How Self-Driving Cars will Destroy Cities (and what to do about it)

Support Not Just Bikes by signing up to Nebula: http://go.nebula.tv/notjustbikes

Watch this video ad-free and sponsor-free on Nebula:
https://nebula.tv/videos/notjustbikes-how-selfdriving-cars-will-destroy-cities

Patreon: https://patreon.com/notjustbikes
Mastodon: @notjustbikes@notjustbikes.com
NJB Live (my live-streaming channel): https://youtube.com/@njblive

—
Related Videos

Stroads are Ugly, Expensive, and Dangerous (and they’re everywhere) [ST05]
https://nebula.tv/videos/not-just-bikes-the-ugly-dangerous-and-inefficient-stroads-found-all-over-the-us-canada-st05

More Lanes are (Still) a Bad Thing
https://nebula.tv/videos/notjustbikes-more-lanes-are-still-a-bad-thing

Cities Aren’t Loud: Cars Are Loud
https://nebula.tv/videos/not-just-bikes-cities-aren-t-loud-cars-are-loud

—
References & Further Reading

Safe Street Rebel – Not without a fight
https://www.safestreetrebel.com/

Person dies after 2 driverless cars block ambulance on way to hospital
https://www.scrippsnews.com/us-news/person-dies-after-2-driverless-cars-block-ambulance-on-way-to-hospital

Explore: See the 55 reports — so far — of robot cars interfering with SF fire dept.

Explore: See the 55 reports — so far — of robot cars interfering with S.F. fire dept.

What is Vision Zero?
https://visionzeronetwork.org/about/what-is-vision-zero/

Uber Takes Us for a Ride
https://www.powerswitchaction.org/resources/uber-takes-us-for-a-ride

Robotaxis Won’t Get Us There, So Let’s Stop Being Used to Sell a Future that Doesn’t Serve Us

Robotaxis Won’t Get Us There, So Let’s Stop Being Used to Sell a Future that Doesn’t Serve Us

Segregation by Design
https://www.segregationbydesign.com/

cars.destroyed.our.cities
https://www.instagram.com/cars.destroyed.our.cities/

Bamberg merchants, city officials say downtown open for business
https://thetandd.com/news/bamberg-merchants-city-officials-say-downtown-open-for-business/article_fec5d451-61bd-5fe2-904c-bdfecdb19392.html

Historical photos of Utrecht
https://hetutrechtsarchief.nl/

Historical photos of Fake London
https://search.ourontario.ca/
https://images.ourontario.ca/london/search

De Markt in gebruik als parkeerplaats voor auto’s.
https://hdl.handle.net/21.12115/NL-DtAD211303220

The number of references far exceeds the maximum length that YouTube allows in descriptions, but you can access the full list of references on Nebula or at this link:
https://notjustbikes.com/references/selfdrivingcars.txt

This video was made with video filmed on location by Not Just Bikes, stock footage from Getty Images, and specialty scenes filmed by videographers who were hired for this project. The 3D scenes were created by a professional 3D artist.

No generative AI or AI voices were used in the making of this video

—
Chapters

0:00 Intro
1:17 Reasons for optimism & skepticism
2:33 Moving fast and breaking people
8:50 Fatal Uber Crash
10:34 The Real Road Safety Issue
14:27 Promoting car-centric cities
17:03 Cheaper taxis (yay?)
21:07 Traffic congestion will be even worse
25:19 The promised future (past and present)
28:04 How AVs will destroy cities
29:56 Eliminating public transit
32:37 Consuming all streets
34:02 Eliminating pedestrians
36:05 Eliminating speed limits
38:05 Pollution and noise
39:41 Eliminating traffic lights
42:16 Do we actually need AVs?
43:28 Utrecht vs Fake London
48:33 What should we do about it?
51:53 Where to learn more & Nebula

A lot has been said about the 
future of self-driving cars.  Some people take a very negative position, 
claiming that autonomous vehicles are a pipe dream and will never become a reality.
They’re just another example of the   overhyped nonsense that’s typical 
of tech companies these days. And that may be true, but 
in several US cities today,  you can literally open an 
app, select your destination. And a car with nobody in it,
will pick you up.  And drive you there, automatically. 
So self-driving cars are here today. Which leads to the other kind of commentary often 
I see online. That self-driving cars are real. They will be coming to more cities soon.
And they will solve all of   our transportation problems.
They’ll make the roads safer. They’ll eliminate traffic congestion. They’ll be environmentally 
friendly They’ll provide mobility to everyone  And our cities will be better than ever.
I actually disagree with both of these positions, which is why I wanted to make this video.
Because I do believe that self-driving   cars are real and they will be 
in every city in the near future. The technology is getting 
getting better every day, and,   if these improvements continue on this track,
they will fundamentally destroy the fabric of our cities.
[NJB Intro] There is a cautious optimism among 
many urbanists about self-driving cars.  After all, the majority of problems 
within cities are caused by cars: traffic, pollution, noise, dangerous streets …
And there never seems to be enough space   available for sidewalks, bike lanes, parks, 
or anything else because it’s all been turned into wide roads and parking lots.
Anybody who has tried to get around   without a car in a car-centric city knows 
that human drivers can be really scary. Especially when they’re speeding or driving 
distracted, which happens way too often. So having cars driven by computers can’t 
possibly be worse than what we have now, right? And it is possible that, with the right incentives 
and regulations, that could be our future. But given the course we’re on, I believe 
that the future is increasingly unlikely. Especially when you consider the history of how 
automobiles destroyed cities in the 20th century, but also the effects of money and capital.
It’s estimated that over $160 billion dollars has already been invested into the 
development of autonomous vehicles.  So there is a LOT of money riding on this 
technology, and therefore a lot of pressure to get self-driving cars to market as soon as possible.
When people see these vehicles cruising around the streets of San Francisco, they might believe 
that this technology is right around the corner.  You can go there right now and 
see them dropping off passengers. They follow the speed limit, 
and appear to drive cautiously.  And it’s amazing to see that there’s nobody 
in the driver’s seat. But that doesn’t mean they’re always driven by computer.
As of last November, Cruise confirmed   that their robotaxis required human 
assistance every four to five miles, I don’t know how often this is needed today, 
but you don’t have to watch these cars for very   long to see that humans definitely still 
need to take remote control sometimes. This stupid robotaxi got stuck behind a 
construction toilet trailer, until someone   in the call centre could steer it around.
This one sat here for several light cycles, because it didn’t think it could 
turn right. The drivers behind it   eventually gave up waiting and just drove around.
Now, you might be wondering about these “remote assistant agents” who take control. Who are these 
people? Where are they located? How much are they paid? What are the working conditions like?
But nevermind all that,   it’s a technological miracle, OK?
Just shut up and get in the car. Now of course, this technology will inevitably get 
better. But nobody knows how long it will take to sort out the long-tail of rare occurrences that 
a self-driving car will encounter in a busy city. And as any experienced project 
manager will tell you, the first 90%   of a project takes 90% of the time.
And the last 10 percent of the project, will take the other 90%.
The technology is advanced, but fundamentally, self-driving cars are stupid. They’re 
mindless. They make a bunch of dumb mistakes. Recently in this parking lot for self-driving 
cars, Waymo vehicles started driving in   circles and randomly honking at each other.
This happened for days, and this particular video was taken at 4AM. The people who 
lived next door were not impressed. There have been dozens of official 
reports of driverless cars getting   in the way of rescue operations,
From blocking ambulances to crashing into fire trucks. But it gets worse.
Last October, a woman was hit by a car in San Francisco, driven by a human driver, 
and knocked into the path of a Cruise robotaxi. Cruise said the vehicle, quote, “braked 
aggressively to minimize impact,”  And while the robotaxi couldn’t stop 
fast enough to avoid hitting the woman, it still reacted faster than any human could.
So that’s a win for self-driving cars, right? Except that the woman then 
fell under the car, so as   far as the robotaxi was concerned, she was gone.
So it continued to drive, dragging her underneath! A key detail that Cruise initially tried 
to hide from the press and regulators. [It’s a first for police and fire 
departments involving a driverless car]  [The victim was found to be under 
the left rear wheel of the vehicle] Thankfully the woman survived, and Cruise 
cars were taken off the roads, but this is   a scary example of how these cars will make new 
mistakes that we haven’t even considered yet. They are not intelligent.
They don’t even have the   object permanence of a toddler.
But that hasn’t stopped them from being beta tested on public roads, and the 
citizens of these cities did not agree to this. Several groups in California have protested 
against being the involuntary test subjects of   multi-billion dollar tech companies.
And a Waymo car was even set on fire earlier this year.
Self-driving cars have caused   a lot of problems in San Francisco
[Waymo cars in San Francisco] [reportedly caused a big traffic jam]
[just as baseball fans were leaving]  [Oracle Park Monday afternoon]
[A KTV viewer shared this video] [you’re looking at showing]
[two self-driving cars stopped]  [in the middle of the Embarcadero]
The advocacy group Safe Street Rebel has published video evidence of 
self-driving cars running red lights,  cutting off drivers
cutting off pedestrians clogging up intersections
blocking streets  blocking public transit
parking in bike lanes and … uh … what the hell is 
this car doing to that old woman?!  So the idea that these cars will always 
obey the rules of the road is a fantasy. Safe Street Rebel also organises events like 
the Week of Cone. Since self-driving cars are programmed to avoid traffic cones at all 
costs. One of the best ways to incapacitate   one is to put a traffic cone on its hood. 
This is called “coning”, and it’s fucking hilarious. I’m sure these megacorporations 
will lobby to make this a criminal offence, so do it now while you still can.
Safe Street Rebel does a lot of great These protests may slow adoption, but let’s face 
it: there is no way in hell we’re going to stop these things from coming to our cities when there 
are hundreds of billions of dollars on the line. And to be clear, most of these incidents would not 
even have happen if the self-driving car companies still had a safety drivers like they used to,
but they’re under so much pressure to deliver that they’d rather fuck up the city, and even 
potentially kill people, rather than to admit   that the technology isn’t ready yet.
Tech companies proudly say that they “move fast and break things.”
And maybe that broken thing   will be your spine under a robotaxi, but 
that’s a risk they’re willing to take. Self-driving car companies have already 
lobbied politicians to allow them to   operate based only on “self-certification”,
And they’ve successfully stopped laws that require them to report extensive safety data,
Recently California passed a bill that says police can’t even ticket self-driving car 
companies when their cars violate traffic laws. It may be that eventually, on average, 
self-driving cars will be safer than   human drivers in the future, but it is 
incredibly naive to believe that they will not introduce new safety problems of their own.
A self-driving car will try to avoid hitting people and objects, and there are dozens of 
sensors to make sure that doesn’t happen. But these sensors are not infallible, and 
they require interpretation by software. It’s important for self-driving cars not to 
crash into things that are there. But it’s also   important that they don’t stop things that aren’t.
Like this Tesla in self-driving mode that suddenly came to a stop on the highway 
and caused an 8-car pile up.  You don’t want a self-driving car to stop 
just because there’s a bit of garbage blown onto the road, or it detects something that 
won’t cross the path of the vehicle anyway.  So the software needs to evaluate 
every object that it detects. The first person to ever be killed 
by a self-driving car was a woman   in Tempe, Arizona in 2018.
She was crossing this road at night while pushing a bicycle and was hit by
a modified Volvo SUV driven autonomously by the automated driving system developed by Uber.
So why didn’t all of this advanced technology prevent killing her?
It’s not that the car didn’t detect her,   it’s just that it didn’t think she 
was anything worth stopping for. The crash report from the NTSB provides a 
detailed explanation. The car flipped between   identifying her as a vehicle, to a bicycle, 
to a general ‘other’ category. Sometimes it detected her as being in another lane and other 
times as a static object that wasn’t moving. The software didn’t consider her to be an important 
object that was going to cross the path of the   SUV until 1.2 seconds before the crash, when 
it was too late to stop without killing her. So when a self-driving car starts 
driving toward you, you’d better hope   it hasn’t detected you as ‘other’.
But all of this talk about sensors and self-driving obfuscates the 
real root cause of this crash.  Which is the atrocious state of American 
traffic engineering. This area is designed like a freeway with a high speed limit, and 
almost no accommodations for people walking. Despite being directly between a 
rail station and hiking trails.  At the time of the crash, the median had 
a paved roadway designed only to be used by cars if one of the bridges was closed, but 
it clearly looked like a walking path and it   was regularly used by people crossing this road.
Instead of any actual pedestrian infrastructure, the city just put up these “use crosswalk” 
signs that are still there today.  It’s disgraceful that traffic engineers put 
up these signs at a natural place where people would want to cross the road, instead of, 
y’know, actually building a crosswalk here. But what’s worse is that in 2018 it 
wasn’t even physically possible to   walk to the crosswalk because of the trees 
and greenery that had been planted there. Since the crash, the “solution” has been 
to turn the brick roadway into a rough   stone path that is hostile to pedestrians.
But they’ve now cut a hole through the cacti so that it is at least theoretically 
possible to reach the crosswalk now.  Though you still have to walk 
over 200 metres out of your way, Plus the time it takes to wait for 
your turn to cross this highway.  American traffic engineers routinely 
bend over backwards to make driving as convenient as possible,
while making these places   actively hostile to pedestrians, then blame 
the pedestrians when they’re hit by a car. In the US, around 42 thousand people 
are killed in car crashes every year,   and hundreds of thousands are seriously injured,
sometimes with debilitating injuries that will affect them for the rest of their lives.
So it’s not surprising that the most prominent   feature promised by these American 
self-driving car companies is safety. But the US also has some of the most 
dangerous roads in the developed world. The most deadly kind of road in the US is the 
stroad, a road designed for high-speed travel, that also tries to act as a destination.
Stroads are rare in other countries, especially inside of cities, but 
they are all over the US and Canada.  The non-profit organisation Strong Towns has been 
warning about the dangers of this kind of road design for over a decade, and I have a previous 
video about stroads if you’d like to learn more. There has been decades of research into road 
safety and we know how to make the roads safer, such as the “Vision Zero” method, 
developed in Sweden in the 1990s.  In fact, if the US had roads as safe as Sweden 
they would cut their road fatalities by over 80%. Without any new technology at all.
But many of these design changes   require slowing or restricting cars, 
and that’s a step too far for the US. There’s no money to be made in making the streets 
safer, so the only solution to the dangers caused by the old cars has to be buying new cars.
Self-driving car companies do not care about road safety, they care about 
selling cars and technology.  They care about safety insomuch 
as it allows them to sell cars. The moment they can convince regulators and 
the general public that their cars are “safe   enough” they will stop caring about safety.
So you can expect the messaging to pretty quickly go from “preventing road deaths” 
to being “yeah, sure, they kill people,   but our marketing department says it’s less 
often than human drivers would, so it’s fine.” And even if they are safer than American drivers 
on American roads, the rest of the world is going to get these cars whether we want them or not.
And that’s an interesting thing to consider because right now, these 
vehicles are being trained in   some of the most car-centric places on Earth.
The first locations were undoubtedly chosen because they have very good weather.
But I guarantee you that Phoenix,   Arizona was also a top choice because it 
is literally like driving on easy mode. The roads are insanely wide, there are 
almost no restrictions on where cars can go,  it’s made up of square grid blocks a mile 
on each side, and pretty much everybody drives everywhere all the time already.
To say that Phoenix is car-dependent is a massive understatement.
If anthropomorphic cars could   design cities, they would design Phoenix.
It’s also one of the most dangerous places for pedestrians in the entire developed 
world, but don’t worry about that.  Self-driving cars will fix everything.
San Francisco is certainly more urban. There are people who walk and cycle,
and there are public transit vehicles as well. But I’ve lived in the Bay Area, and I can tell you 
that even San Francisco is pretty car-centric and car-friendly compared to most cities in Europe.
Which I think is going to be a problem when this technology is rolled out internationally.
These cars are trained to drive like Americans on American roads, where cars are considered superior 
and above all other forms of transportation. Have you seen an American when they try to drive 
in Europe for the first time? It’s not pretty. Of course, these tech companies could spend 
years re-training their vehicles to drive   like Europeans on European streets. But you 
and I both know that’s not going to happen. They’re going to set American-trained 
cars loose on European cities.  I hope I’m wrong, but I wouldn’t 
be surprised if European cities are pressured to “Americanize” their streets, 
to make them more self-driving-friendly.  Which brings us to another potential 
issue with these cars: they are not so much programmed as they are trained.
They learned how to drive by monitoring   what real human drivers did through 
millions of miles of driving. So let’s say that in the future, a city wants 
to change the way their streets work. To improve   safety, to speed up public transit, or to make 
more room for walking and cycling. Whatever. It won’t be enough to just put up a 
sign and update the rules of the road,  governments will need to convince every 
self-driving car company to re-train their cars. Given that these are multi-billion 
dollar international corporations with an   army of lobbyists, how likely do you think 
it is, that those changes will go through? This could result in our cities getting “locked 
in” to what they are today, with any new changes being dictated by car companies.
Which I guarantee you is not going   to favour anything that results in fewer cars.
Of course, there are a lot of problems with our current transportation systems, but I am sceptical 
that self-driving cars will truly solve any of them. Because apart from the supposed safety 
claims, when you boil it down, self-driving cars are really just promising to be cheaper taxis.
Yeah, they’re not driven by a human,   or at least they’re not driven by a human who 
is currently in the car with you, but otherwise they’re just cheaper taxis, right?
Like, do you seriously believe that   all of our urban transportation problems would 
be solved tomorrow, if only taxis were cheaper? And I’m even sceptical about the “cheaper” part.
We saw this happen with Uber and Lyft,   when app-requested taxis, sorry I 
mean, “ride-sharing” was introduced. Uber was cheap. That is, when 
venture capitalist money was flowing.  In fact it was so cheap that several 
cities in the US reduced investment in public transit in favour of subsidising 
rideshare rides. And it worked … for a while. But taking an Uber isn’t cheap anymore. And 
I’ll leave a link to this great article with   an analysis of why those price increases 
can’t be explained by an increase in costs, rideshare legislation, or inflation.
They raised prices because the owners and investors of those companies wanted 
a big return on their investments.  The exact same thing is playing out with 
self-driving cars. Hundreds of billions of venture capital are flowing into the 
development of autonomous vehicles,  and companies like Waymo are using 
that money to make their robotaxi rides price-competitive to traditional taxis.
And I find it wild that people online will unironically say things like, “in 
the future, AVs will be cheaper than   taxis because there’s no driver to pay.”
Like, do you honestly believe that the prices that companies charge are based on what 
they cost to deliver? Of course they aren’t! They charge as much as possible to maximise 
profits and once they’ve eliminated or   colluded with the competition they 
can charge whatever they want. And I shouldn’t even need to say this, but 
turning your transportation system over to a   handful of multi-billion dollar corporations 
does not result in good outcomes for anybody except the owners of those corporations.
It’s commonly claimed that self-driving cars will reduce the amount of cars on 
the road and eliminate parking lots.  The idea is that you will be dropped off 
at the door by a self-driving car and you won’t need to park. And that’s true. But 
you can literally do that today, right? You could just take a taxi to work 
every day instead of owning a car,   but very few people actually do this. Sure, it’s expensive to take taxis everywhere of 
course, but even wealthy people don’t do this,   because people just like owning their 
own stuff and being in their own car. There’s this vague promise that your 
car could earn extra money for you   as a taxi while you’re at work,
but how many people want their private car to end up looking – and 
smelling – like the inside of a taxi? So those cars are going to need to park 
somewhere, and their owners will want   them parked nearby so they don’t take too 
long to arrive when they’re ready to go. And you can forget about ever charging for 
parking, because if it’s cheaper to just let your   car cruise around the block while you’re shopping 
or at work, then that’s what people will do. Another common claim is that these 
cars will allow universal mobility   to everyone, especially the disabled,
But many people with disabilities require help from others to get in and out of 
a vehicle, or to secure a wheelchair,  so an attendant will likely 
be needed on-board anyway. And given that you need a credit card and 
a modern smartphone to book a robotaxi,   this puts it out of reach of many people.
Sure, the rich Boomer with a bad hip will be taken care of, but they can already take a taxi today.
This is really just an extension of the often-repeated myth that cars are the only 
transportation that works for disabled people,   and that as long as you have cars, 
disabled people are taken care of. This is a great article by disability advocates 
about how self-driving cars are a distraction to the solutions they actually need, and 
I’ll leave a link to it in the description. But of all of the claims made by 
proponents of self-driving cars,   the one that I find the most ridiculous is the 
idea that they will reduce traffic congestion. There is absolutely no reason to 
believe that this is true, and lots   of historical evidence to suggest that it is not.
Every single time that we have done anything to make transportation cheaper or more convenient, 
the inevitable outcome is that people travel more often, and they travel longer distances.
There is decades of well-documented evidence that every time we widen roads or build new 
highways, it increases the amount of driving, and I talked about that in more detail 
in my video about induced demand.  When Uber and Lyft came to cities, they also 
claimed that they would reduce traffic congestion, but the exact opposite happened, and vehicle 
miles travelled increased significantly This will only become worse with 
robotaxis because people will take   them even more often than they do Ubers today, 
especially while they’re heavily subsidised. I’ve seen some people claim that robotaxis will 
reduce the number of cars in cities because   they will park in parking lots on the edge of 
town, and only come in when they’re requested. But I don’t believe that for a second, 
because one of the major competitive   factors between robotaxi companies is going to 
be how quickly they can get to their customers. Consider how impatient people get 
while waiting for an elevator.  There’s no way in hell that people 
in the future will want to wait 14 minutes for a robotaxi pickup.
Which is why the self-driving car parking lots today are right in the heart of San 
Francisco. But the future could be even worse. There’s no cost to driving around most cities.
Gas taxes are usually the only thing that even comes close to a “user fee” 
for city streets but regardless,  the robotaxis of the future will be 
electric and charged at private facilities. As long as the roads are free to use there 
is zero incentive for these companies not to have robotaxis constantly circling 
the block so that they can respond   to customers as quickly as possible.
And when these cars are significantly less space-efficient than buses or trams, 
that alone will result in less space   available in our cities. But it gets worse.
Consider today a suburban parent who drops off one child at elementary school, another 
at highschool, and then drives to work.  With self-driving cars, those kids will be sent 
in their own robotaxis, so that one car trip today will become three car trips in the future.
This is already happening in San Francisco, and it will only get worse.
It’s great for kids to have   independent mobility, but we need to do that 
without just putting more cars on the road. In an earlier video I talked about 
the problems that many European   cities have already experienced after the 
introduction of 10 minute delivery services. And there are stories today from 
Doordash drivers who were called   to pick up trivially small orders, 
like a single tube of toothpaste. As robotaxis become cheaper and more convenient, 
this is only going become more common. They’re going to be delivering anything within a 
few minutes, with a car for every delivery. When cars become common, people didn’t 
need to live near their work anymore.  And cities started to sprawl as more and more 
car-dependent suburbs were built on the periphery. This resulted in huge volumes of traffic funnelled 
into arterial roads and highways, because there were more cars driving longer distances.
Self-driving cars will only make this worse. When people are freed from having to 
pay attention to the road they’ll be   able to go through their emails or take 
conference calls from their autonomous vehicle. So why live in a suburb at all?
Why not live in a cottage in the middle of nowhere where land is plentiful and taxes are cheap?
This will result in an exponential increase in the amount of vehicle miles travelled, 
with millions of cars driving very   long distances regularly and putting even 
more strain on our transportation system. And it’s going to make the existing problem of 
financing suburban infrastructure even worse.  See my Strong Towns series to learn 
how suburbanisation and car-centric development is literally bankrupting 
North American suburbs already.  But if you believe in the marketing of 
autonomous vehicles, then this technology will advance exponentially.
Vehicles will communicate   with each other to travel efficiently.
Autonomous pods will merge seamlessly onto high-speed self-driving highways 
for even higher throughput and speeds.  And vehicles will be more efficient than 
ever through platooning, or whatever. In this vision of the future, the transportation 
problems of today will seem laughably archaic, and we will achieve levels of 
efficiency we can only dream of.  Here’s a marketing video from Ford, showing what 
they think the city of the future will be like. Look at all of these problems 
with the modern city! Of course, these problems were caused by 
people using their cars in the first place, but don’t worry about that, because buying their 
new cars will fix everything, so it’s fine! Look! There are ebikes and parks and 
futuristic trains and less parking. People can cross the street whenever they 
want and magic crosswalks will appear. There are vacuum trains, pedestrian 
plazas, blue skies, and lots of trees. Every single problem will be 
solved. What’s not to like?  But before we get too excited about our utopian 
self-driving future, it’s worth looking back at previous promises of the automobile industry.
At the 1939 New York World’s Fair, General Motors presented Futurama, the 
high-tech city of future in the year 1960. This exhibit was absolutely huge.
They had a ride that would take you   over the city, and even a life-sized 
model of an urban neighbourhood. In Futurama, car traffic would be eliminated 
completely with advanced motorways. These would be separate from all other traffic and 
allow cars to drive faster than ever before. Pedestrians would be able to 
comfortably walk through the   city on elevated walkways while car traffic 
was safely whisked away on the roads below. But the reality was quite different.
The motorways promised by Futurama   were built, but they just induced demand 
for more driving and more car ownership. Which was the car industry’s 
real purpose in the first place,   of course. They wanted to sell more cars.
Instead of being the salvation of the city, elevated freeways divided 
and destroyed neighbourhoods.  They flooded city streets with car traffic, 
increased pollution, and left municipalities with massive tax burdens and maintenance liabilities.
To the point where many cities have spent billions to remove urban freeways 
in an attempt to repair the damage. There are many accounts online that show 
the destruction that happened to American   cities because of car infrastructure, and 
I’ll leave some links in the description. My point here is that new transportation 
technology can and will fundamentally change our cities, especially if there’s money to be made.
So we need to seriously consider how AVs could negatively affect cities, so that 
we can prepare for them today.  As AV technology improves, it will 
start to arrive in personal vehicles. Just like with cars, the first mass 
adoption will happen in suburbia,   especially in the US and Canada.
Suburbanites spend too much time sitting in traffic as it is, and it will 
allow those people to live even farther   away from the city, where property is even 
cheaper, so it will definitely be worth the cost for any reasonably wealthy suburbanite.
And designing AVs for suburban roads is a much easier problem than for busy cities, 
so they’ll be faster to market, too.  AVs will be proven safer than human drivers, so 
AV companies will lobby to have AV-only lanes installed on highways and throughout the suburbs.
For those who can’t afford their own autonomous vehicle, suburban municipalities will subsidise 
robotaxis, in place of public transit. Entire suburban neighbourhoods will be built 
that are only accessible by autonomous vehicles, which has the added benefit of 
keeping any undesirable people out. This is the next-logical step to the gated 
communities that are already common across the US. Suburbs will rapidly expand to 
consume even more valuable land.  Power lines, water pipes, sewage pipes and 
all other infrastructure will become even longer and more spread out, but there will be so 
much money and debt flooding into suburbs on the   promise of building the city of the future 
that nobody will stop to think about it. And besides, the bills to 
maintain it all are decades away.  In other words, literally exactly what happened 
in the post-war suburbs of the last century. Autonomous vehicle companies may be making record 
profits from suburbanites, but the line must go up, so they’ll look towards the cities.
And their primary objective will be   to eliminate the competition.
In the 1920s, Los Angeles had the largest electric streetcar network in the world.
But service declined significantly as the streetcars got stuck in traffic behind cars,
And the streetcars couldn’t compete with the heavily subsidised road and highway system.
Private streetcar companies were bought out by National City Lines, a company that was 
financed by General Motors, and related companies, and was specifically founded with a goal of 
tearing up streetcar lines throughout the US. Because well-functioning public 
transit is a threat to car companies. History is going to repeat itself here: the 
companies who finance autonomous vehicles   are going to do everything in their power 
to cancel transit projects, lobby against funding for public transportation, and replace 
public transit with their private robotaxis. There will be autonomous bus concepts 
promised, and some of these will go into   operation to replace today’s buses and trams.
But the real money will be in private AVs and robotaxis. Robotaxis will be pitched as 
a more efficient form of transportation, by offering door-to-door service.
And any train or transit lines   will be converted to AV lanes. But there’s 
nothing more space-efficient than a train. This 18 lane freeway in Toronto is over 100 
metres wide, but it moves fewer people per day Than this single subway line, because trains 
are so much more space efficient than cars. In the early 1900s the Brooklyn Bridge in New 
York moved over 420 thousand people per day. But in the 1950s, the trolleys 
were removed to make way for cars.  Since then, the Brooklyn Bridge has never moved 
more than 180 thousand people per day. Well under half as many, as when there was public transit.
The same thing is going to happen to our cities with AVs. The decrease in space efficiency moving 
from shared transit to private transportation, coupled with the massive increase in 
demand for cheap, subsidised, AV rides,   will result in a massive increase in 
the number of vehicles in our cities. AV companies will lobby for some roads 
to be designated as autonomous-only. This will be pitched as a way to 
increase safety and efficiency,   but the ultimate goal will be to eliminate public 
transit and human driving, and get everybody to sign up to an AV subscription instead.
Here’s how a concept of this called “Loop NYC” was pitched by a New York 
engineering company a few years back.  In this video they assign certain streets 
in New York to be self-driving-only. They claim that this will be so much 
more efficient that other streets   could be blocked off to cars completely 
and turned into a giant linear park. But AV companies aren’t going to be happy 
giving up any roads, as their success is dependent on offering door-to-door service.
So AVs will be required to go everywhere. They’ll even argue that AVs should also be allowed 
in pedestrian areas, because they’re safer than human drivers, and will offer an important 
service for people with mobility issues. When there’s no price put on driving, 
competing robotaxis will circle the block,   in order to have the quickest response times.
Private AVs will drive around automatically to get the cheapest parking rates 
while their owners are at work.  The parking lots that we were told would 
become obsolete will be converted to solar AV charging stations for 
both private AVs and robotaxis. And they’ll jump between stations for 
whichever one has the cheapest prices.  So despite what the car company 
propaganda might show you, AVs are not going to let us replace roads and 
parking lots with parks. Autonomous vehicles will demand even more space in our cities.
And they’re going to get it.  The next problem will be pedestrians.
As AVs become more common, people will realise they can cross the street wherever 
they want, because AVs will always stop. And there’s no way that AV companies 
are going to put up with that. They’ll claim that people crossing the street 
are crippling the transportation system,   and they’ll demand that fences be put 
up along important streets and roads. Bamberg is a small town in 
the US state of South Carolina  Like pretty much every American town 
they had a traditional main street. But this main street also became the 
main highway through town, and all of   that downtown activity, and especially 
all of the people crossing the street, was interfering with the flow of car traffic.
So they decided to widen the street and install fences along the sidewalks to 
prevent people from crossing mid-block. After construction was done, car 
traffic moved much more efficiently!  And within a few years, every single shop 
in downtown Bamberg went out of business, and today it looks like a ghost town.
Because believe it or not,   people don’t enjoy being on a cramped 
sidewalk next to high speed traffic. This was a monumentally stupid idea of course, 
but I bring this up because before the project, the director of the chamber of commerce said that 
the project will, quote, “be a major boost to the economy for the county and downtown Bamberg.”
There are a lot of people who genuinely believe that if car traffic moves more 
efficiently, then business will thrive.  When the exact opposite is true. 
Cars don’t buy things, people do. So if you make a place inhospitable 
to people, nobody will buy anything. Expect the same thing to happen 
with autonomous vehicles.  As the fences go up, there will be no reason 
for AVs to drive slowly, so they won’t. Streets everywhere will be just like downtown 
Bamberg: inhospitable places where people won’t want to be, and street life will slowly die.
With public transit, human drivers, and pedestrians out of the way, AV companies 
will demand an end to speed limits.  They’ll argue that these restrictions 
designed for human drivers are irrelevant for vehicles driven by a computer.
This will result in a huge increase   in vehicle throughput, and will be 
considered a tremendous success. But when AVs are driven faster, pedestrian 
and cyclist fatalities will increase. AVs may have lightning-fast reaction times, but 
at high speed, the laws of physics win every time. There will be outcries whenever someone is killed, 
and people will demand to slow down AVs to make them safer, but history will repeat itself again.
Today, cars are a reality that we deal with just about everywhere.
But when automobiles were first introduced,   people did not want them in their cities.
Cars were clogging up city streets and killing thousands of people every year.
By the 1920s, the general public was very anti-car. Several cities had plans to require 
speed governors on all cars entering the city, And they introduced high 
fines for dangerous driving.  But the automobile industry was not about 
to go down without a fight. They fought to get car-friendly politicians elected 
and vote down any anti-car legislation. And did they do anything to make their cars 
less likely to kill people? No, of course not. They invented and spent millions promoting the 
concept of “jaywalking” to make it the fault of the victims for not looking where they were going.
This changed cities almost overnight. Streets went from a common place for everyone …
to a place exclusively for cars just a few decades later.
The “jaywalking” of the 21st   century will be the transponder that everyone 
will be required to wear when walking outside, if they want to avoid being killed by a car.  When someone is hit by an AV, people will ask, 
“Well, were they wearing their transponder?”. And the news will report, “the victim 
was not wearing a LIDAR reflector.” These faster cars will also 
create more pollution and noise.  Cars are a major contributor 
to pollution within cities. Air quality will improve as 
cars transition to electric,   but electric cars will not solve this problem.
Because a major source of local pollution from cars comes from the erosion of the 
tyres, brakes, asphalt, and road markings. And all of these emissions increase as cars 
get heavier and drive at higher speeds. Cars also create a lot of noise pollution in 
cities, and that is not just just uncomfortable,   it literally creates a stress response 
that leads to physical problems, and I’ve talked about that in a previous video.
There’s a common belief that electric cars will solve this noise problem, 
but that’s not really true.  Because once a car goes above about 50 
kilometres per hour, the sound of the tires becomes louder than the sound of the engine, 
and this is worse as cars get heavier. Plus the faster the car goes, the louder this 
noise becomes. Many cars in Oslo are electric, but it’s still very loud on this 
pedestrian bridge over the highway. With all of the extra electronics and computers, 
AVs will be heavier than today’s cars, and they’re going to be driven faster, too.
Which will make the air toxic near   autonomous vehicle highways.
And the noise from the tires on the road surface will be deafening, 
making these areas uncomfortable and   dangerous for anybody walking nearby.
Eventually, AVs will consume every street in the city, and push out 
every other form of transportation.  The final step will be to remove traffic lights.
About a decade ago, I saw this video, produced by the University of Texas at Austin.
They proposed a future protocol called AIM, that would provide Autonomous Intersection Management,
for high speed intersections without traffic lights. This is the animation that made 
me really think twice about the future   of self-driving cars. Because when I looked at 
it, the first thing that crossed my mind was, “huh. It would really suck to have to cross 
this intersection while walking or cycling.” Of course, there will be people 
who protest and complain.  They won’t want their neighbourhood 
being cut in half by an AV highway. But we can’t have these Luddites 
denying us our transportation utopia.  As a compromise, we’ll be promised pedestrian 
bridges to keep neighbourhoods connected. Though due to cost-cutting, only a 
few of them will actually get built. When highways came to New York 
City, they completely destroyed   some neighbourhoods and separated others.
A handful of pedestrian bridges like this one were built to re-connect neighbourhoods.
These bridges are awful. They require walking up long ramps, they feel like 
dirty concrete trenches, and they’re really ugly. The chain link fence separates you from 
the high-speed traffic below, but the   noise and pollution from cars is unbearable.
These were clearly built to benefit drivers by keeping pedestrians out of the way, and 
not for the benefit of people walking. In the AV city, you won’t 
be able to cross the street.  You’ll need to go out of your way 
to the nearest pedestrian bridge, walk up the ramp, cross the bridge, and 
walk back down to where you want to go.  All while enduring the deafening sound and 
toxic air of the high-speed traffic below. This will be the city of the future.
Streets that are completely   consumed by autonomous vehicles.
Intersections that are impossible to cross. Cities that are carved up by self-driving 
expressways, creating islands that are   infeasible to leave on foot.
With air and noise pollution so bad it makes travel outside unbearable.
We’ll have a transportation system that consists entirely of autonomous vehicles.
So that you won’t be able to do   anything – go to work, go to school, or 
even buy food for your family – without paying an AV company to get you there.
And once every other form of transportation is either eliminated or infeasible, the 
autonomous vehicle companies will ramp up prices. They’ll be able to charge whatever they 
want because we’ll have no alternatives.  So what can we do about it?
Well, this brings me to my fundamental issue with the entire concept of self-driving cars.
I grew up in a car-dependent city called   London. London, Ontario, Canada.
And despite its Thames River Oxford Street, and
Covent Garden Market,   it’s nothing like Real London.
A kind of Fake London, if you will. When I arrive in Fake London by 
train I’m on the side of a high-speed   stroad across from a surface parking lot.
There isn’t much left of downtown and my only option to get anywhere is a taxi.
The city is very spread out and   designed almost exclusively for cars.
Bus service is highly unreliable and slow. And there is no way I would feel 
safe enough to ride a bike here.  I have to drive pretty much 
everywhere to do pretty much anything. Which I gotta say, kinda sucks.
Too many people drive really dangerously And there are regular news stories about 
people being killed in car crashes. So when I’m sitting on a giant stroad, stuck in 
traffic, the idea of having the car take care of all this boring, dangerous driving seems 
pretty great. But then I think about Utrecht. Utrecht is a city in the middle of the 
Netherlands, and I have to go there   fairly regularly for one reason or another.
The population of Utrecht is very similar to that of Fake London. But otherwise, 
these cities are very, very different.  So you’re probably thinking that it’s almost 
comical to compare Utrecht to London, right? It seems ridiculous because these 
cities have almost nothing in common.  And yet, a hundred years ago, 
they were pretty similar. Sure, the architecture was 
different, but the fundamental   design of the city was pretty comparable.
Both were compact, walkable cities full of mixed-use neighbourhoods and connected 
by streetcars, because both cities were   founded long before automobiles were invented.
Over the 20th century, both cities embraced the automobile, and both cities 
built wide roads and highways.  Utrecht filled in a canal in their city 
centre and turned it into a highway. And London bulldozed a neighbourhood to 
build a wide road straight into downtown. Utrecht and London both even bulldozed part 
of their city centre to build a suburban-style shopping mall. And these malls were both the 
first of their kind in their respective countries. Utrecht was about a decade behind London, 
but otherwise, these cities were on exactly the same path … that is, until about the 
1990s when Utrecht decided to change course, due to growing concerns about the negative 
impact of car-centric development.  They made a fundamental decision to make the 
city friendlier to people walking and cycling, to reduce car traffic, and to 
create better public spaces.  The highway around the city centre …
Was turned back into a canal. You’d never know that this used to be a highway.
And today that downtown shopping centre is connected directly to a new train 
station via a pedestrianised area, And closely integrated with the rest 
of the city centre on the other side.  It’s such an incredible difference to arrive 
in this place by train compared to Fake London. Despite having almost the same population as 
my hometown, it hosts the busiest train station in the Netherlands, transporting 
over 200 thousand people per day. With over a thousand daily departures 
on national and international trains. There are hundreds of destinations 
within walking distance of the station.  And there are great connections to other forms of 
public transit, including this new tram system. Underneath the station is the world’s 
largest bicycle parking garage,  with spaces for over 12 thousand bicycles.
And I can pick up one of the hundreds of rental bicycles using my transit card,
and cycle to any destination.  I can cycle anywhere I want quickly and easily, 
and because of the high-quality infrastructure, I never feel unsafe.
It also helps that car   volumes are low and so are the speed limits.
Near the train station is this street accessible only to walking, cycling, and public transit.
This is one of the busiest cycling paths in the world and I probably see more people 
cycling in 5 minutes here than I have in   my entire life in Fake London.
Walking around the city is also convenient and comfortable.
The streets are lined with   interesting shops and restaurants, and there 
isn’t a surface parking lot in sight anymore. There are many nice residential neighbourhoods
And it’s remarkably quiet, too. Because as I’ve realised over the years, 
cities aren’t loud, cars are loud. I don’t feel unsafe walking and cycling here, 
so the safety promises of autonomous vehicles   don’t really seem to matter that much.
There isn’t a lot of traffic because there are viable alternatives to driving.
People aren’t forced to drive here,   so the only people in cars are those who need 
to drive, or those who really want to drive. Self-driving cars are supposed to provide mobility 
to children, the disabled, and the elderly.  And yet I see all of those people 
getting around just fine in Utrecht. Because universal access to 
mobility isn’t a fundamental issue:   it’s a problem caused by car dependency.
Would this guy be better off in a self-driving car? Maybe, I didn’t ask him. But it’s not like 
he’s unable to get around independently here, like would be the case where I’m from.
Now, Utrecht isn’t perfect, And the city still shows the scars 
of that 1960s car-centric development  but it also just doesn’t have most of the problems 
that autonomous vehicles are supposed to solve. And yet they did it without 
any advanced technology.  Without spending hundreds of billions of dollars. And without letting their city 
be controlled by the whims of   multi-billion dollar corporations, either.
If everybody here was in self-driving cars, this place would be much, much worse.
So it seems what we really need is not driverless cars, it’s car-less drivers.
I think Utrecht, and cities like it, can provide a template for what we should 
be doing to prepare for self-driving cars. We need to limit where cars can go 
and that includes autonomous vehicles. Utrecht has many places that are off-limits to 
cars, but motor vehicles are extremely useful sometimes, so access is still 
provided for delivery trucks,  emergency vehicles, and people with disabilities.
We need fewer cars in cities, so we should definitely tear down urban freeways that 
divide neighbourhoods and turn the space   into parks, shops, and houses.
Cars should go around cities, and not through them. We should make it impossible 
to drive through the middle of the city by car. The most direct route should only 
be accessible by walking, cycling,   and public transit, enforced by modal filters.
We also need to lower speed limits now, to make the streets safer today, but 
also to reduce noise and pollution. Lowering speed limits can sometimes be 
controversial, but after it’s done there   is nearly universal support to keep them low.
So let’s do it now, before AVs take a foothold. We need to remove parking, 
especially surface parking lots.  We’re never going to be able to charge for 
parking once cars can drive themselves, so let’s start removing parking spaces now, 
and turning them into something more useful. And let’s stop building any new parking 
garages too. After all, they tell us they   won’t even be necessary once we have autonomous 
vehicles, right? So let’s not build them at all. Instead, let’s build more mixed-use walkable 
neighbourhoods, so that the places where people   live, work, and shop are closer together,
making it easy to get to where you’re going without needing a car at all.
And safe sidewalks and bike lanes   should be part of the standard street design 
guidelines, so that they’re built automatically with every new construction.
We need to invest in functional   and viable public transit, especially 
infrastructure that is difficult to remove. History has shown us that bus routes will get 
overrun by cars, so we need more trams and trains, and a lot more dedicated transit lanes.
And we have to ensure that public transit remains under democratic control, so that it 
can never be captured by private corporations. And most importantly of all, and I cannot 
stress this enough, we need to put a price on driving. We absolutely must keep AVs from 
circling around and clogging up our streets, The gas tax is going away as cars 
transition to electric anyway,   so let’s replace it with something better.
Cars should be charged based on how much they drive, but also when they drive too, with a 
higher price during periods of peak congestion. If we do all these things to fortify 
our cities, then we might just be   ready for the onslaught of autonomous vehicles.
Then we can allow AVs into the city on our terms, to provide transportation where it’s 
genuinely needed. Not just where it   profits tech companies the most.
And if the naysayers are correct, that self-driving cars really are just a pipe 
dream, and they never reach mass-adoption,  then I guess we’ll do all that work and just 
be left with beautiful, functional, efficient, and sustainable cities accessible to all. 
And that might just be the best outcome.