Hacker News Reader

Comment by Maarten88 | Hacker News Reader

I attended a presentation by Travis Kalanick recently, where he told he got the idea for Uber when he could not get a Cab while visiting Paris.

I thought that was quite funny: Paris has this awesome ridesharing service, that is even mostly driverless, called the Subway. Everyone there uses it; only a US tourist would use a taxi there. Still that poor idea led to a great company.

On a more serious note, this made me think: private companies and startups doing self driving cars and other transport innovations gain nothing when public transport gets more popular. They compete with it. I now wonder if that has something to do with the decrease in rides in US cities outside NY.

View on HN

In theory, they could coexist. There was some talk that Uber was going to create a "hybrid" transportation solution, where it would handle the beginning and end of your transit, dropping you off at a train for the majority of the trip and automatically picking you up at the other end for the last leg.

Only Uber could re-brand the bus system as an 'innovation'.

A bus coming directly to my house, with no intermediate stops on its optimal route? Yes, please. I'm only sure it's not going to be $2.75 flat rate.

It will happen within 10 years somewhere, and it will probably not be a flat rate, but most rides will be less than $2.75. The per-mile cost for self driving electric vehicles is very low.

That's an extremely aggressive presumption to make without a source/explanation.

Not really. It already happened in some countries where with high development but still low salaries, like big cities in China. How/if cities will be able to deal with all the increase in traffic from self-driving cars without significant congestion charges is of course still unclear.

I think self driving cars and public transportation could ultimately have a symbiotic relationship. If self driving cars result in significantly decreased car ownership, there will be plenty of traditional car trips replaced by the self driving variety, but there will also also be traditional car trips replaced by public transportation.

If I live 20 minutes outside the city, but not within close walking distance to a transit stop, I am currently forced to drive to my destination, even if it is located somewhere along the public transportation network. With a self driving car, I could be taken to the closest transit stop, and then ride the subway to my destination to avoid congestion.

Berlin's rail system is so good I never even considered a taxi or uber.

> Everyone there uses it; only a US tourist would use a taxi there.

Or someone wanting a ride after, say, 12:30AM ..

Even in cities with good public transportation an autonomous vehicle will be faster than public transportation. Especially when you don't have to park.

Gridlocked Manhattan is an exception. But for most trips in most places a car will be faster. Especially if public transportation requires switching one or more lines.

Be prepared to be downmodded by HN'ers who seem to think that the entire purpose of municipal transit is to move the largest possible number of people per hour from one train station where they don't want to be to another train station where they also don't want to be.

Of course autonomous vehicles are going to render fixed-rail systems irrelevant... or at least, they will if they're implemented properly. "Gridlocked Manhattan" is only gridlocked because the vehicles doing the gridlocking aren't talking to a central authority, or to each other.

But what seems like common sense to me has proven to be a surprisingly hard sell among technologically-literate people who I'd expect to understand it intuitively. People on HN really, really like trains.

> Of course autonomous vehicles are going to render fixed-rail systems irrelevant...

I'm not sure everyone is willing to take that as a statement of fact. I'm sure not.

Can you point to a single transportation expert who believes that centrally orchestrated single-occupancy vehicles are a solution for transportation capacity problems?

The difference in lane throughput between single occupancy vehicles and mass transit is staggering. Part of the benefit of rail is that it can use dedicated facilities that are deconflicted. Autonomous vehicles still need to deal with red lights, pedestrians, cyclists, and limits on safe acceleration/deceleration.

Here in Seattle our light rail system is hobbled because of non-grade-separated parts of the line - which despite having signal priority is restricted in its operation (station dwell times, # of trains in operation, train separation) because of conflicting cross streets. A central authority is NOT a panacea.

With human drivers, lane density tops out at somewhere around ~900 vehicles per hour. Let's say automatic cars can somehow double that - to 1800 vehicles per hour. Even dedicated bus lanes can move 6000 people per hour.

Forget Manhattan which is a beast of its own scale - look at Chicago mass transit which moves over a million people a day. You're sure you could replace the capacity of rush-hour Red Line trains at crush capacity with the surface streets? Really? Where would you park these cars? Where would everyone get in and out of the cars? (On/off boarding is the scaling limit for bus lanes - that alone is enough to sink your plan.)

You do realize the irony in having faith that centralized coordination between autonomous vehicles will lead to these improvements?

Even dedicated bus lanes can move 6000 people per hour.

But not where and when they want. Passengers-per-hour figures are simply measuring the wrong thing.

Considering that there won't need to be as many cars, parking won't be a huge issue. Cars that spend most of the day and all night sitting in parking spaces are a waste of both resources and real estate. In an optimal scenario there wouldn't be a need for any parking spaces at all.

Here's a thought experiment for you: take all the taxicabs off the streets of NYC, and replace them with privately-owned cars. Hell of a mess, isn't it? Nowhere to park between Boston and Philly. Now go the other direction, and imagine that every single passenger car that isn't a taxicab simply isn't there at all. It should be obvious that there would be plenty of room in that scenario, both on the streets and in the parking garages.

Every single objection you raise is an example of "fighting the last war." You're pitting the worst aspects of the status quo against the best aspects of your favored solution. This happens time and again when these threads come up.

Can you point to a single transportation expert who believes that centrally orchestrated single-occupancy vehicles are a solution for transportation capacity problems?

No, but I'm less than impressed with the "transportation experts" I've encountered around here (Seattle area as well.) It seems self-evident that planners who devise fixed-rail systems with 25-35 year construction horizons while not even considering the arrival of autonomous vehicles are not competent at their jobs.

> It should be obvious that there would be plenty of room in that scenario, both on the streets and in the parking garages.

No! It's absolutely not obvious. My biggest problem with your argument is that you're taking it for granted that streets can move as many cars as you think they can. They can't. For the sake of argument, I'm willing to entertain the idea that you can double throughput through automation, and that simply isn't enough for such a low-density form of transportation.

Again - picture your system in Chicago where it moves 300,000 people into the downtown loop during the commuting rush-hour.

Nobody owns any of these cars, so there isn't any parking to worry about. Fine.

Where do these 300,000 people get in and out of their cars? Remember - bus lanes can't even scale past 6,000 people an hour per lane because of the need to get people on and off.

Don't forget that these throughput numbers are under best-case scenarios, for smooth flowing traffic. We haven't even discussed intersections...

You're using present-day definitions of words like "bus" and "car" as if they're still going to hold in 20-30 years. I'm not sure how that fits into the accepted taxonomy of logical fallacies, but it's in there somewhere.

Hint: An ensemble of coordinated autonomous vehicles is a train. (And no, there's no contradiction there -- you behave in a locally-autonomous manner when you drive your car, but your actions are still centrally coordinated by the traffic laws you have to obey, locally coordinated by the network of traffic control devices that your city maintains, and ultimately constrained by the other vehicles and pedestrians around you.)

What's the difference between a light rail system and a chain of buses that can drive at high speed, nose-to-tail, along any highway in the city, capable of starting and stopping like a train to let people on and off but also capable of arbitrarily diverging from and rejoining individual routes like passenger cars? The only real difference is that the latter is going to be better at meeting its users' needs, because unlike rails, roads are everywhere. How cool will it be when your city can grow and reshape itself organically, instead of being tied literally and metaphorically to some existing rail infrastructure that was installed decades ago?

The notion that there have to be some metal rails bolted to the right of way is just something some people made up 100 years ago. But damn, it's influential for some reason.

I can't point to a single transportation expert period. Nor can I point to a single city-scale simulation of autonomous vehicles.

I wish I could. Because I too live in Seattle. And before we decide to spent 30 _billion_ dollars on light rail over 30 years I'd really like to know that's the best way. What if we instead spent 30 billion adjusting infrastructure to support autonomous vehicles? What would those adjustments even look like? God I wish I knew. I'm pretty sure Uber/Google would even build the study for us for free.

I live a mile and a half from a light rail station. In another 6 years I'll live 3/4 mile from a light rail station. And when that day comes it will STILL take twice as long for me to walk and rail to work than to drive. It takes three times as long by bus. And that's with me working right beside a transit center. God help me if it required a second transfer or a walk on the other end.

Reply on news.ycombinator.com