And they pollute much more and take more space.
I'm also a daily user of public transport in a major world city, however, and whilst buses and trains work they also have the fundamental problems highlighted above that are in my view insurmountable, and which (self driving) cars solve. So I'm in a dilemma over whether I think they'd be an improvement, but I ultimately believe they would be.
But yes, absolutely pollution and space saving would have to be addressed. I'm optimistic they could be:
Pollution: hopefully solvable with electric. Yes, of course, there are efficiency gains that trains/buses would perhaps always have (having to move many small things rather than one big thing carrying the same number of people) but I can see electric reducing the absolute level of pollution from cars to a point where it basically doesn't matter. Maybe overly optimistic, but I hope that'll happen.
Space: Driverless cars, particularly electric driverless cars, can be much smaller and can drive much closer together. Different cities have different challenges, but I don't see why we couldn't get to a situation where roads might effectively have capacity for 2x vehicles across and 2-3x vehicles along if the vehicles can be coordinated to drive closer together and are physically smaller (remember that on-demand unlocks certain design constraints - you don't have to provision for peak - if the average number of passengers is 1.5, then most cars on the road could be ultra small 2 seaters). Again lots of conjecture and optimism here, and I am probably assuming that 'manual' cars would eventually disappear or be banned at least from cities too, but nonetheless I think there is hope here.
There is a minimum amount of space occupied by the vehicle's own matter. The ratio of that material to the occupant is a scaling law: the larger the vehicle, the higher the volume-to-surface, the denser the passenger-to-vehicle ratio can be.
A standard R160 subway car carries up to about 250 people in an area that would be about the same as 5 Corollas positioned nose to tail, a starting ratio of about 10:1. If cars are made twice as dense, it's still 5:1.
This is ordinary batching. The only thing that can carry more people more densely than a bus or train is the footpath.
It's also important to point out that of course, buses and trains shouldn't go away - they could indeed also have a bunch of problems solved by being self driving too, most notably the 24/7 running thing - they'd just be another option in the mix.
While that's true, that's also just an argument for something we already understand. Are there possible arguments to be made about the number of busses and routes because it's not taking people point to point. What about the lul-time of day, where on-demand model is more efficient than driving empty busses around. They reduce schedules obviously. But now you're punishing passengers that start a shift at noon. There are more things to solve than just "the least amount of pollution and space."
Totally unrelated but similar argument. The FAA introduced a new system and process for creating standard flight paths a few years ago. Flights around the nation started doing new things. TONS of people are now inundated by jet noise. They're now discussing just going back in certain areas to the old routes. They ONLY optimized for efficiency of route and didn't consider anything else.
The argument I was responding to was that self-driving cars would be sufficiently more efficient users of space that it would obviate the need for public transport. This was demonstrably false.
> But now you're punishing passengers that start a shift at noon.
Sure, but (a) that's a financial question, not a capacity question and (b) rush hour is when both finance and capacity collide.
Public transport is a more efficient option per passenger-kilometre and per square metre of road or cubic metre of tunnel than self-driving cars. It is always going to be. Self-driving cars will have other virtues and will reshape transportation, but ascribing magical powers to them does nobody any good, especially if it leads to defunding of subways and buses.
You're not accounting for the space between trains. They will only take up 5-10% of the available rail space because you need gaps between the trains.
At highway speeds, 4 seconds of following distance is 120 meters.
NYC's least efficient system is block train control, requiring 300M separation between trains.
In 300M you can fit about 66 Corollas, again packed bumper to bumper. That's 66 * 5 people = 330 people, or approximately 1.5 subway cars.
Subway trains are 6 or 8 subway cars long, depending on track and time of day.
With CTBC some platforms can run at 60 trains per hour (30 local, 30 express). That's 8 subway cars every 2 minutes, or approximately 1,000 people per minute.
To move 1,000 people at all with the Corollas you will need 250 * 4.55M of them, or approximately 1.1 kilometres of cars, bumper to bumper.
Unless your trains are going at 10 kph you will have more of a gap than 300 meters if you see a train every 2 minutes.
You should recognise this argument: it's the same logic given for why self-driving will solve everything.
If you're deploying a fleet of robotaxis, most of your fleet can be vehicles that only carry one or two passengers at a time on short trips. If a family wants a ride to the airport, then they'll specify that they need a larger vehicle for more passengers and luggage, etc. So you can buy a fleet of mostly smaller cars with a shorter range, and compliment them with a few minivans for long range trips. Overall your fleet will be quite efficient.
Let's say, there is an object completely blocking the middle lane, and partially blocking the two outer lanes. What will the driverless cars decide to do?
My guess, is they wait there until the road is completely unblocked, there's no way for the cars to understand the proper course of action. It will be great fun to bring absolute gridlock with an empty cardboard box in downtown NYC once this all kicks off. Naturally, by then, card board will be illegal there.
How do you make this work? There's no standard in self-driving systems, no standard in sensors, and no standard for vehicle to vehicle communication.
I've been on double-length public buses where I was the only passenger for most of the journey. And many times where there were only 1 or 2 other people. This is necessary e.g. at night where a bus only runs every 30 minutes, there has to be a bus, but there just aren't that many people.
Obviously in this case individual cars would pollute much less, and even take up less space.
Any realistic solution for a city is going to combine self-driving cars for the long tail of origins/destinations and times, with public transit for the busiest routes at the busiest times, including links between the two.
I'd love to be able to take a self-driving car from a house in Queens directly to the subway, the subway to Brooklyn, and then another self-driving car to my final destination that's 20 blocks away from the subway.
The buses need to run at night so the people can rely on them during the day -- it's just the price of public transit, the public uses transit way more if buses run on 10-minute interval during off-peak rather than 30-minut interval. (another study I can't find right now).
I think as programmers we tend to try and optimize problem for least pollution, but the problem here lies in the domains of sociology and urban planning; we might not be the most qualified people to solve it.
For example if buses pollute 3% because of the extra buses on off-peak routes but now more people use them, this would have net effect of reducing pollution.
All your points are correct, but you're not responding to the point that SDCs can be superior to buses for those low-capacity times. If it were one or the other, then yes, buses could be less pollution on average, but that doesn't address the case the parent was talking about.
To the extend low cost ride sharing is a short term unicorn phenomena, that will again become true, although self-contained electric buses might change that a bit. Maybe someday self-driving cars will change this, but that's not in the predictable, foreseeable future.
It's just a fascinating optimisation problem. I think the answer is far from clear, and certainly couldn't be reliably approximated without lots of real world data.