Hacker News Reader

My reading was different, in that the article uses "elite" for the kinds of decision-makers in positions of power whose exceptional attainment and status makes them some of the least likely to be consumers of the transit anyway.

I think both you and the author would agree that in the absence of a reliable, versatile mass transit network, many people with access to cars will use cars instead. But you and the author appear to disagree on the viability and transformative power of self-driving cars and whether they serve as a meaningful instrument of mobility for all with the kind and amount of transport infrastructure available to us now.

I believe that the form factor of a 2-to-7 person car common today isn't sufficiently dense to provide appropriate amounts of throughput to high traffic routes to completely replace something like a subway, commuter rail, or the like. Although the experience of the (figurative) first and last mile would be superior, the sheer number of such vehicles would clog the arterial roads and freeways where the point-to-point routing converges, the same way this happens with cars today.

But the form factor of the 2-to-7 person car changes significantly with the shift to autonomous electric vehicles.

With computer control, you eliminate the gap needed between cars - just hook it directly to the car in front, let them draft for better aerodynamics, and detach when you need to make a turn, saving 70-100 feet of road space. With electric motors, you eliminate the hood, engine, and transmission, saving another 5 feet or so - independent electric motors can go directly onto the wheels. In an urban environment, electrify the streets (as with trolleys) and eliminate the battery pack - that's another 4 feet and 1200 pounds or so. Once everything moves to car services and people don't own their own cars, you can eliminate the trunk, saving another 4-5 feet - in the majority of trips you aren't carrying anything big, and you can hail a different model pod for those road trips or grocery runs where you need the cargo space (or just summon the groceries and forget about driving it yourself). For trips without the kids, eliminate the other seats as well - you can just summon a different vehicle when you need to take a family trip.

Pretty soon the vehicle is just a thin shell around its occupant, taking up no more room than the personal space they would've taken up on a bus.

This analysis ignores the fact that the two modes (bus and AVs) load and discharge passengers very differently. The problem is not only the physical space AVs take when in motion on the road. See, e.g. https://www.theatlantic.com/technology/archive/2018/10/bus-b...

“How many people’s doors can a driver get to in an hour, including the minute or two that the customer spends grabbing their things and boarding? The intuitively obvious answer is the right one: not very many. An Eno Foundation report promoting microtransit could not cite a case study doing better than four boardings an hour of service. John Urgo, the planner of demand-responsive service for AC Transit in Oakland, California, has said that seven boardings an hour is “the best we hope to achieve.” Few fixed-route buses perform that poorly. Across sprawling Silicon Valley, for example, fixed-route buses carried 12 to 45 people an hour in 2015. In a dense city such as Philadelphia, the number can exceed 80”

Door-to-door transit simply cannot achieve the throughput of buses! Which is not to say door-to-door would not be profitable or even have a role in a mass system. It's probably an improvement over the current dominant mode (although my suspicion is that it's simply a slightly less bad local minimum).

But the point of the original article is to argue that certain groups of people (decision makers in cities, mostly), should not extrapolate from their own needs and preferences when designing a system designed for mass use, because they're by definition unusual and they're each N=1

That whole Atlantic article is predicated on the main cost of a ride being the driver's time. Get rid of the driver and you get rid of that cost. Get rid of the internal combustion engine (which has large economies of scale - one big engine works more efficiently than many small ones) and you lose the fuel advantage - many small electric engines are just as efficient as one big one.

The boarding issue in the article is mostly in reference to the driver's time, but there's a simple solution for the rider: docking, not board. Board a pod once, at your door, and travel inside it until you reach your destination. If you need to switch modalities - like latching a number of pods together into a train for long road trips, or putting them on an aircraft - you load the pods onto the plane, you don't transfer the people between pods. It's like intermodal containers for passengers, with the nice benefit that you can continue working on your laptop or watching a movie or sleeping throughout modality switches.

>With computer control, you eliminate the gap needed between cars - just hook it directly to the car in front, let them draft for better aerodynamics, and detach when you need to make a turn, saving 70-100 feet of road space

You will still need a way to turn off at speed, leaving you with interchanges at each potential detachment point taking up a lot of space.

Grand Central Terminal in NYC has 44 platforms in 48 acres. Tens of Thousands of people per hour can switch between subway, regional rail, and bus lines, as well as access the surrounding city on foot, and access the on-site amenities and offices above. 700,000 people pass through daily. 10,000 people visit each day just to have lunch there. It works because the most space-efficient way to route people is to have them get out and walk. A tiny person-carrying vehicle will always need more space and supporting infrastructure than a person walking.

That may be true if you can eliminate human drivers from the picture. I'd be shocked, though, to see numbers that support this being feasible at that kind of scale on public access roads. And that just brings you back to needing some sort of separate right of way for transit -- sure it's potentially interesting and useful, but it's just another alternative form factor for transit as it exists today.

>> as people who rely on transit in the US are either big city urban elites, or the poor stuck in crappy burbs with shitty mobility.

Clearly we don't ride the same bus. A lot of the urban elites on my bus route are heading downtown to service jobs in hotels, restaurants, and perhaps day-care centers.

More likely you two just don't live in the same city.

FWIW, my home city's bus service (Charlotte NC) is exactly like the commenter above describes. There is a city-wide bus service that runs about every hour or half-hour if you're lucky, but it's terrible (unreliable, slow, poorly maintained) and mostly used by people too poor to own a car or those who can't drive due to disability or DUI.

Then there's the nice "commuter" bus for a buck more, which only runs during peak commute times, only picks up from park-and-rides in the suburbs, and is mostly occupied by people who work nice office jobs in the big high-rises downtown.

The middle-of-the road workers all pretty much have to drive everywhere.

My suburban city is the same and I drive, walk or bike everywhere. Estimated travel time for going by bus here is .. identical to that of walking, according to Google.

But that makes the "transit is for the rich" argument rather, uh, rich. Transit is for the rich ... now. Make a decent bus service and it wouldn't be.

For sure, but it's a matter of investment priorities too. Operating fleets of buses, especially if you need enough to cover a city that sprawls all over like Charlotte does, is expensive. And it's not like much of that expense will be recouped in fares if we want it to be priced reasonably. Meanwhile car ownership is still relatively affordable - even people working on minimum wage usually have a car. And once people are already paying for that car, they aren't going to want to spend extra money on a bus, no matter how nice they make bus service.

The new light rail service though is starting to change things though. I've noticed more young professionals and college students who are open to using the light rail and ride shares for their travels, regardless of whether they already own a car, due to convenience and it's ability to get through traffic faster.

>The article's idea of a "elite" is a person who relies on a car to get around. I find it unintentionally funny, as people who rely on transit in the US are either big city urban elites, or the poor stuck in crappy burbs with shitty mobility.

Well, if we look at an example, like the NYC metro area, we see that car-owners have a higher annual income on average than non-car-owners, regardless of whether they live in Manhattan, the outer boroughs, or the suburbs:

http://www.tstc.org/reports/cpfactsheets.php

"Basically, everything should be a battery skateboard riding on asphalt by way of rubber." What a wildly narrow view on transportation.

Latency and efficiency matter. Take two extremes. Air vs rail. Air (chopper) is wildly inefficient (maintenance and energy). Rail has twice the efficiency of a hybrid car.

If you make everything a "car" you end up with a world where the ambulance and the big rig are both stuck in rush hour traffic.

Dedicated lanes help for some things (such as buses) for considerably cheaper than rail.

I have seen these going up in Toronto. What do you mean by cheaper?

Maintenance of vehicle? Doubt it. Especially if its an electric train.

Land acquisition for lane? Doubt it. The dedicated lanes also have full station stops.

Energy usage? Doubt it. Electric trains dont have round trip energy loss in batteries.

Operations? Doubt it. Its an electric train is usually autonomous.

Flexibility? Ya its far more flexible than a train that requires rails. Also you can reroute buses.

I was thinking cheaper capital costs, but I'm not an expert.

The article uses transit as an example, as that is the area of expertise of the author. It is tangential at best to the idea of 'elite projection,' and yet it is far easier to make arguments about transit than to consider whether we are looking at problems from an 'elitist' POV.

¯\_(ツ)_/¯

I don't understand why you are suggesting that autonomous cars will make buses and trams obsolete. They are the closest to autonomous vehicles out of everything on your list (they run around on meta-organized and efficient routes, you don't drive them yourself, etc).

They have the huge disadvantage of not always taking you to your exact destination and not doing so on your schedule. Neither is really necessary in the world of computerized autonomous transport and small electric motors, since computerized transport could easily get you from your destination to a moving transporter without it having to slow down.

It would need to slow down due to traffic.

I'm totally sympathetic to the cause. But I also agree with the author, we can't ignore the fundamental of transit, or we will just build something even worse.

Assuming everyone could have a single passenger autonomous vehicle, ideally as small as possible, while still spacious enough to satisfy our elitism desires. And it were to pick you up and drop you off at your exact destinations.

How would traffic jams not be as huge an issue as they are today? If not more, assuming you got rid of mass transit?

Are we suggesting to massively increase the size of roads, or tunnel roads and layer them, but still, there's gonna be point of contention within the network, causing traffic again.

I feel that's pretty much the author's claim. That unless a radical idea to this problem is figured out, which I guess hasn't, then its just foolish to think we can all be picked up and dropped off wherever, riding solo at that. It won't scale. That's why its elites projection. From the problem itself, you can show it can only work for a small number of individual. But if you are that individual, you might think, I can just give this to everyone else, and problem will be solved. Forgetting the main purpose of why mass transit exists in the first place. Which isn't affordability of a car. But scale of efficiently moving large number of people around in limited space.

You're still assuming a human driver. Most of the carrying capacity of roads is dictated by human reaction times.

A human takes 150-250ms to perceive a change in the environment and react to it. A car travels about 15 feet/second for every 10 mph of speed. Hence the rule of thumb to stay one car length apart for every 10 mph of speed; that gives you about a second of reaction time, a safety factor of 4. It would be suicidal to drive one foot away from the car in front of you at 60 mph, as a human.

It takes a computer maybe 10ms to detect and react to a change in front of it, and that's assuming fancy computer-vision algorithms (it's much easier if you just put a laser beacon on all vehicles that rapidly flashes an encoded representation of the vehicle's ID, velocity, and acceleration). And computers are much better at measuring velocity, acceleration, and position with the help of a LIDAR or doppler radar. In 10ms, the vehicle will move roughly 2 inches per 10mph of speed. Suddenly it's reasonable to drive 60mph with 2-3 feet of separation between vehicles.

At that range, though, you might as well just put down a hook and latch the vehicles together. And I think that's what the OP means by cars <=> buses <=> trams <=> trains existing on a continuum, and autonomous vehicles making this clear. We'll likely get something that works like a scooter for last-mile transit, and works like a train for long-distance trips. As it enters a main road, the computers will automatically adjust speed so that they can "dock", and we'll fit an order of magnitude more vehicles onto the road.

Such an idea sounds like it would be pretty easy to simulate, and know its capacity. I'm still not convinced until I see data. You're optimizing the road network to its maximum, but that might still not be enough to prevent traffic given a point to point system. But, maybe it is, I think we'd need to see data. Also, I'm not sure all the assumptions make sense, I'd be surprised if safety was all about reaction time, there seems to be an upper cap here, where even if you had instantaneous reaction times, you still need a good distance between the vehicles to avoid a collision.

They are, which is why I'm suggesting that instead of thinking of "transit vs private cars" everything will essentially melt together. IE, if you take a bus with 30 riders today and replace it with 5 AV-minivans each with 6 occupants driving behind each other, is it any different? What if instead of the city, I own one of these minivans? What if I pay 6 fares and have the minivan take my family to somewhere? What if 6 passengers split a minivan, but pay extra to use a dedicated lane?

>IE, if you take a bus with 30 riders today and replace it with 5 AV-minivans each with 6 occupants driving behind each other, is it any different?

Yeah, it is different. You have five engines instead of one, and 20 tires instead of four, meaning maintenance and fuel costs are some multiple higher, and wear-and-tear on the road surface is higher, noise and air pollution are higher. Among other things.

>What if instead of the city, I own one of these minivans?

Then you have to go through an additional middleman, the city, to find parking for your 1/5-of-a-bus and permission to drive on certain roads, in certain lanes, etc. Among other things.

When I last did some armchair thinking about transport modes my conclusion was only a little bit removed from parent's basic hypothesis, just with one critical detail: It's not that one mode will rule them all, it's that the friction of switching modes is likely to go further and further down until we have successfully "containerized" or "packet switched" transit.

So, instead of envisioning it as "5 minivans vs 1 bus", it's a system that smoothly transfers you from street to car to bus to train to plane and back. You get the "now" of low latency, and then you transfer up a step to get speed and efficiency.

Right now we have very primitive notions of multi-mode: every transfer is time consuming and difficult on the psyche, often involving new fare systems, security checks, navigating transfer points, delays, tolls, parking, and so on. This is a big factor in our tendency to bias towards a single mode and overbuild infrastructure for that mode.

Portable electrics are a big step forward in moving up the baseline: they're roughly person-sized without too much additional weight. An imagined standard for docking them in a larger vehicle or running them in a convoy is not too far-fetched, but will probably still take a lifetime to be realized with all the involved political factors.

But while ganging or docking is nice, you don't really need it to get most of the benefits; low-friction transfers by eliminating the huge parking lots that you half to walk from at the interface of small personal last-mile (or last-few-miles) transport with larger long-distance transport gets you much of the way to the kind of network you seem to want, and only really takes autonomous on-demand vehicles and integrated payment and reservation systems, not ganging/docking. You've got some buffering at transfer points still, but much less friction than in the status quo.

>eliminating the huge parking lots that you half to walk from at the interface of small personal last-mile

We don't need autonomous vehicles to do that. Valet parking exists.

> We don't need autonomous vehicles to do that.

Yes, we do.

> Valet parking exists.

Valet parking is labor intensive, doesn't scale on-demand easily, (because it is labor intensive), ands a lot more latency on both ends than autonomous (or chauffered, but that's also labor intensive) drop off/pickup even if the labor is provisioned optimally (and gets worse very fast with slightly underprovisioned labor.)

It's decent for providing an up-charged lower-latency option for an elite, higher paying subset of customers, or optimizing parking density with less concern for latency for a larger customer base, but it's not nearly as good as on-demand last-mile transport for reducing endpoint transfer latency.

Your objections are actually based on ICE car assumptions. ICE cars are going away soon, they're irrelevant when talking about the future of transportation.

There are few, if any efficiency gains in one vs many electric motors. They don't produce noise or air pollution. Wear and tear on road surface is actually much worse for heavy busses than cars, they actually deform the asphalt and can cause accidents: https://www.thestar.com/news/2007/07/05/road_ridges_made_by_...

> Then you have to go through an additional middleman, the city, to find parking for your 1/5-of-a-bus and permission to drive on certain roads, in certain lanes, etc. Among other things.

I already have parking for my 1/n-th of a bus. Running an exchange would be a much more valuable use of a city's resources rather than operating bus depots, etc.

> You have five engines instead of one, and 20 tires instead of four, meaning maintenance and fuel costs are some multiple higher, and wear-and-tear on the road surface is higher, noise and air pollution are higher.

Road wear and tear per vehicle-mile scales with something like the fourth power of vehicle weight, so many small vehicles actually have less impact that one large vehicle even with the fact that they tend to be heavier per passenger.

OTOH, land use impacts are higher since you need more lane capacity for many small vehicles than one large one.

Read the authors other article: https://www.citylab.com/transportation/2017/12/what-elon-mus...

He makes the argument the reason that doesn't work is geometry and utilisation of limited city space.

Autonomous vehicles don't solve the problem of there not actually being enough space on the road (a bus takes up less space than a fleet of cars with the same number of passengers). See another article by the same author: https://humantransit.org/2016/07/elon-musk-doesnt-understand...

> as people who rely on transit in the US are either big city urban elites, or the poor stuck in crappy burbs with shitty mobility

What city are you talking about?

Los Angeles?

> I find it unintentionally funny, as people who rely on transit in the US are either big city urban elites, or the poor stuck in crappy burbs with shitty mobility.

In cities such as Amsterdam, Utrecht, Aachen, and various others it is very difficult or expensive to park your car. Using your car in such cities is therefore elitist, but because most people just use bicycle and public transport it is of high quality (due to high demand, becoming a circular reasoning).

I'd love to see autonomous "pods" with docking behavior. You'd take one of the appropriate size, i.e. there could be scooter-sized pods for the solo commuter, cart-sized pods for when you've got luggage or groceries, family pods, sleeper pods for long trips. They would drive autonomously for the last mile (perhaps even being able to dock into elevators and take you directly to your destination), but could also merge onto arterials and dock with larger pod-carriers, the modern equivalent of buses. Thus, you'd maintain your privacy and last-mile flexibility, but still get the economies of scale of better aerodynamics and powerful engines for longer trips. You could even dock onto airplanes, i.e. instead of needing to drive an hour to the airport and wait 2 hours at the gate, your pods arrives at your apartment door an hour beforehand, whizzes you through town, docks into a bus for the trip to the airport, and rolls you onto the tarmac and into the plane. No need to manage baggage (it all comes with you), the pod scans you for security during the commute, and you maintain your privacy on the plane.

I find it unintentionally funny, as people who rely on transit in the US are either big city urban elites, or the poor stuck in crappy burbs with shitty mobility. Thus, "more transit" is itself the elite projection, as it supposes that what's good for the urban elite, must be good for everyone else.

This is kind of bizarre reasoning. Suppose there's mass starvation in a nation and I were to say "I find it unintentionally funny to call for more food, as people who are able to eat regularly are the urban elites. Thus, 'more food' is itself the elite projection, as it supposes that what's good for the urban elite, must be good for everyone else."

Sure, suburban transit is terrible but that's argument for "transit only matters to people who already have."

Busses as we know them today will become obsolete, as they are functionally equivalent to a heterogenous mix of autonomous vehicles driving in the same direction.

"My imaginary solution is obviously better than something that's been tried and with proper planning and funding work now".

> I find it unintentionally funny, as people who rely on transit in the US are either big city urban elites, or the poor stuck in crappy burbs with shitty mobility

So everyone not living in a big city center in the US are now considered the poor. And everyone living in a big city center with a car is now the Elite?

>> as people who rely on transit in the US are either big city urban elites, or the poor stuck in crappy burbs with shitty mobility.

Clearly we don't ride the same bus. A lot of the urban elites on my bus route are heading down to service jobs in hotels, restaurants, and perhaps day-care centers.

There are a lot of places where trains (above or below ground) and trams will still make sense. It's going to be nearly impossible to automate past the fundamental limits of street space in cities like New York for automated cars to surpass the sheer volume of people serviced by the subway.

Well, you always have to take in account sunk costs. Places like NY already have tunnels and trans and tracks, so it will likely never make economic sense to replace those with something new.

But think about what a tram is - typically, several linked cars attached to a wire running on tracks. Why are they linked? Because you need a driver to keep costs low, otherwise you'd unlink those cars and have more frequent service. Why are they attached to a wire? Because batteries are expensive. Why are they on a track? A bit cheaper, but in many places (Lyon, Montreal) they are not.

What is changing is the price of battery and soon, the price of driver. Trams are built on assumptions that are changing, so would they not as well?

Yeah they certainly benefit from already having the infrastructure already in place, but most large cities are only going to get more dense as time goes on and adding additional roads to urban cores is always going to be very hard. After a certain point the only way to provide enough transportation is to go underground just to avoid tearing up the existing city.

"Busses as we know them today will become obsolete, as they are functionally equivalent to a heterogenous mix of autonomous vehicles driving in the same direction."

This is not an opinion any experts on mass transit share (including those who are bullish on autonomous vehicles).

For an outline of just one of the arguments why this is a naive view (at best!), please read the author's "The Bus Is Still The Best" (https://www.theatlantic.com/technology/archive/2018/10/bus-b...) and an interview with him in the Financial Times "A conversation about how public transport really works" (https://ftalphaville.ft.com/2018/01/26/2198114/a-conversatio..., free FT account required, but not paywalled).

Some relevant quotes from those articles:

“How many people’s doors can a driver get to in an hour, including the minute or two that the customer spends grabbing their things and boarding? The intuitively obvious answer is the right one: not very many. An Eno Foundation report promoting microtransit could not cite a case study doing better than four boardings an hour of service. John Urgo, the planner of demand-responsive service for AC Transit in Oakland, California, has said that seven boardings an hour is “the best we hope to achieve.” Few fixed-route buses perform that poorly. Across sprawling Silicon Valley, for example, fixed-route buses carried 12 to 45 people an hour in 2015. In a dense city such as Philadelphia, the number can exceed 80”

“No, of course, the driverless car people will say, no, cars will fit closer together and they’ll be smaller and so we’ll fit more of them over the bridge but that’s a linear solution to an exponential problem. The other dimension of this problem that you must keep in mind is the problem of what we, in the business, call induced demand. And induced demand is the very simply idea that when you make something easier, people are more likely to do it and this is why, for example, when you widen a motorway, the traffic gets worse or it fills up to the same level of congestion that you had before. It’s because when you actually create new capacity, people use the capacity and you end up back in the same point.”