I think this post represents a bit of a failure of imagination. Robotaxis don't need the form factor of cars; they could be buses as Bill Lovotti asked or tiny one or two person vehicles more akin to motorcycles. This alone could significantly increase efficiency. They can drive more densely than human driven cars, increasing efficiency. They make it trivial to enforce traffic rules, since they can be enforced at a software level. There are fewer equity concerns where you're pitting the concerns of individual drivers against individual riders or communities. They also could do longer distance trips that are not possible or economically viable for a driver who needs to get home to sleep.
Obviously we're not doing all this yet, but they seem like fairly obvious opportunities for improvement.
One thing I will say about robotaxis, for all their faults, is that they are probably easier to electrify than fleets of buses. An electrified bus fleet means batteries and/or roadside charging stations need to be sized to ensure that every bus can complete every scheduled run, all day, every day, under all weather conditions, no matter what. An electrified robotaxi car, on the other hand, can just drive until its batteries get low, then go back to base, charge up, and do it again.
If colder weather means the vehicles have more energy consumption and less range, the robotaxi system will have less capacity (due to more vehicles being offline than normal for charging), but there is no risk of unlucky people being stuck, waiting for a bus that will never come because the bus had to end its run an hour early to charge its batteries. To prevent this, an electric bus fleet must dramatically oversize its batteries and charging stations in ways that an electric robotaxi fleet might not need to, especially if the coldest days of the year will typically have fewer than normal people up and about, anyway (people don't want to walk even from car to building when it's -20 outside if they don't have to).
There is also another issue that, at least today, in the U.S. market, the cost premium of electric vehicle vs. fuel burning vehicle is dramatically higher for buses than cars, way beyond that which can be explained simply by the size of the batteries. I suspect that a lot of this has to with the fact that the electric car industry has big competition, with lots of companies trying to compete on price, while the electric bus industry is dominated by a couple of companies that behave like lazy monopolists, overcharging those few jurisdictions with zero-emissions mandates because they can, while selling everybody else diesel buses for more money than what Chinese battery buses are going for in the rest of the world. Bottom line, the U.S. transit bus market just isn't big enough to entice a new competitor to enter the market, while steep tariffs and "Buy America" FTA rules work well to keep competition from the rest of the world out. And, as long as public transit is treated as such an afterthought by national politicians, nothing about this will ever change. (Republicans would rather see transit wither and die than be more efficient; Democrats would rather see more union jobs making buses than see transit be more efficient; there is no real constituency from either party for making public transit function efficiently, so overpriced stuff is what we get).
I know this isn’t the point of the piece, but for those who haven’t experienced such a ride, we took a couple Waymos during a SF trip last month. Don’t think I’ve ever enjoyed a smoother, more comforting taxi ride. Once we got the hang, it worked quite well and was price-competitive.
As far as double-parking, our only problem was that it WOULDN’T; we had to chase the thing down once when it couldn’t decide whether it was in a good spot to pull over. Likewise, there was a moment when the vehicle froze in place at a busy tourist-trap taxi line. Actual taxi drivers were yelling at us to “move,” which we of course couldn’t make happen.
Numerous friends were aghast that we would try such a service. “Nope, never,” was the refrain. I don’t think it’s because they were concerned about downstream urban traffic issues…
Matt is not just telling people afraid of public transit to eat their spinach. The whole point of his article is to hinder and burden the mode of transportation people choose, so they will be compelled to use the mode of transportation they should. People will hate you when you do this. Public transit has to compete on cost, comfort and convenience and that means ruthlessly eliminating costs, nuisances etc. People will love you when you do this.
This is a weirdly close-minded, doomed take from Slow Boring, obviously not from Matt.
The author assumes that traffic congestion is the big problem with cars in cities but in reality, it’s parking.
If robotaxis can get cheap enough that people give up their personal cars, there a huge amount of parking that could be turned into something more productive.
SF’s main goal should be to become a city where people don’t want or need a car, on NYC’s level. This will probably require a combo of cheaper taxis (robotaxis) and improved public transit.
I think robotaxis will favor walkable/public transport centered cities all on their own. The robotaxis will finally truly enable people to not own cars -- even with Uber and Lyft things like vet appointments or big grocery/shopping trips are awkward to impossible not to mention all the times you really don't want to deal with a stranger (sick kid or sick you to doc).
And switching a service from mostly upfront fixed costs with small hidden marginal costs (extra miles may add maintenance costs or hurt resale but it's not salient) to a system with larger marginal costs that directly take dollars out of your bank account has an absolutely huge effect. This is why pay per article and micropayments never took off online -- people have a barrier to spending money.
Once people start thinking "ohh do I really want to spend $15 so I don't have to take the bus" they'll start taking the bus way more. Indeed, I think it's critical not to impose taxes on them to make it affordable to go car free.
One modification I would make to a VMT tax is to factor in a vehicle’s weight. Heavier cars are massively more destructive to our roads, and also tend to be bigger (and therefore less economical as a single-occupancy commuter vehicle). They should not be taxed at the same rate as a sedan or other smaller car.
I’m not sure if the linked NYT article mentions that, but I’m not in a position to break the paywall right now.
One thing missing here: the old. We don't like to give up our licenses. We don't like to drive anywhere any time. We like our routines and almost always drive around our locality.
These factors would seem to suggest: there's a great potential market for selling robottaxis to the old. I suspect the extent of the market depends on the balance between our desire for control versus the possible drop in costs of robotaxis as drivers are eliminate and presumably the number of robotaxis begins to exceed the number of taxis we have had.
I have an honest question that I never see addressed by urbanism advocates/opponents. Is there an urban planning model that doesn't correlate with birth rates falling below replacements levels? Hearing about various demographic crises out there has made me wonder what research exists about urbanism's role in population decline has been.
I don’t believe a definite correlation between urban living and low birth rates has ever been found, as there is just too many confounding variables (education levels and inflated urban rents being primary ones).
I haven’t researched this extensively though, so I’m happy to be corrected if it has.
But the point is that this doesn’t define a causal link between urban living and lower birth rates, as living in cities also implies higher education and higher CoL (the latter of which is not a necessary fact of cities, just a truth about our current ones). Both of those could be the cause of lower birth rates
My impression is that birth rate declines happen basically anywhere women have decent rights and birth control is available.
It really is a difficult issue, and I don't know that there's a great solution. Many people suggest improved social services, but Sweden and France are pretty much the world standard for welfare states and have about the same birth rate as the US or Canada.
It's interesting how both self-driving trucking and personal AV ownership are being left out of this discussion. The latter might be a few more years down the road, but could revolutionize suburbia and rural area transportation. The big barrier on the former is probably regulation and laws.
I just got back from a trip visiting family in SF, and took a Waymo for the first time--it was incredible. One of the handful of times in my life experiencing awe at a technological advancement (another one that happened recently: having a Turing-test-passing conversation with chat GPT).
For me it was one of those moments where you realize the world is going to be fundamentally different, and your children won't have a memory of the world as it was before.
As someone who loves trains, metros, and all manner of grade separated fixed guideway transit, it feels to me that its era (at least in the US, limited as it was) is rapidly coming to an end. There are broadly two rational reasons to build high capacity transit today:
1. Geometry, as alluded to in the article: the city is so dense and requires so many people to be moved into such a small area that doing it via surface roadways becomes impossible.
2. Scaling operational costs: building fixed guideway rail massively increases the number of passengers per driver, since vehicles are much longer. So you invest some money now to reduce your labor costs over the long term.
With self-driving technology rapidly becoming mainstream, #2 will soon be dead letter: if an automated bus and an automated train have equivalent operational labor costs (ie roughly zero), why would you ever build the train? And, at least in the American context, there are very few cities dense enough that surface roads with maybe some dedicated bus lanes are insufficient with enough bus service, so #1 is out too. What reason is left, outside of maybe 5-7 truly super dense metros, to build a subway?
AVs don't have to be 5-seater SUV/truck abominations occupied by usually one person. The networks that run these can be smart. You can run cheap busses that dynamically adapt to demand to do curbside pickup and dropoff (or maybe a slight compromise for routing speed and efficiency, so you're never asked to travel more than a block or so) based on people whose routes happen to overlap. For solo or pair travelers, you can dispatch microcars. These are obvious cost optimizations to do in a competitive market.
The "car as lifestyle signaling / prosthetic genitals" thing completely goes away if the cars aren't yours. Nobody cares what the bus looks like. So vehicles can actually be the appropriate size for the trip.
AVs can give you something that is *systematically better* than bus-based public transit. No sending busses that run mostly empty sometimes to fill out the timetable so people can plan around them. If there's no demand, the network wouldn't send anything. If there was slight demand, it would send a smart car. AV networks could have much higher average vehicle utilization *and* provide an experience more similar to rideshare (hit a button in an app and get where you're going in a reasonable amount of time) that is much better for consumers.
Outside of the content I need to say this is an amazing lede: "In San Francisco, robotaxis are like naked guys. The first time you spot one, it feels like a big moment. After that, they fade into the background."
I think this post represents a bit of a failure of imagination. Robotaxis don't need the form factor of cars; they could be buses as Bill Lovotti asked or tiny one or two person vehicles more akin to motorcycles. This alone could significantly increase efficiency. They can drive more densely than human driven cars, increasing efficiency. They make it trivial to enforce traffic rules, since they can be enforced at a software level. There are fewer equity concerns where you're pitting the concerns of individual drivers against individual riders or communities. They also could do longer distance trips that are not possible or economically viable for a driver who needs to get home to sleep.
Obviously we're not doing all this yet, but they seem like fairly obvious opportunities for improvement.
One thing I will say about robotaxis, for all their faults, is that they are probably easier to electrify than fleets of buses. An electrified bus fleet means batteries and/or roadside charging stations need to be sized to ensure that every bus can complete every scheduled run, all day, every day, under all weather conditions, no matter what. An electrified robotaxi car, on the other hand, can just drive until its batteries get low, then go back to base, charge up, and do it again.
If colder weather means the vehicles have more energy consumption and less range, the robotaxi system will have less capacity (due to more vehicles being offline than normal for charging), but there is no risk of unlucky people being stuck, waiting for a bus that will never come because the bus had to end its run an hour early to charge its batteries. To prevent this, an electric bus fleet must dramatically oversize its batteries and charging stations in ways that an electric robotaxi fleet might not need to, especially if the coldest days of the year will typically have fewer than normal people up and about, anyway (people don't want to walk even from car to building when it's -20 outside if they don't have to).
There is also another issue that, at least today, in the U.S. market, the cost premium of electric vehicle vs. fuel burning vehicle is dramatically higher for buses than cars, way beyond that which can be explained simply by the size of the batteries. I suspect that a lot of this has to with the fact that the electric car industry has big competition, with lots of companies trying to compete on price, while the electric bus industry is dominated by a couple of companies that behave like lazy monopolists, overcharging those few jurisdictions with zero-emissions mandates because they can, while selling everybody else diesel buses for more money than what Chinese battery buses are going for in the rest of the world. Bottom line, the U.S. transit bus market just isn't big enough to entice a new competitor to enter the market, while steep tariffs and "Buy America" FTA rules work well to keep competition from the rest of the world out. And, as long as public transit is treated as such an afterthought by national politicians, nothing about this will ever change. (Republicans would rather see transit wither and die than be more efficient; Democrats would rather see more union jobs making buses than see transit be more efficient; there is no real constituency from either party for making public transit function efficiently, so overpriced stuff is what we get).
I know this isn’t the point of the piece, but for those who haven’t experienced such a ride, we took a couple Waymos during a SF trip last month. Don’t think I’ve ever enjoyed a smoother, more comforting taxi ride. Once we got the hang, it worked quite well and was price-competitive.
As far as double-parking, our only problem was that it WOULDN’T; we had to chase the thing down once when it couldn’t decide whether it was in a good spot to pull over. Likewise, there was a moment when the vehicle froze in place at a busy tourist-trap taxi line. Actual taxi drivers were yelling at us to “move,” which we of course couldn’t make happen.
Numerous friends were aghast that we would try such a service. “Nope, never,” was the refrain. I don’t think it’s because they were concerned about downstream urban traffic issues…
Matt is not just telling people afraid of public transit to eat their spinach. The whole point of his article is to hinder and burden the mode of transportation people choose, so they will be compelled to use the mode of transportation they should. People will hate you when you do this. Public transit has to compete on cost, comfort and convenience and that means ruthlessly eliminating costs, nuisances etc. People will love you when you do this.
I don’t want to be woken up by robotaxis.
This is a weirdly close-minded, doomed take from Slow Boring, obviously not from Matt.
The author assumes that traffic congestion is the big problem with cars in cities but in reality, it’s parking.
If robotaxis can get cheap enough that people give up their personal cars, there a huge amount of parking that could be turned into something more productive.
SF’s main goal should be to become a city where people don’t want or need a car, on NYC’s level. This will probably require a combo of cheaper taxis (robotaxis) and improved public transit.
I think robotaxis will favor walkable/public transport centered cities all on their own. The robotaxis will finally truly enable people to not own cars -- even with Uber and Lyft things like vet appointments or big grocery/shopping trips are awkward to impossible not to mention all the times you really don't want to deal with a stranger (sick kid or sick you to doc).
And switching a service from mostly upfront fixed costs with small hidden marginal costs (extra miles may add maintenance costs or hurt resale but it's not salient) to a system with larger marginal costs that directly take dollars out of your bank account has an absolutely huge effect. This is why pay per article and micropayments never took off online -- people have a barrier to spending money.
Once people start thinking "ohh do I really want to spend $15 so I don't have to take the bus" they'll start taking the bus way more. Indeed, I think it's critical not to impose taxes on them to make it affordable to go car free.
One modification I would make to a VMT tax is to factor in a vehicle’s weight. Heavier cars are massively more destructive to our roads, and also tend to be bigger (and therefore less economical as a single-occupancy commuter vehicle). They should not be taxed at the same rate as a sedan or other smaller car.
I’m not sure if the linked NYT article mentions that, but I’m not in a position to break the paywall right now.
One thing missing here: the old. We don't like to give up our licenses. We don't like to drive anywhere any time. We like our routines and almost always drive around our locality.
These factors would seem to suggest: there's a great potential market for selling robottaxis to the old. I suspect the extent of the market depends on the balance between our desire for control versus the possible drop in costs of robotaxis as drivers are eliminate and presumably the number of robotaxis begins to exceed the number of taxis we have had.
I have an honest question that I never see addressed by urbanism advocates/opponents. Is there an urban planning model that doesn't correlate with birth rates falling below replacements levels? Hearing about various demographic crises out there has made me wonder what research exists about urbanism's role in population decline has been.
I don’t believe a definite correlation between urban living and low birth rates has ever been found, as there is just too many confounding variables (education levels and inflated urban rents being primary ones).
I haven’t researched this extensively though, so I’m happy to be corrected if it has.
Birth rates are consistently lower for cities than for rural areas. Towns and small cities have rates that are in the middle between the two.
This has been found pretty consistently, not a one off.
But the point is that this doesn’t define a causal link between urban living and lower birth rates, as living in cities also implies higher education and higher CoL (the latter of which is not a necessary fact of cities, just a truth about our current ones). Both of those could be the cause of lower birth rates
My impression is that birth rate declines happen basically anywhere women have decent rights and birth control is available.
It really is a difficult issue, and I don't know that there's a great solution. Many people suggest improved social services, but Sweden and France are pretty much the world standard for welfare states and have about the same birth rate as the US or Canada.
It's interesting how both self-driving trucking and personal AV ownership are being left out of this discussion. The latter might be a few more years down the road, but could revolutionize suburbia and rural area transportation. The big barrier on the former is probably regulation and laws.
I just got back from a trip visiting family in SF, and took a Waymo for the first time--it was incredible. One of the handful of times in my life experiencing awe at a technological advancement (another one that happened recently: having a Turing-test-passing conversation with chat GPT).
For me it was one of those moments where you realize the world is going to be fundamentally different, and your children won't have a memory of the world as it was before.
As someone who loves trains, metros, and all manner of grade separated fixed guideway transit, it feels to me that its era (at least in the US, limited as it was) is rapidly coming to an end. There are broadly two rational reasons to build high capacity transit today:
1. Geometry, as alluded to in the article: the city is so dense and requires so many people to be moved into such a small area that doing it via surface roadways becomes impossible.
2. Scaling operational costs: building fixed guideway rail massively increases the number of passengers per driver, since vehicles are much longer. So you invest some money now to reduce your labor costs over the long term.
With self-driving technology rapidly becoming mainstream, #2 will soon be dead letter: if an automated bus and an automated train have equivalent operational labor costs (ie roughly zero), why would you ever build the train? And, at least in the American context, there are very few cities dense enough that surface roads with maybe some dedicated bus lanes are insufficient with enough bus service, so #1 is out too. What reason is left, outside of maybe 5-7 truly super dense metros, to build a subway?
I think you're missing the point.
AVs don't have to be 5-seater SUV/truck abominations occupied by usually one person. The networks that run these can be smart. You can run cheap busses that dynamically adapt to demand to do curbside pickup and dropoff (or maybe a slight compromise for routing speed and efficiency, so you're never asked to travel more than a block or so) based on people whose routes happen to overlap. For solo or pair travelers, you can dispatch microcars. These are obvious cost optimizations to do in a competitive market.
The "car as lifestyle signaling / prosthetic genitals" thing completely goes away if the cars aren't yours. Nobody cares what the bus looks like. So vehicles can actually be the appropriate size for the trip.
AVs can give you something that is *systematically better* than bus-based public transit. No sending busses that run mostly empty sometimes to fill out the timetable so people can plan around them. If there's no demand, the network wouldn't send anything. If there was slight demand, it would send a smart car. AV networks could have much higher average vehicle utilization *and* provide an experience more similar to rideshare (hit a button in an app and get where you're going in a reasonable amount of time) that is much better for consumers.
Outside of the content I need to say this is an amazing lede: "In San Francisco, robotaxis are like naked guys. The first time you spot one, it feels like a big moment. After that, they fade into the background."
Great hook.