Traffic Wouldn’t Jam If Drivers Behaved Like Ants | Nautilus

Tom Vanderbilt (author of Traffic) writes about traffic again in “Traffic Wouldn’t Jam If Drivers Behaved Like Ants” for Nautilus, discussing Ant Colony Optimization and route choice. Our work gets a mention:

These “traces in the environment,” in a process generally referred to as “stigmergy,” help explain how simple agents, without a larger meta-ability to communicate, coordinate, or structure their transport networks, can display such “surprisingly intelligent collective behavior.” Humans are not so ruthlessly efficient—the University of Minnesota transportation researcher David Levinson found, in one study, that only 15 percent of daily commuters were on the shortest route possible.

Automation and the last land use (On First Space, Second Space, and Third Space)

In the beginning, people lived in caves. And then they built huts when they ran short of caves. But at first there were no developed land uses other than housing. Sure, land was used, it was used for hunting and it was used for gathering. But it wasn’t much used for manufacturing.

An historic house in Sydney

Eventually enough was gathered (and agriculture was developed) that warehouses of a sort were constructed to store foodstuffs for future use.

Enough was traded, that a blanket on the ground became a table became a tent became shops (retail land uses). Places of worship moved out of the living room into a common structure, or communal fire pits were enclosed. Manufactured goods moved from the home to a workshop to a factory. Bureaucracy itself required its own spaces.

The modern world was created.

Sociologists talk about First space (home), Second space (work or school), and Third space (social gathering spots). Some, such as Ray Oldenberg in the the Great, Good Place, bemoan the loss of third space in the US as First space has gotten larger and larger. There is less reason (and it is costlier) to go to a pub or a coffee shop when you have your own wet bar in the basement and your houses are spread out in a low density environment. I like my house. I may have felt cooped-up in a one-room apartment in the city.

Importantly, today’s large homes with information technologies are so much more entertaining, people don’t want to leave. The notion of free-range children is the exception, and this is not just because parents are overly protective [my neighbor returning my child home because he was playing standing guard outside her house across the street], but because children aren’t as bored in the home with their devices as they are outside the home without them. And when the neighbor kids stay at home, the outdoors is even less entertaining as there is no one else to play with. It is a vicious cycle with the natural end state that no one plays outside. With the internet and its billions of cat videos, plethora of video games, and any music they could want, along with instant virtual connectivity to anyone they might want to socialize with, there is no need to go out.

In recent years, we have seen obsolete land uses revert back to housing. We have seen conversion of old industrial lofts to housing, transformation of offices to housing. We are adding far more housing space as a share of new development then in previous decades. Retail spaces are being abandoned and repurposed.

The End of Traffic and the Future of Access: A Roadmap to the New Transport Landscape. By David M. Levinson and Kevin J. Krizek.
The End of Traffic and the Future of Access: A Roadmap to the New Transport Landscape. By David M. Levinson and Kevin J. Krizek.

Automation means that (a) manufacturing and warehousing have even fewer ties to the labor force than before, and so are even more footloose (to locations where the cost of robots and transportation of materials to the site and transportation to end users is lowest), (b) that transportation costs drop, and depend primarily on the value of time of the goods being moved, wear and tear on vehicles, and the costs of energy (which are falling), (c) agriculture and food production are are more like manufacturing than artisan craft-work. While manufacturing and food production will require space, they don’t necessarily require space inhabited by people.

To the extent that people work, they will less and less need to work any particular place. They may need to meet with others, but that need not be at one another’s cubicle. Second space will disappear. This may result in an increase in Third space as humans, at least the extroverts among you, and even introverts in extrovert mode, still want some face-to-face interaction. But what is to say that won’t occur in people’s large large homes? In a world with authenticated people identifiable from facial recognition by your smart home and with established identities and reputations, the threat of inviting “a stranger” into your home, or the public portion of your home may no longer be so frightening, as no one is really a stranger anymore. The greater threat will be meeting someone in a public place where you don’t control the environment.


In the end, just like the in the beginning, the only built structures people require is housing. Second space will diminish. Third space may or may not fill the gap.

Elements of Access: Roundabouts

1_vancouver roundabout.png

Most people I talk with about roundabouts probably fall somewhere on the love-hate spectrum between extreme dislike and hate.  One reason for such an unenthusiastic assessment – especially for Americans – can likely be traced to some common misconceptions about what we are actually talking about when we talk about roundabouts.

Having grown up near Boston, I spent countless Friday evenings on hot summer nights sitting in traffic with everyone else trying to get to Cape Cod.  The Cape Cod Canal cuts the area off from the rest of Massachusetts, so the only way across by car meant traversing one of the two bridges.  For years, both bridges also had what those from New England called a “rotary” on one side or the other.  Traffic would routinely back up for miles at both bridges, making what could have been a 75-minute drive considerably longer.  The culprit was often these multilane rotaries.  Such intersections proved to be inefficient – both in terms of traffic flow and land consumption – as well as dangerous.  You may also remember Clark Griswold getting stuck in a London traffic circle in the 1985 movie European Vacation. “Hey look kids, there’s Big Ben!  Parliament!” Homer Simpson was in a similar situation in a 2003 episode and nearly killed the Queen of England!  Examples like these end up giving all circular intersections a bad rap.  Not surprisingly, circular intersections were essentially phased out of most U.S. design toolboxes and the minds of many Americans.

2_sagamore rotary.png

Modern roundabouts, not developed until the 1960s, refer to something quite different.  For one thing, roundabouts are much smaller than the old rotary intersections.  Instead of outside diameters exceeding 300’ or 400’, modern single-lane roundabouts typically range between 90’ and 180’.  Another thing is that the cars entering must yield to the cars already in the roundabout (although this was usually, but not always, the case with the older traffic circles and rotaries).  The main defining characteristic of modern roundabouts, however, has to do with speed deflection.  Speed deflection refers to angle at which cars enter the roundabout.   With the old rotaries, there was little to no horizontal deflection of through traffic so cars could easily exceed 30 mph.  A well-designed modern roundabout typically has enough deflection in the angle of this approach to actively manage vehicle speeds to less than 20 mph.  It can also still handle truck traffic with design features such as a traversable apron that skirts the inner circle, which can be seen in the top image from Vancouver.

Elements of Access: Transport Planning for Engineers, Transport Engineering for Planners. By David M. Levinson, Wes Marshall, Kay Axhausen.
Elements of Access: Transport Planning for Engineers, Transport Engineering for Planners. By David M. Levinson, Wes Marshall, Kay Axhausen.

So what does the research tell us about modern roundabouts?  In most contexts, they move traffic more efficiently and are safer than conventional intersections.  Why would this be the case?  In terms of efficiency, there is no waiting for the light to turn green when there is no cross traffic.  In fact, single-lane roundabouts have been shown to reduce delays as compared to conventional intersections and effectively manage traffic flows as high as 25,000 cars per day.   Less idling also means fewer emissions.   In terms of safety, roundabouts eliminate conflict points and the most dangerous types of conventional intersection crashes; while you may get more sideswipe or rear-end crashes, such crashes are far less likely to be fatal or severe injury.  Also if the roundabout is designed with adequate deflection, these crashes tend to happen at slower speeds.  This reduces crash severity to the tune of 78-82% fewer serious injury or fatality crashes as compared to conventional intersections (AASHTO Highway Safety Manual).

There are valid concerns about pedestrians and bicyclists in roundabouts, but splitter islands, setback crosswalks, and sidewalks – when combined with slower vehicle speeds – help tremendously.  Interestingly, many places allow bicyclists to act as either a vehicle or a pedestrian in roundabouts.  My own concerns center more on effectively serving those with impaired vision, which is still an issue with most roundabouts.

While I would argue that multilane roundabouts are unnecessarily used in many situations where a one-lane roundabout would work well, multilane roundabouts still offer many of the same advantages.  Compared to single-lane roundabouts, however, they do: lose some speed deflection when volumes are low; introduce a new crash type to the mix (i.e. sideswipe crashes due to lane changes); and make things more difficult for pedestrians and bicyclists. You can also include neighborhood traffic circles – which are even smaller than most modern roundabouts – in this overall discussion of circular intersections.  The example below from Berkeley, CA combines 4-way stop control with a circular intersection.  While not quite a roundabout, it is a good example of using a small traffic circle to help manage speeds and improve safety.

3_berkeley traffic circle.png

Compared to signalized intersections, roundabouts are generally less expensive, more efficient, more environmentally friendly, and perhaps most importantly, safer.   Furthermore, you never have to worry about a power outage with a roundabout.  While there are legitimate reasons not to use roundabouts in some situations – such as highly unbalanced traffic flows or ROW limitations – many get eliminated as an option due to our cultural biases against them.  All we are saying is give roundabouts a chance.

What impacts will e-shopping and telecommuting have on our transportation future? | CTS Catalyst


Information and communication technologies (ICT) are changing how and where we work, shop, and live our lives by making virtual activities a viable alternative to traditional physical activities. The rapid growth of these technologies has important implications for the transportation system.

“Unsurprisingly, transportation policymakers and planners are hoping ICT will replace some activities that require travel between places and help alleviate the transportation challenges that many large metropolitan areas face,” says Jason Cao, an associate professor in the Humphrey School of Public Affairs. “However, the impact of ICT on activity participation and travel is more complex than it appears at first glance.”

In a project sponsored by the Minnesota Department of Transportation (MnDOT) and the Minnesota Local Road Research Board, U of M experts synthesized the effects telecommuting and teleshopping have on travel behavior and predicted their potential impacts on our transportation system in the decades to come.

Both e-shopping and telecommuting are growing, and this growth is expected to continue. In 2010, about 16 million employees worked at home at least once a month—an increase of 62 percent from 2005. Currently, e-shopping makes up more than 6 percent of total retail sales, and experts estimate it may eventually account for up to 50 percent. Other ICT such as telemedicine and telebanking are also likely to see growth.

Cao found that previous studies on the travel impacts of ICT show mixed results. Telecommuting offers the potential to reduce travel during peak hours, lower VMT slightly, and help mitigate freeway congestion in the short term, but it can increase non-work travel and influence the travel of other household members. It may also enable commuters to move farther away from their work and become auto-dependent.

For e-shopping, the literature shows that people who buy online also buy in person more often, so e-shopping may—or may not—reduce travel to stores in the short term. Online buying increases delivery traffic and freight transportation, Cao says, and if the share of e-shopping grows enough to change the distribution of commercial land uses in a region, it would have a profound effect on shopping-related travel.

“We believe that transportation planners should be realistic about the relationship between ICT and travel,” he says. “Although the short-term effect of ICT on travel may be substitution, leading to slight reductions in VMT, in the long term travel demand is likely to grow as ICT induces new travel.”

“This work clearly shows the complex relationship of information and communications technology with travel behavior and impacts on the system,” says Ken Buckeye, program manager with MnDOT’s Office of Financial Management. “As transportation planners and managers, we must be mindful that the short-term effects of these policies and trends may mask more significant long-term effects like the potential for increasing travel.”

Cao’s research is part of a multi-pronged study that analyzed the technological shifts altering surface transportation and the implications for Minnesota. Other contributors included Professor David Levinson (principal investigator) and assistant professor Adam Boies of the Department of Civil, Environmental and Geo- Engineering, and Humphrey School associate professor Yingling Fan. Their high-level white papers are compiled in a final report: The Transportation Futures Project: Planning for Technology Change.

Related Links

Reprinted from CTS Catalyst: What impacts will e-shopping and telecommuting have on our transportation future?

Funding transport in Minnesota

I was interviewed by the Heartland Institute about transportation funding in Minnesota. Being leery given their reputation, I asked for the questions in writing. I was right to be leery, they still managed to misattribute opinions to me in the posted article despite the interview being in writing. I have asked for a correction and will update if they provide one. Update June 17, 2016: Posted article is corrected.

These are the actual questions and my actual answers.

Fresno Gas Prices (Dec 2001)

On May 27, 2016, at 13:43, Jeff Reynolds <> wrote:

Hi David,

Thanks for taking a bit of time for my article. I have just a couple of basic questions.

1.       What are the pros and cons of this particular user fee that was proposed in MN?

Why we should raise gas taxes now:

• Road quality is not where we would like it and will only get worse if insufficient revenues are raised to maintain and reconstruct existing infrastructure, as existing roads continue to age and deteriorate.

• Raising gas taxes is more administratively efficient in the short run than implementing tolls or mileage fees, especially tolls requiring lots of infrastructure. It presently has lower collection costs. There is no guarantee that will remain true as new technologies change the cost of collection.

• Currently roads are only partially funded with gas taxes, general revenue is a major source. This charges non-users as well as users, and sends no signal about the appropriate amount of roads that should be built or how scarce road space should be allocated.

• So, states like Minnesota can (they are perfectly capable of, the price of gas tax is a small share of the price of gas, and a 5 or 10c increase is well within weekly volatility of gas prices) (and should for the reasons above) raise the gas tax before using property taxes or general revenue to pay for roads.

We do need to be careful that the money gets spent on maintaining the valuable parts of the existing system, not building wasteful new facilities. There is a serious lack of trust in existing institutions. Minnesota, like all states, has built roads for political reasons rather than because demand warranted it.

In contrast, the vehicle sales tax or property tax (in its various forms) favors travelers who drive more miles per vehicle (i.e. rural travelers), and provides no incentives to reduce demand. It encourages people to keep vehicles longer, which keeps older technology around longer (whether this is an environmental cost is unclear, as the environmental cost of building new vehicle is not small). Dedicating transportation-sourced sales tax to transportation expenditures is a hidden subsidy, since those funds are no longer able to pay for the general costs of government. Why should sales taxes on gasoline or motor oil only pay for roads, but sales taxes on everything else pay for schools and police?

2.       Is more revenue necessary, or would MN be better served to reprioritize its spending?

Minnesota, like all states should reprioritize its spending, favoring maintenance and preservation over new construction, but it should also reform where it gets its funds for transportation. There is no good reason that transportation should be subsidized by general funds dollars, which is true now for more than half of all state and local transport expenditures. Roads, like electricity and natural gas should be thought of, and managed like a utility, and the beneficiaries should pay proportional to use. This requires significant governance reforms, but would help reduce politically-driven investments.

3.       Why did the legislature reject these proposals?

Republicans are afraid of voting for gas tax increases. After the I-35w Bridge Collapse in 2007, a gas tax hike was brought to legislature in 2008.

The “Override Six” are Republicans who voted with the DFL (Minnesota Democratic Party) to override Gov. Pawlenty’s veto of the gas tax bill. Four of them lost their seats due to Primary challenges, while the Republicans lost two of those seats to the DFL in the 2008 General Election. This lead to the rule that voting in favor of a gas tax increase can be dangerous to your political health, if you are a Republican.

4.       What’s next for MN and its infrastructure?

The Governor and some are discussing a special session of the legislature, but in the absence of that, it will come up again next year, after the election, and the DFL might recapture both Houses and then in that event will likely pass some form of gas tax increase statewide and raise the transit sales tax in the Metro area to pay for transit capital expenditures for new LRT and Bus Rapid Transit lines.

— David

Untethered: How communities are physically drifting apart

In the dark age before electricity, great mills were located adjacent to waterfalls to provide direct energy. This is the origin story of many early industrial revolution cities including Minneapolis. The development of the electric grid, first DC, then AC, untethered milling from the falls.

St. Anthony Falls and the Mill City Museum

In the age before the streetcar, people lived within walking distance of their jobs. Downtown was very important. With the streetcar and subway, downtown remained important, as the destination of a radial commuting, shopping, or entertainment trip beginning farther out in suburbs. But with the automobile, not only residences, but first shops and then workplaces could become untethered from their downtown anchor at the head-end of the transit system. Downtowns in many US cities haven’t added employment in many decades, and almost all have lost market share.

Yet the individual’s daily activity pattern itself was still confined to a roughly 30 minute radius (sometimes referred to Marchetti’s constant, but identified by Zahavi and others earlier). This helped glue cities together.

With soon-to-be-deployed mobility technologies like Automated Vehicles (AVs)  this commuting budget range can expand. Thirty minutes of actively engaged traveling (driving, biking, walking) is not the same as thirty minutes as a disengaged passenger in a (commuter train or) autonomous vehicles.

With new and better telecommunications technologies — which admittedly people have been talking about this for ages, but which are steadily getting better — the requirement for in person meetings steadily drops. With fewer in-person meetings, there are fewer days per week that one needs to “go” to work, which means a weekly commuting budget may be a more appropriate concept than a daily one. It also means off-peak travel is more likely.  Non-work trips are likely to substitute in part for work trips, but they won’t be as long or as peaked. With less congestion in any case due to AVs and changes in demand patterns, roads are faster. All of which suggests even more decentralization of residences and less physical tethering between home and work.

Telecommunications also no longer requires wires, as wireless gets more efficient. So the need to be on the wired telecom network is not required.

Elements of Access: Transport Planning for Engineers, Transport Engineering for Planners. By David M. Levinson, Wes Marshall, Kay Axhausen.

Rooftop solar energy is increasingly becoming feasible. Without the need to attach to an electric grid (though maybe still wanting to due to load balancing – though with enough energy conversion efficiency this doesn’t matter), and with more power available on large rooftops where land is less scarce, score one more for decentralization.

Even in the absence of flying cars, though most certainly with it, the costs of living in isolation and physically unconnected from the city drop. The consequence is a greater trend toward decentralization. Sorry urbanists.

Now the urbanists will say, rightly, that cities are getting better too. While life for the disconnected may be improving, the quality of life for highly tethered urbanites might also be rising. Urban pollution will drop as renewable energy and electric vehicles become standard and social amenities will always be closer in terms of travel time.

Where anyone will live depends on their preferences and the opportunities available to them. The good news is advances in technology suggests more opportunities will be available. The bad news is your opportunities depend on the preferences of others. I can’t be alone in wanting to live a city of 100 million people (imagine the specialization in food, stores, and entertainment possible at that scale) and expect to be able to satisfy that want. I need more than 99 million of my closest friends to agree.

Posters for the A Line

The new MetroTransit A Line will take you to Rosedale, the State Fairgrounds, Hamline University, the new Soccer Stadium, Snelby, Macalester College, Highland Golf Course, Highland Village, and Minnehaha Falls. There should be posters to that effect around the city.

The general principle is of course true for all MetroTransit routes. Some posters from the London transit system below for inspiration.


For more on the A Line, see: A Review, and Accessibility Evaluation.

The Era of big Infrastructure is Over

In his 1996 State of the Union Address, then President Clinton said twice “The era of big Government is over.” Clearly it was not. While government spending ebbs and flows, big government continues to be a feature of American society.

A Political Economy of Access: Infrastructure, Networks, Cities, and Institutions by David M. Levinson and David A. King
A Political Economy of Access: Infrastructure, Networks, Cities, and Institutions by David M. Levinson and David A. King

If I were President I would claim the “Era of big (civil) Infrastructure is over” in the US. Not that we don’t have big infrastructure, we do, and it isn’t going anywhere soon. The size of the paved area in the US is on the order of the State of Virginia. That’s pretty big, and just looks at one measure of one infrastructure (admittedly a large one).

Once upon a time we did deploy big infrastructure. The railroads in the 19th century, and the interstate in the 20th were BIG. Turnpikes and canals were other large technical systems of the 19th century, as were the US Highway system, airports, container ports, and the like in the 20th. But they have been deployed, and many of them area already shrinking.

Instead, because the existing infrastructure systems are mature (built out), they need little expanding (and likely some contracting).

Certainly there are potential new infrastructure for surface transport.  The most widely discussed would be intercity High Speed Rail and urban transit projects. Similarly there are proposals for water (rebuilding the water and sewer networks) and for energy (massive investment in renewables as well as smart grid technologies).  I think the transport investments are unlikely, the water investments are mostly piecemeal replacements, and the energy investments will be a set of many small, decentralized power generators rather than large facilities. In short change is likely to incremental rather than comprehensive.

In part the question turns on what you mean by “Big”, and I think we mean system level deployments, like the interstate, or a national HSR network, and not individual segments that are adding to an existing system or replacing an existing system element in-kind with some added functionality.

On the transport side, there is little interest in large new systems. The last great window for high-speed rail was the 2009 Recovery Act, which did not achieve that aim. Seven years in and there is no actual high-speed service to show for it. Even if the California line were ever opened, we are decades away at the earliest from the onset of a national network in the US.

Proposals for new Interstates appear from time to time (like this for I-87:  [10 points if you knew that was Norfolk to Raleigh without looking it up]), and occasionally one actually opens, and even a second or rejuvenated Interstate 2.0 system has been proposed, but again there is no strong push for such a thing, and the advent of new technologies gives such proposals a ghost-like feel.

There are new systems emerging. The internet and wireless telecommunications are pretty important. Combine these with transport and we can construct an on-call ride-hailing system that has updated the traditional taxis. This may eventually become substantial with Autonomous Vehicles. But this latter element is not a conventional physical infrastructure investment (not much of one, some servers, some software), rather it redeploys existing (and soon new) vehicles in a useful way.

The new information-enabled systems that ride on-top of the classic physical layers are the products of Electrical Engineers and Computer Scientists, not great Civil Engineering works. We can imagine some things that might become useful. For instance we can think of  space civil engineering, things like Space Elevators and Dyson Spheres. But these are not on the near horizon.

Unless we can find an infrastructure that increases connectivity massively the way the railroad and the interstate did (doubling speeds, e.g.), there is no point in spending resources for that given the increasingly high costs and diminishing returns that civil infrastructure faces. We have enough trouble maintaining what we have with its proven connectivity (or lack thereof), the value of future infrastructure systems is speculative at best.

Urban Scaffolding: 6 transport technologies which will be largely removed in coming decades

Elements of Access: Transport Planning for Engineers, Transport Engineering for Planners. By David M. Levinson, Wes Marshall, Kay Axhausen.
Elements of Access: Transport Planning for Engineers, Transport Engineering for Planners. By David M. Levinson, Wes Marshall, Kay Axhausen.

Like buildings under construction, cities are built with scaffolding. Remove the scaffolding and cities remain. Yet what is “scaffolding” and what is “permanent” is not at all clear. Yesterday’s permanent structure is today’s scaffolding.

Take for instance the deployment of streetcars in the late 1800s and early 1900s. These streetcars enabled (not coincidentally) suburbs from which their customers, resident travelers, would use on a regular basis to commute to jobs and journey to shops. Yet in the mid-20th century these streetcars (urban scaffolding if you will) were removed (just as horsecars before them), and the city itself remained. Those streetcar suburbs still exist sans streetcars. That which enabled their construction and occupancy was eventually unnecessary and removed.

Ports were the reason d’etre for many cities, yet in today’s era of containerization, ports that failed to make the transition for whatever reason withered. The city that port enabled remains. These include such places like the City of San Francisco, New York City, and London, which today lack significant port operations, yet have maintained or gained in status. The port scaffolding was removed, and the rest of the city was self-sufficient without. However in the absence of an initial port, those cities may never have been more than hamlets.

The scaffolding enables the construction of an urban web of social, economic, and technological elements that eventually becomes thick and secure enough that the initial framework can disappear without taking the system down with it. But at the time, no one envisioned the port or streetcar as temporary. They seemed permanent.

So what today seems permanent but is simply the scaffolding for tomorrow’s city?

If I were to offer speculations …

  1. Start with the at grade (or elevated) urban interstate. Without the urban interstate, today’s cities would look much different. Many would argue better, but no one would dispute different. It is not that there will be no surface passenger transport goods movement, but that it will look very different in 50 years (30 years) than it does today. For freight, we will be in a world with self-driving delivery vehicles and aerial drones. Through traffic could always avoid the central city, to traffic can be carried on surface streets as needed. Vehicles are considered a nuisance if they are seen and heard, even with pollution levels down to almost zero and vehicles much safer. Roads which carry those vehicles enable the nuisance. These roads will be bridged over (land for driverless (and especially passengerless) vehicles is too valuable to expose it to the sky), or rebuilt as high capacity tunnels under the city. Construction of course is costly (and disruptive), but with advances in tunneling technology, further automation, and the right economic model, this will be justified in larger areas.
  2. Move on to traffic lights. Without traffic lights, we never would have managed to maintain the urban auto-mobility we do have, it was critical scaffolding for the twentieth century auto-dominated city. In tomorrow’s world, they will preserved in selected districts only for their historic authenticity, not as actual control devices. Traffic will continue to need to be controlled in places, but that control can be conducted invisibly through radio spectrum, rather than with colored lights designed for the human driver’s eyes. The difficulty will be dealing with pedestrians in mixed, congested environments (if it’s not congested, pedestrians will just have right-of-way, if it’s not mixed, there isn’t a problem). More discrete in-ground sensors will detect and guide pedestrians across busy intersections. Most downtowns will be largely pedestrianized, so these locations will be at the edge of the core.
  3. We will probably kill off most of the remaining mass transit in all but the largest cities. Large lumbering vehicles showing up every 10, 20, or 60 minutes provide a poor level of service compared with the equally if not more cost-effective shared single passenger AV, which will be nearly on-demand and will be point-to-point. Only the highest frequency routes serving the most passengers (think urban subways or the El) will remain.
  4. We will certainly kill off the remaining gas stations with the transition to electric vehicles. While trucks may switch to LNG for a period, cars will move to EV status rapidly over the next two decades. EVs powered by renewables will be cheaper than gasoline-fueled cars even with low gas prices. The gas station sites will be converted, as many already have been to shops and convenience stores. Other auto-related uses, like dealerships and service stations, will also become scarcer and scarcer. Land is too pricey in the central city for these to be visible, especially when cars can get themselves serviced, or drive out to you for a test drive, or are shared in the first place.
  5. We will eliminate most urban parking lots and garages, both because cars can drive farther away to park, and because shared vehicles can be in motion for much more of the day. Again space is too valuable.
  6. We will eliminate most larger cars and passenger trucks. People will be able to get the larger vehicles on-demand, which turns out to be not that often for most people, and will use the right size car for the trip. While most vehicle forms will remain in the population, the share of single person vehicles will grow enormously, and the larger vehicles will shrink as people choose to save money once they realize the opportunity cost of not having the large vehicle for the occasionally.

It’s not that cities will be unrecognizable, it’s just that the parts of the cities we like will become more and more prominent, and the parts we hate will tend to shrink in size. Plenty of twentieth century changes (skyscrapers, skyways) will remain part of the landscape, and the huge expansion seen by the twentieth century will largely remain in place so long as urban populations continue to grow.