Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[10ebbor10]

It's a bit different for me. But yeah, I suppose my view on public transport is skewed as I get to use busses and streetcars for free.


*Belgium gives out free bus passes to large swathes of the population. Thanks to a mild variant of Autism, I fall in one of those.

[forsaken1111]

It's a bit different for me. But yeah, I suppose my view on public transport is skewed as I get to use busses and streetcars for free.


*Belgium gives out free bus passes to large swathes of the population. Thanks to a mild variant of Autism, I fall in one of those.

So let's say that this auto-taxi service arrives in your town, wherever that may be. This service will send a taxi to pick you up at your door and drive you anywhere for free.

Would you still use the bus/subway instead?

[GlyphGryph]

Unless it's significantly faster than driving, I would still use the subway. Not the bus. I find the bus system to be incredibly confusing, despite being a lot simpler now than it was even a few years ago thanks to technology, and it's still slow and unpleasant.

Subway is great for getting downtown quickly though.

[10ebbor10]

It's a bit different for me. But yeah, I suppose my view on public transport is skewed as I get to use busses and streetcars for free.


*Belgium gives out free bus passes to large swathes of the population. Thanks to a mild variant of Autism, I fall in one of those.

So let's say that this auto-taxi service arrives in your town, wherever that may be. This service will send a taxi to pick you up at your door and drive you anywhere for free.

Would you still use the bus/subway instead?

Yes. Mostly because I'm stubborn like that, and I greatly favor routine.

But anyway, you're making a very flawed comparison. This taxi service will not go everywhere, from everywhere, and won't be free. Not in Belgium anyway. Unions are way to strong for that to happen. Hell, Uber got itself almost booted out of the country (Taxi services have to follow certain regulations.)

[GlyphGryph]

The free part I'll give you, but I don't see why it wouldn't go to/from everywhere cars already go, at least.

[DJ]

If I'm going to a densely packed place, I'll take the subway because it'd be much faster.

[forsaken1111]

It's a bit different for me. But yeah, I suppose my view on public transport is skewed as I get to use busses and streetcars for free.


*Belgium gives out free bus passes to large swathes of the population. Thanks to a mild variant of Autism, I fall in one of those.

So let's say that this auto-taxi service arrives in your town, wherever that may be. This service will send a taxi to pick you up at your door and drive you anywhere for free.

Would you still use the bus/subway instead?

Yes. Mostly because I'm stubborn like that, and I greatly favor routine.

But anyway, you're making a very flawed comparison. This taxi service will not go everywhere, from everywhere, and won't be free. Not in Belgium anyway. Unions are way to strong for that to happen. Hell, Uber got itself almost booted out of the country (Taxi services have to follow certain regulations.)

I was making an argument based upon the theoretical service that is being discussed here. Why exactly wouldn't the robot taxi be able to go anywhere?

[LordBucket]

I can think of no very good reason why robot taxis wouldn't have better coverage than other forms of mass transit. Human-driven taxis are already more versatile in their coverage. Subways and buses have extremely fixed routes. A robot taxi would presumably be able to go anywhere the GPS navigation system is aware of, and existing systems are already vastly more open in where they can go. Check your local subway/bus route maps, then pick any two arbitrary destinations and plug them into google maps.

Google is specifically talking about vehicles picking you up at your front door. No subway is going to do that, and there's no reason why the taxis couldn't drop you off at the specific front door you wanted to go, as well.

As for costs, there are a lot of reasons why they could be cheaper. Obviously it would depend a lot of implementation, but subway systems cost billions of dollars and take years to build. Whereas robot taxis would use already existing roads. So long as they can keep their unit costs down, there'd be a huge cost benefit. Admittedly, google is presently spending $150,000 per vehicle. In comparison, a New York taxi is a $20,000 to $35,000 vehicle. It would take an awful lot of savings from not having to pay for the driver to make up that difference. But costs would presumably go down with mass production. It's notable that the current driverless cars are standard drivable cars that are being modified. That doesn't need to be the case. Again, they're specifically talking about the option of making little one and two man vehicles for this rather than the four seat vans being used as cabs. If they decide to make these two-seat podcars, with the usual driving interface equipment like steering wheels and brakes all under the hood, I think it's reasonable to suggest that they could be produced far more cheaply.

Then, once they exist, expansion becomes trivial. A subway, once built, has a very fixed location. Opening up the subway to new locations means spending billions more dollars over years all over again. With robot cars they can easily introduce them in small quantities into a specific city that's suitable for them. 100 to begin with, maybe, in an existing market where taxis are already successful. Simply make fares cheaper, and people will use them. Then add to the fleet incrementally and expand the coverage area. Unlike a taxi service, which needs to train and accommodate drivers and dispatchers and have conveniently centralized parking garages so drivers can return their vehicles then go home, a robot taxi fleet could be far more flexible about its range and location.

[Frumple]

It would take an awful lot of savings from not having to pay for the driver to make up that difference.

Going by barely-any-effort research, a cab driver makes on average (which is probably a terrible measure -- median would be better but effort *shrugs*) around 25k a year (in the states). Cutting out the lowest cost of the car itself (which would be sunk regardless), with that 150k vehicle you'd be looking at a bit over five years of work to break even on the cost of a driver. Probably less,* since an automated vehicle can run 24/7, which would take two or more cab drivers to manage. Depending on vehicle lifetime and demand and whatnot that could be an incredibly attractive investment, especially if you phase them in slowly and they last long enough.

*Half it, then add a little because of low traffic times... say maybe 3, 3.5 years. Add maybe another half year to account for potential maintenance cost increases. So... four years to break even, after which is pure profit? Maybe less, considering the company itself would probably pay more than 25k per driver per year, even ignoring potential training and recruitment costs. E: Other factor would be how much the company offloads maintenance costs (such as gas) onto the driver, though. That could move the break-even point back a bit.

E2: Which is to say, from the financial angle if I were a taxi cab service owner and had the capital to invest in automated taxis, I would be looking at google et al's work incredibly intensely right now.

[alway]

It would take an awful lot of savings from not having to pay for the driver to make up that difference.

Going by barely-any-effort research, a cab driver makes on average (which is probably a terrible measure -- median would be better but effort *shrugs*) around 25k a year (in the states). Cutting out the lowest cost of the car itself (which would be sunk regardless), with that 150k vehicle you'd be looking at a bit over five years of work to break even on the cost of a driver. Probably less,* since an automated vehicle can run 24/7, which would take two or more cab drivers to manage. Depending on vehicle lifetime and demand and whatnot that could be an incredibly attractive investment, especially if you phase them in slowly and they last long enough.

*Half it, then add a little because of low traffic times... say maybe 3, 3.5 years. Add maybe another half year to account for potential maintenance cost increases. So... four years to break even, after which is pure profit? Maybe less, considering the company itself would probably pay more than 25k per driver per year, even ignoring potential training and recruitment costs. E: Other factor would be how much the company offloads maintenance costs (such as gas) onto the driver, though. That could move the break-even point back a bit.

The other big factor is insurance. Most of your car insurance payment goes to paying for humans screwing up. If you had about the expected decline in accidents from self driving cars, you would see the insurance costs drop to about 1/4 their current rates. The basic part failure costs are waaaaaay less than the costs incurred when someone drifts into the wrong lane, totals both cars and causes injuries and lawsuits.

So yeah, even at current rates of $150k per vehicle (which is custom prototype manufacturing, not mass-production costs), it would likely still turn a profit.

And the fact is, once it got into mass production, it would probably not even cost half that amount. Computers are cheap (in fact, a big part of a recent NVidia presentations have been their chips designed for cars), and sensors are cheap as well, especially if mass produced. And this is precisely why automation is such a big threat to employment numbers: everything you need to do it is relatively dirt-cheap aside from a few million dollars in fixed hardware/software development costs.

[LordBucket]

snip

1) The $150,000 figure is for a development prototype of which rougly ten exist. The final product is not going to cost that much.

2) There are costs for having employees besides wages. HR costs, FICA, worker's comp, healthcare, benefits...these are all costs paid by the employer that wouldn't apply to a piece of software. In the example in the link, these costs are ~40% of wages. So your true cost for that $25,000/yr employee is $35,000.

3) I have no particular insider knowledge of the taxi industry, but I assume that a typical taxi driver doesn't work 24 hours a day. More likely they work ~8 hour shifts and multiple drivers share each car over different shifts. Meaning your $35,000 cost per employee is being paid multiple times per vehicle.

I wouldn't want to be looking at a career in driving right now.

[Frumple]

Re: 1: I was, actually 2 and 3 were both mentioned, if far from rigorously.

Should have caveated m'post with "this is lowballing basically everything," "probably highballing the autotaxi side of things," and a little bit of "massively overestimating cabbie work hours" -- basically, worst case scenario for the 'bots. Point I was making is that right now, even in a worst case scenario, owners can recoup the cost of drivers in a relatively short time -- almost certainly shorter than the numbers I used. Basically, there are a lot of savings from not having to pay a driver, and they'd make up the difference fairly quickly.

And as you and alway note, the cost outlay is just going to get worse for manned vehicles. Poor cabbies, I guess. In today's economy, you just don't keep a job when that's the facts of things.

[Helgoland]

@Autotaxis outcompeting subways: This won't happen - once all the commuters who previously used public transportation use autotaxis, the streets will be so crowded during the rush hours that the subway will be faster once more. As far as I can see, the most plausible solution is a taxi/subway or taxi/rail combination, with taxis doing the first and the last mile and mass transport systems doing the bit inbetween.

@Cost of building subways: Streets need to be built too, you know...

[Propman]

Streets need to be built too, you know...

Not to mention they need require maintenance per year, being exposed to the elements and all. Superfluously-placed roads eat up cash faster then the commerce they generate (especially in areas that receive snow), something which no amount of optimization would fix, and sometimes your better off simply building rails or even an airport/dock to make a low-trafic, out of sight area accessable.

Ultimately, autotaxis won't obsolete subs anymore then cars have obsoleted the passenger train; a sub can only get you to a set amount of areas, but those areas it can get you to, it gets you there faster bar none.

[Helgoland]

Ultimately, autotaxis won't obsolete subs anymore then cars have obsoleted the passenger train; a sub can only get you to a set amount of areas, but those areas it can get you to, it gets you there faster bar none except giant catapults.

FTFY