Anytime you make something cheaper, easier to use, more useful and/or more efficient you get an increase in usage. Autonomous electric cars are all 4, and will lead to a dramatic increase in the number of miles driven. Some of those miles we can imagine (for instance sending your car on a dead-head journey somewhere for somebody else to use it), but most of the increase will probably be from new usage models that we don't expect.
There may be a decrease in the number of cars; the increase in car sharing may be higher than the increase in miles, but I doubt it.
But congestion? That's going to get a lot worse, it's not going to be the utopia that people here seem to expect.
I worry that self-driving cars are going to be used as an excuse not to expand public transit, especially in the Bay Area. The problem is scale -- once the population reaches a certain size, the roads are going to be gridlocked even if every car is self driving. I'm not sure if the Bay Area is at that point yet, but NYC certainly is, as are most of the major cities of the world.
I worry that self-driving cars are going to be used as an excuse not to expand public transit, especially in the Bay Area.
By reducing people's sunk cost in cars, it may well free up the public mind to alternative forms of transportation. A lot of political inertia comes in the form of personal sunk costs. Also, self driving vehicles could themselves be used to expand public transit. IIRC, In Portland in the 90's, there was this thing called the "Custom Bus" where an late-night adapted bus route could be formed by people phoning in a request for a pickup within 2 miles of a normal bus route. Automated vehicles and networked computers could let us do an even better job of adaptive transportation.
I love this conversation. The US new auto sales run rate is ~17 million vehicles/year, with an average price right around $35k. That's $595,000,000,000 per year in capital being allocated to automotive mobility alone. Think what better things we could do with that capital (or everyone could work far less!) if we only spent a fraction of that on automotive capital costs.
You may love that conversation, but how many Americans are willing to give up their vehicles?
In America, the privately owned vehicle isn't just transportation, it's a status symbol. That's not going to change just because someone crunched some numbers.
> but how many Americans are willing to give up their vehicles?
That's not the question. Can you afford your own car if the cost efficiencies are no longer there because there are so many fewer cars on the road?
How many people own their own Cessna versus buying tickets on Southwest? Even with a private pilot license, I would never consider owning a plane myself. Too expensive!
I have a hard time believing this will be as big an issue as many believe.
The biggest reason is generational die off.
We're already seeing younger people forgoing even getting a license. Urban population growth is rendering it less valuable.
The only cohort that licensing is up is boomers. 55-75 are more likely to have a license than they were a generation ago.
Licensed 16 year drivers is way down. Something like half the number are licensed versus a generation ago.
Some of it is "blamed" on cost of owning a car. Those that can afford it but don't buy say they just don't need it. Rides are readily available; friends, Uber/Lyft, and pub trans are enough for a lot of folks as urban density increases.
I'd post references but this is easily Googled from multiple sources. And I'm on a slow mobile connection right now.
Younger folks tend to see cars as noisey, smelly, expensive, and passé already. Not sure that status symbol thesis holds up
> Younger folks tend to see cars as noisey, smelly, expensive, and passé already. Not sure that status symbol thesis holds up
That might be some of it, but Millenials on average are also less well off than older cohorts; they simply can't afford a new car, or possibly even a car at all (but might require one; not everyone lives in a metro with good public transit).
Shoe leather wasn't cheap in the middle ages, and leather soles used to wear out all of the time. Transport is something people are willing to pay for.
I haven't checked the numbers but it doesn't surprise me. A car is the most expensive or second most expensive asset that most people own. (And/or they spend a lot of money on taxi/Uber/Lyft/etc.) And most people spend a decent chunk of their days getting from Point A to Point B. Only some of that is commuting.
The cost of not being a subsistence farmer, I suppose. (Somewhat joking, but a big part of what makes modern civilization possible is mobility and a lot of that requires cars. In some situations, they're less important but they're still common pretty much anywhere.)
Not expanding public transit already happens. Google/Cisco/Apple/Facebook was running its own shadow public transit system for free to the cities, but the anti-techie lobbies demanded that SF regulate it to a shadow of its former self and tax the tech companies every time a bus stops to pick up passengers.
In Seattle, Google just gives its employees a bus pass, because their public transit system isn't full and dysfunctional.
In Seattle Microsoft also runs its own substantial bus service, "the Connector". There are so many Microsoft employees commuting in and out of Redmond every day that it makes a lot of sense.
Hopefully Uber/Lyft start a van/bulk transit service. This would help reduce traffic by combining trips while being more efficient than buses because the server can dynamically route the driver to optimize pickups and departures.
Perhaps even adding discounts for riders willing to trek the last mile on foot.
Then why isn't that being done today? As you increase the number of passengers, the cost of the driver becomes relatively trivial. It seems as if this were a service people were willing to pay for these money-losing companies are missing a significant market opportunity. I know there's UberPool but as far as I can tell it isn't a wildly popular option.
Routing. Turning an efficient set of start > points to a combined path is a complex traveling salesman problem. But this is worse as you need to find multiple such sets from raw data. It's not hard to get some overlap but if you go past 2-3 passengers everyone gets a longer trip.
That's probably true. Plus a lot of people just don't want to share vehicles. At some point, you're basically creating most of the disadvantages of a bus in a smaller vehicle.
Congestion? I don't think you're taking into account the huge advantage that autonomous cars have with respect to coordination among themselves. Much, if not most, congestion is caused by human drivers attempting to accrue advantage to themselves through maneuvers that disadvantage everybody else on the road.
No most of congestion is caused by the spacial inefficiency of car based infrastructure. Cars are big and roads have fixed capacity. Autonomous cars will not significantly change this.
Right we all expect this sort of thing, but so far it's all theory and I haven't seen anyone seriously test and study the idea based on real world factors.
For example I doubt we'll see any benefits of decreased safe separation in cities where there are cyclists and pedestrians and other factors that are hard to predict.
My prediction would be that the efficiency gains will be marginal and will be dominated by other factors that affect congestion.
> For example I doubt we'll see any benefits of decreased safe separation in cities where there are cyclists and pedestrians and other factors that are hard to predict.
Well, yeah, that applies to roads with exclusive use of autonomous vehicles, which clearly doesn't apply to roads shared with cyclists and pedestrians, but limited access freeways. But even within cities those can serve as arterial routes.
My amateur expectation is that if 100% of vehicles are autonomous, they can be FAR better drivers and you could get far more cars on the same roads with considerably less stopping and waiting.
The problem will be when fewer than 100% of vehicles are autonomous, then it will simply remain as awful as it is today.
I wonder if road signs will evolve too. The other day a few red lights were put in blinking mode because of roadwork. While asking for way more attention, it also avoid many 50sec pauses.. I found it very very nice. With collaborating SDV you could get rid of many stuck intersections that creates bloat and useless consumption (even though EV can just cut power on stop ...)
Certainly many slowdowns on highways and major roads stem from awkward, inefficient merging traffic - I can imagine autonomous cars would come together like the teeth of a perfectly smooth zipper.
Yeah I don't understand this enthusiasm for autonomous cars. It seems like the HN-crowd seems to wax enthusiasm for solutions that don't yet exist, and absolutely reject alternatives (bikes, public transit) that already exist and work. With new solutions comes new problems, and I can see congestion being a big one.
I guess it's not sexy to implement the tried-and-true, but it's a lot more sexy to dream.
They are not ready to be widely deployed but are close, which is why we're talking about them. It's an exciting technology with great potential impact on the future and a lot of that impact is positive.
As far as bikes go: no one is complaining about them. It's all in your mind.
As far as public transit: autonomous cars will improve it. By my calculations, autonomous car will be cheaper, unsubsidized, than a $2.50 bus ride subsidized by the city from our taxes (SF has ~$1 billion budget to fund money-loosing bart, buses etc.).
And that's cars. Things will be even cheaper if we get buses to be autonomous. At which point it's likely bus service will be cheaper and provided by a commercial company (the only reason public transport is government owned is because it looses money, so it wouldn't exist as private company).
> They are not ready to be widely deployed but are close, which is why we're talking about them. It's an exciting technology with great potential impact on the future and a lot of that impact is positive.
If you've ever implemented anything in your life, you should know that completely new solutions often come with unexpected failure modes. Very well thought out solutions often address all of these, but more often than not we have to address failure modes through iteration. Unfortunately, physical infrastructure isn't like software: iteration isn't nearly as cheap or fast.
> As far as public transit: autonomous cars will improve it. By my calculations, autonomous car will be cheaper, unsubsidized, than a $2.50 bus ride subsidized by the city from our taxes (SF has ~$1 billion budget to fund money-loosing bart, buses etc.).
And how will that scale? When SF's streets are filled grid-to-grid with vehicles that can only fit 5 people at once (and will probably only accept 4 due to customer sentiment), what will we do then? Also, care to share your calculations?
> And that's cars. Things will be even cheaper if we get buses to be autonomous. At which point it's likely bus service will be cheaper and provided by a commercial company (the only reason public transport is government owned is because it looses money, so it wouldn't exist as private company).
A bus driver is far from the main cost of a bus service. Many buses in SF/Berkeley are already electric or hybrid, so it's not even fuel anymore. Buses are expensive because they are large vehicles and because maintaining them is expensive. Autonomous driving tech won't make a bus cheaper to purchase or maintain.
I'd argue that the most useful aspect of driverless cars is going to be as short-range taxis. I live in San Diego. We're not LA but we're not tiny. I'd take the trolley basically everywhere if I could but that would increase my ride time by somewhere around an hour and a half. The VAST majority of that time isn't waiting for the trolley. It's sitting on a bus and waiting for it to take me to the trolley station which is only two miles away from where I live. If I drive to the trolley station, my A-to-B ride time becomes about fourty minutes total. If I could just buy a bus pass that included autonomous pickup that shuttled me quickly to the trolley station then I would give up my car in a heartbeat.
it's not going to be the utopia that people here seem to expect
Look for the emergent problems. It's the insightful thinker who can observe cars then predict strip malls and parking lots. It's the emergent problems which will point the way to new disruptive business opportunities.
It may be Econ 101 but a lot of people (aka voters) tend not to like fee schemes that make it blindingly obvious that the wealthy are effectively allowed to use public infrastructure in ways that they are not.
We already subsidize roads over public transportation so the subsidies to the wealthy are already there. We're just attempting to internalize externalities of a resource with physical limits.
The convenience of a public transportation systems is priced into the housing near its stations, meaning it mostly benefits the wealthy. The road network, on the other hand, reaches into neighborhoods which are far enough away from the center to be affordable to working-class families.
The people who end up driving despite the tolls, parking elimination, etc. are those who can't afford the rent near the public transit system. Cars are a significant cost for them, but not as much as the rent premium they'd have to pay for a well-connected location.
Washington state does this on the lanes that are otherwise HOV on I-405, which runs N/S through Bellevue (Expedia, Microsoft) and meets up with 520 E/W to Redmond (Microsoft). The tolls are proportional to demand, and they often hit the upper max of $10 each way. [1]
It's astonishing how much some people are willing to pay for the pleasure of driving free of traffic... We should charge them as much as they are willing to pay...
Thank you for pointing this out, it seems like it's often forgotten. We're probably about to make all the same mistakes we did when cars first showed up in cities. Of course, if we actually get coordinated driving that allows long-distance 100+ mph runs mere feet apart, the city will expand to cover... well, even more of what little remains of wild habitat. We're already urbanizing an area the size of Britain every year. Why pay high rent to be close to work when you can have a villa in what used to be Big Sur and nap during your hour commute to work?
Pity the motorcyclists, walkers, and cyclists who dare to get in your way.
Predicting future human behavior is hard, but I disagree (under the assumption that self-driving cars will be as good as perfect)
Firstly, I disagree about your congestion claim. It is 'easy' to program self-driving cars to create road trains, where cars ride at speed with centimeters distance. That way, road capacity can easily quadruple. Traffic may slow down in rush hour, but traffic jams due to driver misjudgments could become a lot rarer, if not a thing from the past.
Traffic jams due to cars breaking down also might be avoidable. Cars could be programmed to push other cars out of the way, if one were to break down.
I also disagree with your claim the number of cars will go down. Reason? For many people, cars are too much of a status symbol. They also serve a bit as an extension to one's home (both psychologically and practically, for example to store a child's seat, umbrella, tennis rackets).
So why, if everybody able to afford a car today will still be able to afford one, would they do away with it?
Also, I don't see lots and lots of possibilities of decreasing the peak demand for cars. Many people still will want to go to work between 7 and 9 (all else being equal), and most cars will not be able to make two trips between suburbs and offices in a single rush hour (even without congestion, as I expect people to move further out of town in response, where gardens are larger and houses cheaper). Also, parents who can afford it will have third or fourth cars to drive their kids to school.
> Anytime you make something cheaper, easier to use, more useful and/or more efficient you get an increase in usage.
This isn't inherently true. You only get an increase in usage if people want/need that usage but previously couldn't have it. (i.e, if there was already a shortage -- if we were severely undeserving the need previously, or a new undeserved need was discovered).
If there is no shortage already and/or no newly-discovered need for something, then making something cheaper or easier does not increase it's usage.
(This is the big common mistake nearly everyone makes when talking about "Induced Demand". If you don't understand that Induced Demand is really about critical shortages further up the chain, then you trap yourself into doing nothing under the false belief that any new capacity added will automatically "induce" a failure of the new system).
That's certainly true but I have to believe that, for most people, effectively having a private driver at their beck and call would increase the amount of time they spend in a car.
You're right that it's not infinite; I don't want to spend 8 hours a day commuting just because someone else is doing the driving. But, speaking as someone who lives about an hour outside of the nearest large city, I would absolutely go in more for the evening--as well as take more weekend trips--if I didn't have to drive.
That's likely typical of a lot of people, albeit not everyone.
Could you expand on "critical shortages up the chain?"
It seems like if slightly less-preferable alternatives are already being used, this would be enough to induce demand when the price drops a bit. Many car trips are optional; consider shopping and the ability to combine trips or order online, if you're motivated to do so. Also, there are many alternatives for entertainment, whether by going somewhere else or staying home. And people do take traffic into account when choosing when and where to drive.
A good example is to focus on things that aren't cars first.
For instance, no one says "we shouldn't make the air cleaner. If clean air became too cheap or too easy to breathe, people might breathe too much". Such a thought is ridiculous. It's understood that, once we have enough so that there is no longer any shortage for anyone, demand simply won't rise, no matter how much cheaper or easier to use that becomes. To the extreme that it's infinitely easy and has absolutely no cost for me to breathe extra clean air right now, but I still don't over-breathe because it would be pointless, there's nothing to gain by doing that.
Another example is a hospital. No one says "we should never build a new hospital. If healthcare becomes too cheap or too easy to get, people might become overly-healthy." Or "If we build a new hospital, we'll just encourage people to be more sick". Someone will even print up some misleading statistics. They'll say, "see, last year we had a 500-bed hospital, and it was always full of sick people. Now we have two 500-bed hospitals, and they're both full. Sickness has doubled since last year! We have more sick people now than ever before". It's understood that a new hospital does not create sickness, it does not induce sickness. These people were still sick, they just weren't counted because they weren't in the hospital. They were getting ill at home or at work, and not getting recorded at the hospital because it was already full.
---
Transportation works the same way. But it's been so under-built, poorly-built, and under-serviced for so many years, that everyone approaches it with starvation mentality to scrape by with, instead of approaching it as a problem to be actually solved well.
Cheaper and more convenient cars do not create demand, they expose shortages up the chain. There is already, today, many people who have a shortage of housing and transportation across the US. Clean/cheap/safer cars allow them to meet needs that previously were unmet. This will create the illusion of new congestion (because transportation will be more accessible to people), but that congestion is not new, it already existed. It just was not being not measured, it was spilling out into other areas where it is unaccounted for (such as people who ordering online who would have preferred to shop in stores but couldn't, or people just giving up on life altogether and not going places they want or need to go, because no transportation exists to make the trip possible).
For example, Autonomous cars will likely increase total car trips because empty cars will be summoned around to places. That's not really new demand though, that's an exposing of a critical shortage of parking. If Autonomous cars were cheap enough, and parking was cheap enough, Autonomous cars would never need to make empty trips. They could wait for people right where people already are without any problems. But because most cities have a critical shortage of parking, that shortage will spill out into the streets as "induced demand", where cars will waste lots of time and energy driving around empty into and out of cities, just so they can avoid parking in cities. Autonomous cars "solve parking" by moving parked cars into city streets and circulating them. It's not new demand, its old demand being pushed around to different places, hot-potato style.
Similarly, autonomous cars will likely increase total driving distances, as they make it easier for people to live further away from cities and endure longer commutes. But this too isn't really induced demand for cars, this is an exposing of the critical shortage of housing that most US cities have. If there was enough quality affordable housing in cities, people wouldn't waste their time commuting further away, even if doing so had zero cost. But because we are not willing to solve that problem, it too will spill out into our transportation infrastructure, and then get blamed on "cars" or "induced demand", when neither is a really accurate representation of the problem, nor an accurate path for finding a solution.
There's an assumption with induced demand that all the extra trips are mostly unnecessary, because there are slightly-less-convenient alternatives. It sounds like you're making the opposite assumption that they're mostly necessary: "critical shortage of parking".
It seems like this isn't likely to be settled via armchair reasoning in an online discussion; someone needs to look at the data (for example, at pricing), and that's likely to lead to a complicated economics discussion.
> But congestion? That's going to get a lot worse, it's not going to be the utopia that people here seem to expect.
Some congestion, but other congestion will get better. Parking shortage is a form of congestion that will get better when cars can drop you off rather than needing to find a parking space. Also, circling and searching for parking when there's little available causes congestion. There's few more frustrating driving experiences than being stuck behind someone going 10mph while looking for parking.
I can imagine one of the initial big selling points to self-driving cars will be a "go earn money" mode where, instead of parking your car, you tell it what time to come back and pick you up and it goes to drive for Uber/Lyft in the interim and accepts any ride that doesn't preclude it from returning in time.
Global adoption of autonomous vehicles is quite far fetched, ofttimes i suspect when people talk of its widespread usage they're referring to developed economies.
Maps and GPS triangulation or pinpointing operates on a margin of error and sometimes routes you to a wrong location even in New York
I’m not sure but I believe the batteries are still too expensive. Also, we might need solid state batteries before we have acceptable range and charging speed for many people.
Finally, the US federal subsidy, Norway’s subsidy, etc have helped the recent models. However, the US subsidy will disappear for Tesla, for example, shortly after they reach the 200,000 car limit.
Definitely not yet. You can buy a new Nissan Versa for under $13,000. $22,000 under the Tesla Model 3 buys a lot of gasoline. Of course the Tesla is a nicer car in many ways, and a "fair" comparison would look at more similar cars, but the relative cheapness of the ICE drivetrain lets traditional cars go much more downmarket than EVs can for now.
Battery costs are dropping steadily and it seems like it won't be too much longer before TCO for an EV is lower than even a cheap gas car, but we're not there yet.
Autonomous cars may dramatically expand the set of people that can drive cars, which will accordingly increase the amount of cars on the road.
Right now the set of potential drivers are limited to 16 to around 80 (though many seniors may be forced to stop earlier). When skill and attention and licensing are no longer requirements for driving, then children and seniors that would have been forced to bicycle, walk or take transit will be able to call an autonomous car.
At the moment electrics have lower marginal cost but much higher capital cost. That implies the average miles per car has to go up in order for them to make sense ...
The capability of "autonomous" is also overstated - at the moment there aren't the Level 6 autonomous cars that would be capable of dead-head journeys.
There may be a decrease in the number of cars; the increase in car sharing may be higher than the increase in miles, but I doubt it.
But congestion? That's going to get a lot worse, it's not going to be the utopia that people here seem to expect.