HN Reader

I like this; it's smart. It's a low tech solution that simply coordinates transit based on demand and self optimizes to serve that demand.

The value of buses and trains running on schedule is mainly that you can plan around it. But what if transit worked like Uber. Some vehicle shows up to pick you up. It might drop you off somewhere to switch vehicles and some other vehicle shows up to do that. All the way to your destination (as opposed to a mile away from there). As long as the journey time is predictable and reasonable, people would be pretty happy with that.

This is an interesting experiment, but I have my doubts about its effectiveness. I hope someone is tracking how well this works and we get some good research out of it. I'm much more interested in any research paper that comes from this than anything else.

The fact that it can go from proposal to route-in-service in just a few days is impressive

Sounds like something we tried in Helsinki metropolitan area 10-15 years ago. I think it was shuttered due to low demand. Existing paths were already following population density, so there already was maximum availability for bus users.

Where I think it is most in use as a separate program is picking up elderly people. Retirement homes have minubuses picking up people and driving them to centrew and back. The users don’t have to abide by a busier standard bus schedule and the bus is more accessible by the elderly.

China is the only modern country that has both the capability and the lack of bureaucracy to just do things like this. It's simultaneously amazing to see and a depressing reminder of how badly western societies are crippled by rules of their own making. It would take years to make a single new bus route in any city, I don't think I've ever even seen that happen.

Funny to see this on the front page two entries below the story about Uber essentially introducing privatized bus lines. Imagine if America took a fraction of the energy and money we spend on the military and actually put it towards quality of life improvements like this.

take this only as a grain of salt.

It has been tried in many cities before like Beijing, Qingdao, Dalian, Hangzhou and Chengdu.

It wasn't a bad idea, it's just a good route gradually became a fixed route.

This is really brilliant — like desire paths, but for transit. Obviously execution will be challenging, but the concept is fantastic, and China/Shanghai seems like one of the few places with the requisite density & state capacity to actually make this work.

Generally I think that the design of public spaces has SO MUCH room to be improved by just responding to the wisdom of the crowd.

Last year Shanghai celebrated the 100th anniversary of the bus system, so they decorated all of the bus liveries to be a modern take on the historical first busses. They are very cute and easy to use, and a lot of the bus stops have little old LCD displays showing how far away the next bus is.

Everything about this makes sense. I've long-called for similar efforts in America. It's painful to watch my local transit agency (Portland) expend so much money on figuring out transit routes. Endless committees, focus groups, etc, when the state has access to a smartphone-enabled population. Just release an app that lets people request routes and then self-optimize to maximize the number of people willing to take transit. This is not a hard computational problem, but instead we end up with endless committees and bureaucracy.

I'm glad that Shanghai has moved to the next level in public transportation in meeting customer demand. Most cities don't have the funds to buy smallish buses and labour available as drivers. They don't have the money or willpower to get frequencies to turn up and go levels (ie frequent) and leave people with long walks to widely spaced routes.

One of the biggest caveats of citizen participation programs like this is that, surprisingly, there's a subset of people who don't want to participate in the hassle and simply want to be served quickly. It really depends on how the majority of people in a specific area think about civic participation.

Sounds like minibus in Hong Kong with extra steps - we have been doing this since eternity. Driver just ask where people would stop in advance, sometimes an entire area would be skipped if no one goes there

Here in South Africa, we have "Taxis", which are individually owned (to a degree) minibuses crammed full of people. Routes are whatever maximises earning potential for the driver, so it is a kind of bottom up solution in a sense.

It is a violent cartel, so certainly not a good thing across the board, but it's just an interesting variant.

The moia service in Hamburg Germany offers virtual stops which is the next step I would argue. The bus follows a different route and stops every time based on the need of current passengers

https://www.hvv-switch.de/en/faq/what-are-virtual-stops/

this is great. hope beijing will adopted this soon

Google already has the daily trip data on a huge percentage of people and could just create and recommend bus and transit routes and times based on people’s existing commutes. Sure privacy issues exist for allowing them to do this but people have given up more personal information for less benefit.

Chiming in from Los Angeles, USA to say wow, must be nice living in a modern society that prioritizes public transit and peoples' ease of movement. I know, I know, it comes with trade offs of living in an authoritarian state, but the absolute abysmal state of infrastructure in this country is maddening. Ever been on a train in Denmark or Japan or Switzerland?

Tangent:

I’ve often thought that it would be great to let people design their own political districts to reduce gerrymandering

At the polling place you’d get a map with your census tract and then be asked “which two or three adjacent tracts are most similar to your community”. Eventually you’d end up with some sort of gram matrix for tract-to-tract affinity, and then you could apply some algorithmic segmentation.

Two problems:

- this is far too complex for most voters to understand, much less trust, what’s happening

- the fact it’s “algorithmic” would give a sheen of pseudo objectivity, but the selection of the actual algorithm would still allow political infouence over boundaries

Taking this even further would be to autonomous dynamically rerouting minibuses:

- you have app, and you enter destination

- optimal minibus reroutes itself to pick you up and take you there with mix of walking, while dropping off other passengers too

- minimizes the door to door time that makes cars so optimal

Train/bus services change every year here in Switzerland, but based on usage data rather than voting, which seems like it could be gamed.

It's a cool idea, but this seems like the kind of thing that will have the users drift away and cease functioning.

People will log into vote on their route when they want one, and then have essentially no reason to ever access the feature ever again. With no active users, there will be no way to get "votes" for a route.

This has been tried in some European countries in the early 2000s, website not app.

People stop using it. Forget to cancel, unreliable service, took too long. As users drop wait times become longer, cascading failure.

Solution was real time dynamic rerouting and bus stop buttons to request the bus. But by then it was no longer wanted and canned.

Busses need a rethink. There needs to a TGV like central hub and spoke fast travel version, with large capacity. And there needs to a a "on demand, collect people to the spoke" mini-bus service. And then there is no - as in "NOOO" option, for any local politician, to make the speed-bus stop at any location else, that is not directly on route and at least 5 kms apart. And the speed bus can not be allowed to be stuck in traffic, so obviously bus lanes it is.

This remind me that road router should be walked by passenger rather than designed by designers.

This is really cool. It's easy to take the number of stops that you think you can support on a given route for the desired schedule and then ensure that you end up with the fewest stops, on average equidistant between the homes of those who use said stops, all based on passenger demand.

I'm still keen to see automated trains in the future - such as the toob in London. We've had the technology for a long time now. So why isn't it possible that trains are routed on a per-carriage basis? If a camera + ML vision sees n people on the platform at 3am, it should be able to dynamically route and link carriages to serve the right number of people + some margin.

It saddens me that such cool/convenient things have been possible for years but that we have no interest in pursuing them since we're all in competition with each other to try to become billionaires. And the real billionaires aren't really interested in making less money.

This kinda solution wont work in India. People will use relatives phones to vote for the route and get the route approved, but in reality there will be only one passenger

Scale down the number of seats one notch, and increase the flexibility fully and you've got self-driving vans.

How does this work in practice? Say someone wants to take a bus to the hospital. But not enough people want to go to the hospital. Will the bus not run and will you be shit out of luck?