk which isn't convenient if I'm carrying something or lazy.

2 - high cost. Taking uber right now is so much cheaper. Imagine if cars were self-driving. Costs would plummet

3 - parking. This can be very hard depending on where I'm going

4 - I have to pay for the length of time I'm out, not for my ride. I also need to return the car when I'm done.

These are the biggest pain points in my opinion. Also, not needing to drive myself is a huge bonus. The leverage of solving these pain points are huge, imo.

zipcar and it's ilk have the same problems as bike-share programs: people don't use the service in such a way that the vehicles get to where the demand will be. And if you finish work at the end of the day and all the cars are gone, you aren't going to trust it again.

the bike shares solve this by driving trucks around, collecting bikes from areas where they aren't going to be needed and moving them to areas where demand is about to spike. it's hard to load up a truck full of cars to meet demand in the business district at 5pm, but it's very easy to send over all the spare capacity in your autonomous vehicle fleet.

That is it, and the ability to use the data to start offering sharing options with either discounts or priority pick up to make people ride share when capacity is full will be a massive change ("we're seeing high demand; you can wait 10 minutes for a vehicle to yourself, or a ride share can pick you up in 5 minutes"). There are plenty of times where I know it will take time to get a car because of demand, and would happily share if it meant I'd get picked up sooner.

Smoothing that out, so that you come to expect a car to be available very rapidly no matter what, with only minor inconveniences, may not end private car ownership but certainly will make a lot more people opt for alternatives - I know for myself (I don't own a car) the occasional lack of predictability in how soon I'll get picked up is the one aggravation that occasionally make me want one.

You missed the point like the GP. None of those services are anything like what's coming. Car2go still has to be parked, has to be close, you have to live in the home zone. Zip car is just... Not even close. It's a rental service and you need to take the car back.

Comparing either of those or public transit to self driving cars misses the point entirely.

The ridesharing systems (like uber pool and lyft line) are meant to alleviate that utilization problem. Zipcar has no experience with the complex logistics automation work for ridesharing that Uber and Lyft are doing, nor do they have a large enough userbase to make that ridesharing system viable and competitive.

As a zip car member, I wish it did all those things, but it is a far cry from the self-driving future GP is talking about. I'm also worried about tesla's market share, but more because I don't know if they can corner all the markets they are about to compete in: automotive manufacturing, batteries and set driving systems are all hard enough problems for an entire company. It's a lot to get right.

ZipCar is very useful in it's own regard. It's problem is that you need to pick it up and drop it off at the same location, and availability during peak hours.

Self driving car networks basically can solve that.

> I think there are a lot of pretty good reasons why the vehicle market is going to change significantly in the next 15 years

But not necessarily shrinking the market. The cost of mobility goes down, consumption will go up. How high? I'm going to guess absurdly high. The reason US cities are sprawling messes is that they were built as the cost of mobility dropped off a cliff (cars and trucks replaced horses). Go read Clifford Simak's "City". People will live in self driving cars and commute in their sleep. It's going to be nuts.

> Cost of ownership is especially high in cities (parking), and the market for ride sharing is highest. So probably many fewer city dwellers will own cars.

If it's really self-driving it can go park wherever. Not very good for the environment, perhaps, but it completely solves that problem.

The farther the car goes to park, the longer the wait for it to pick you up on demand. Not to mention, traveling a few blocks rarely saves much on parking in densely populated areas.

It can travel way further than "a few blocks" and I'd guess most trips are not so unexpected that it's impossible to accept a delay of 30 minutes or maybe even longer.

If your parking spot isn't included as part of the building where you live, then you're going to have to either buy some more property (however small or far away) to keep your car, or pay someone to let you park your car on theirs.

Yes, agreed, but if you live in a major city that's already the case (or you drive around for an hour looking for a space, something the self-driving car could also do).

Right. And so my point is the expense and inconvenience of parking in cities is incentive for you to either (a) not buy a car, and instead use an on-demand (self-driving) rideshare service, or (b) buy a car, and recoup costs at little inconvenience by putting it into a self-driving rideshare pool.

The end result is fewer cars owned by fewer people, but utilized much more fully, so the cars that do exist in cities spend more of their time driving and in-service, rather than sitting parked and empty for 98% of their lifetimes.

But a self-driving car that can either find street parking or else drive to some suburban lot makes that incentive much weaker, not stronger?

A service will be able to keep cars near you, and will have sufficiently higher utilisation of their cars to be willing to pay more for parking spots. Your waiting time to get a taxi service will likely drop substantially at the same time as your waiting time to get picked up by a car you own is likely to go up.

As someone who doesn't own a car, the only potential appeal to me of owning a car is shorter waits and predictability (always there). If the predictabiity and waits drop for rental services, my reasons for considering buying a car would rapidly drop. If the waits to use a car I own go up, my reasons for considering buying one would drop further.

As someone who doesn't own a car I don't think you have that much insight into the mindset of people who own a car.

You're assuming I've lived in a car free household all my life.