Print

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

[Leafsnail]

Those articles both assume that Google's numbers are correct with no basis.  A lot of these are just bizarre.

A driverless vehicle could theoretically be shared by multiple people, delivering itself when and where it is needed, parking itself in some remote place whenever it’s not in use

Like I don't understand the basis for this claim.  Yes, cars could be shared.  This already happens sometimes.  What exactly does the car being driverless change, on a fundamental level?  There's obviously some utility in it being able to deliver itself (that could save a bit of time for the people involved) but fundamentally people own cars so that they can get in them and drive them to places whenever they want.  That won't work if someone else is trying to share the car with you.  There could of course be a rental system with a fleet, but that just takes us back to the "taxis will be slightly cheaper point".

It also assumes without basis that the cars will be super-safe (so safe that apparently people won't even want basic safety features in their cars).  Which seems rather premature considering the technology hasn't even been invented yet and Google is just making up numbers to impress its shareholders.

[freeformschooler]

I still want to know if driverless cars will report people for jaywalking which everyone does in suburbia. If not, hail driverless cars!

[IronyOwl]

And finally, here are a coulpe Forbes articles with some interesting speculation that google might possibly violently disrupt the established auto industry, because they're apparently not even trying to compete with what google is doing. Audi, Nissan, BMV and a lot of other manufacturers are developing self-driving vehicles, but with exclusively highway use in mind, which considerably limits their scope.

Naturally. Capitalism not working particularly well and then working as intended tends to have rather dramatic results. I'll find it hard to be surprised if a lot of them just sort of continue on more or less as normal and then collapse overnight.

On the other hand, it kind of concerns me that Google's currently the only one really pursuing this. Presumably everyone else will (try to) jump on board once it succeeds, but depending on how that goes they could end up with similar issues to the current auto industry or YouTube, in that they're not necessarily the best so much as able to wield a lot of leverage. Still a net gain, but it'd be unfortunate if we had to wait for teleportation technology to render the whole thing obsolete to break the new set of monopolistic roadblocks.

Those articles both assume that Google's numbers are correct with no basis.  A lot of these are just bizarre.

A driverless vehicle could theoretically be shared by multiple people, delivering itself when and where it is needed, parking itself in some remote place whenever it’s not in use

Like I don't understand the basis for this claim.  Yes, cars could be shared.  This already happens sometimes.  What exactly does the car being driverless change, on a fundamental level?  There's obviously some utility in it being able to deliver itself (that could save a bit of time for the people involved) but fundamentally people own cars so that they can get in them and drive them to places whenever they want.  That won't work if someone else is trying to share the car with you.  There could of course be a rental system with a fleet, but that just takes us back to the "taxis will be slightly cheaper point".

It also assumes without basis that the cars will be super-safe (so safe that apparently people won't even want basic safety features in their cars).  Which seems rather premature considering the technology hasn't even been invented yet and Google is just making up numbers to impress its shareholders.

I would imagine most of the cost in the taxi industry comes from having to pay a human being to operate the thing, not in gas or maintenance. That'd result in a far more significant drop in cost than "slightly," which drags the concept of universal taxis significantly closer to feasible, which drags the price down even further, and so on.

Fair point on safety features. Most car accidents are due to human error, which Google's supposedly managed to beat by a large margin, but until you have no other humans on the road at all that's not necessarily all that comforting.

[Alev]

I still want to know if driverless cars will report people for jaywalking which everyone does in suburbia. If not, hail driverless cars!

Wait, people report other people for jaywalking where you live? I live in suburbs, and that never happens.

[LordBucket]

I could buy a lot more food during the year if I could slash vehicle related costs to a fraction of what they are, ha!

According to this this study:



Reading the fine print, looks like they're estimating that it could result in anywhere from about 60% - 90% cost reduction over owning a car.


It's also worth noting that google has a patent that makes it look like they're considering the possibility that they might be able to make it so cheap that they can have transport subsidized by businesses and advertising. You want to go to a grocery store or restaurant? Call the taxi and go. The store reimburses google a dollar for bringing you to them. Not going to a business where you pay for service? That's ok. Just agree to listen to some adds during your trip.

[GlyphGryph]

Leafsnail, anecdotally I could see it making huge changes. Lots of people are reluctant to lend or share cars, extremely so in my experience, because of trust issues and liability and insurance concerns. The driverless car completely negates both of those, essentially. Beyond that, sharing a car is logistical hell unless the people sharing it live on the same block. This negates that as well - needing a car to get to the car is no longer an issue, we wouldnt need to do weird driver tradeoff dances.

Finally, it would mean I could park my car somewhere cheap and out of the way whenever I use it, which means I no longer have to worry about the logistics of finding a place to put a car I *have* borrowed.

[LordBucket]

Yes, cars could be shared.  This already happens sometimes.  What exactly does the car being driverless change, on a fundamental level?  There's obviously some utility in it being able to deliver itself (that could save a bit of time for the people involved) but fundamentally people own cars so that they can get in them and drive them to places whenever they want.  That won't work if someone else is trying to share the car with you.

"Shared" as in multiple people using them, like taxis. Not shared as in a dozen people share ownership of the vehicle. For example, imagine you don't own a car. You want to go to a movie theater. So you call for a robot taxi, and two minutes later it arrives at your door. It takes you to the theater, then leaves. After the movie you call for another taxi to take you home. A taxi arrives and takes you home, then leaves.

It doesn't matter whether the vehicle that took you to the theater is the same vehicle that took you home. And during the times that your'e "not using it" it can be ferrying other people. That's what they mean by shared.

[Helgoland]

A huge ripple effect ould be city planning no longer needing to include parking space. So much area freed up~

[Leafsnail]

Typically taxi drivers get paid about a third of the fare you pay them if they're on a flat rate (although obviously business models vary).  If you factor in the increased expense of installing the driverless systems and the security system you'd need to prevent people from constantly vandalizing or stealing the car I don't see the costs decreasing that much, and a 1/3rd reduction would hardly be a revolution anyway.  And even if they did drop significantly Google would be likely to have a monopoly in this area, which means they'd just pocket almost all of the difference themselves.

As far as I can tell you're just ignoring the actual costs associated with operating a fleet of vehicles.

Leafsnail, anecdotally I could see it making huge changes. Lots of people are reluctant to lend or share cars, extremely so in my experience, because of trust issues and liability and insurance concerns.

How does having a driverless car negate either of those issues?  Unless the driverless system is able to completely override the operator (which is pretty unlikely, as it would result in your car basically kidnapping you) the trust problems would remain, and so would the insurance problems.

"Shared" as in multiple people using them, like taxis. Not shared as in a dozen people share ownership of the vehicle. For example, imagine you don't own a car. You want to go to a movie theater. So you call for a robot taxi, and two minutes later it arrives at your door. It takes you to the theater, then leaves. After the movie you call for another taxi to take you home. A taxi arrives and takes you home, then leaves.

It doesn't matter whether the vehicle that took you to the theater is the same vehicle that took you home. And during the times that your'e "not using it" it can be ferrying other people. That's what they mean by shared.

So it's exactly like a slightly cheaper taxi (or a more expensive version of decent public transport, I guess).

[LordBucket]

I would imagine most of the cost in the taxi industry comes from having to pay
a human being to operate the thing, not in gas or maintenance

There are other factors that could make it cheap too. Just off the top of my head:

 * The long term plan is that they will be solar/electric. No gas costs.
 * Massive reduction in repair/maintenance costs from the above. Combustion engines are dirty, messy and expensive to run. Electric engines have fewer of this sorts of problem, and there are fewer moving parts to wear. Just, for example...regenerative breaking which instead of wearing down pads that need to be replaced, breaking slows the vehicle while simultaneously recharging the battery. Discussion of maintenance implications for Tesla electric vehicles here
 * Vehicle insurance would be be jointly handled by a corporation, rather than individually handled by different people. I expect that a company that insures 100 drivers under a single policy is going to pay less per driver than each of those drivers would to insure themselves individually.
 * If these are safer than human drivers, which seems probable, then insurance costs will be even lower.
 * Wouldn't surprise me terribly if, given the scale here, google made its own insurance division and took out the middle man completely.
 * Reduced overhead. No need for a dispatcher, because it will be handled by smartphone app. No need for a massive parking garage for vehicles, because they wont' depend on taxi drivers who need to sleep.
 * Vehicle design will be smaller, lighter and cheaper. A lot of articles have mentioned this. Taxis are converted ordinary cars. Typically four seaters or minivans.

The current taxi being phased in by New York city is a $20,000 minivan. Whereas, here's google's current prototype:



It's a podcar. There's less to it. Nobody in the US would buy a car like that because so much self esteem is involved with car ownership. But if you're only riding in it for ten minutes to go to work, you don't need to impress anyone. They don't even have steering wheels. They're going to be cheaper.

[Helgoland]

Leafsnail, the consumer doesn't have to maintain his own car on the side. Just comparing the costs of taxis with those of using a driverless cars is misleading.

[sneakey pete]

And finally, here are a coulpe Forbes articles with some interesting speculation that google might possibly violently disrupt the established auto industry, because they're apparently not even trying to compete with what google is doing. Audi, Nissan, BMV and a lot of other manufacturers are developing self-driving vehicles, but with exclusively highway use in mind, which considerably limits their scope.

However, Audi, Nissan, BMW et all will have their technology on the road a good decade or two before Google gets their stuff on the road, and the majority of people would probably feel more comfortable buying it from an established auto maker instead of from Google.

Google has apparently stated that it's targetting a 90% reduction of vehicles on the road. Disruptive, indeed. Which brings us to:

Ripple effects. I commented on this in the OP. When this happens, a lot of things start to become obsolete:

"...a driverless car would slash hundreds of billions of dollars of annual revenue, or even trillions, from all sorts of entities: car makers, parts suppliers, car dealers, auto insurers, auto financiers, body shops, emergency rooms, health insurers, medical practices, personal-injury lawyers, government taxing authorities, road-construction companies, parking-lot operators, oil companies, owners of urban real estate, and on and on and on."

"Add up all the pieces–$450 billion related to crashes, $600 billion of car sales, $200 billion in auto-insurance premiums, the hundreds of billions of dollars of health-insurance that plausibly relate to car accidents, and so on—and you pretty easily get to about $2 trillion in revenue associated with cars each year in the US."

You're never going to free up the 2 trillion dollars though. The cars still need to be bought, parts on them are still going to break and you're still going to want insurance. Not all insurance claims are for vehicle accidents after all.

Those articles both assume that Google's numbers are correct with no basis.  A lot of these are just bizarre.

Let's not bring logic into this....

There are other factors that could make it cheap too. Just off the top of my head:

 * The long term plan is that they will be solar/electric. No gas costs.

So much for cheap then. We've pretty much hit the efficiency limit for solar panels, and the high efficiency ones are expensive as heck. Current focus on solar power is for lower efficiency but much cheaper solar panels. Which is great for a house which has lots of free roof space, but not so much for  a car which is pretty darn small. A true solar powered car is going to be very, very expensive to build, if at all even possible to build at all. Look at the solar challenge cars. light weight, massive size horizontally but very small vertically to reduce wind resistance. not practical for a taxi. Also, they can't run during night hours.

* Massive reduction in repair/maintenance costs from the above. Combustion engines are dirty, messy and expensive to run. Electric engines have fewer of this sorts of problem, and there are fewer moving parts to wear. Just, for example...regenerative breaking which instead of wearing down pads that need to be replaced, breaking slows the vehicle while simultaneously recharging the battery. Discussion of maintenance implications for Tesla electric vehicles here

You still have all of your other non engine related issues. Suspension bushings, wheel bearings, etc. Maintenance is reduced, perhaps by as much as 2/3rds at a random guess, but not eliminated at all.

* Vehicle insurance would be be jointly handled by a corporation, rather than individually handled by different people. I expect that a company that insures 100 drivers under a single policy is going to pay less per driver than each of those drivers would to insure themselves individually.

I thought we were talking about taxi's here? Or joint owned vehicles between 100 people?

* If these are safer than human drivers, which seems probable, then insurance costs will be even lower.

Can't argue with that one, people are terrible drivers. But as i mentioned before, note that not all insurance covers vehicle accidents, but probably over 75% is.

* Wouldn't surprise me terribly if, given the scale here, google made its own insurance division and took out the middle man completely.

All hail our google overlords.

* Reduced overhead. No need for a dispatcher, because it will be handled by smartphone app. No need for a massive parking garage for vehicles, because they wont' depend on taxi drivers who need to sleep.

That's all good assuming everyone has a smartphone. I guess that assumption is probably going to be valid in a few decades though.

* Vehicle design will be smaller, lighter and cheaper. A lot of articles have mentioned this. Taxis are converted ordinary cars. Typically four seaters or minivans.

Its all a trade off. you want your vehicles to be small enough to not haul around the extra weight, but if your vehicle is to small you will limit your capacity size. You are never, ever going to be able to have a taxi fleet anywhere with 100% pod cars, for example.


Edit: I guess what i'm trying to say is that for the "costs reduced by 90%!" thing to come true you're making an awful lot of assumptions about technology, engineering, peoples habbits and situations.

[LordBucket]

Audi, Nissan, BMW et all will have their technology on the road a good decade or
two before Google gets their stuff on the road

Google is predicting mass market availability in 3 to 6 years, not decades. These things have already been on the road for quite a while. As of April 2014, they'd logged 700,000 autonomous miles.

the majority of people would probably feel more comfortable buying it from
an established auto maker instead of from Google.

1) You wouldn't be buying them. We're talking about robot taxis. Paying $5 for a ride requires considerably less faith than $30,000 to buy a vehicle.

2) I don't think so. We're talking technology here. Nothing about car manufacturer screams "I know how to build software for artificial vehicle control based on  3d visual mapping!" Google is a household name, and I don't think it's a stretch to suggest that the average person knows this is up their alley.

3) According to a Cicso survey, about 60% in the US would trust a driverless car. Presumably those numbers will go up in time. I imagine it took a while for people to start trusting carriages without horses too.

the 2 trillion dollars though. The cars still need to be bought, parts on them are
still going to break and you're still going to want insurance. Not all insurance claims are for vehicle accidents after all.

So what? If you free up only one trillion dollars instead of two, is that really so terrible?

I thought we were talking about taxi's here? Or joint owned vehicles between 100 people?

Taxis, yes. End cost to consumer end "money being exchanged in industry X" are different things, but they're both things. And if it costs a company less to provide a service, some portion of that savings might be passed on to the consumer. 100 people with 100 cars need 100 insurance policies. If 100 people only need 20 robot taxis, that's fewer insurance policies. Or, as I suggested in my last post, only one policy that covers 20 vehicles. Pretty sure it's going to be cheaper than the 100 policies.

You still have all of your other non engine related issues. Suspension bushings, wheel bearings, etc. Maintenance is reduced, perhaps by as much as 2/3rds at a random guess, but not eliminated at all.

Let's say you're right. Why is a 2/3 cost reduction a bad thing? Who cares if it's not 100%?

We've pretty much hit the efficiency limit for solar panels,

It was cited earlier in the thread...if I recall correctly the number was an overall 7% increase in efficiency every year. As in multiply by 1.07, not add 7 to the efficiency number. It's slow, but it's still improving.

Also, they can't run during night hours.

Batteries. Also, there's nothing requiring individual vehicles to store all their power for the night. Having external storage is an option. Tesla motors is already working on that.

not all insurance covers vehicle accidents, but probably over 75% is.

I've been finding it surprisingly difficult to find exact figures on this, but according to this only 42% of the insurance industry is property and casualty, of which auto insurance is a part.

National average auto insurance payment is ~$800/year. Using the study linked earlier as an example, the difference between insuring 120,000 privately owned vehicles and 18,000 robot taxis they estimate it would take to replace them is ~$81 million in saving divided by 200,000 people. Even if you double the cost of the 18,000 vehicles because they're driving more (more mileage means more expensive policy) it's still $67 milllion in savings. That's $235 savings per person. If you happen to live in a place where the costs are higher, say...New York, the savings work out to $326 per person.

But again, it seems probable that google would create their own division for this rather than handing money to insurance companies. Insurance loss ratios are typically in the 40 to 60% range. So if google simply did it themselves, even if accidents remained exactly as they are it would still end up being being roughly that much in cost savings.

If we try to figure it out from the other direction, using the $235 per person figure...88% of the us is licensed, 14% are uninsured, 314 million people, that's 237 million people, ~55 billion dollars per year being paid in auto insurance.

I guess what i'm trying to say is that for the "costs reduced by 90%!" thing to come true you're making an awful lot of assumptions about technology, engineering, peoples habbits and situations.

Well, that's what they've said their goal is. If you read the Columbia University study, their case study figured that a city of 285,000 residents with 200,000 cars could replace 120,000 of them with 18,000 robot taxis. So, a 52% reduction rather than a 90% reduction.

That's still a lot.

[IronyOwl]

But as i mentioned before, note that not all insurance covers vehicle accidents, but probably over 75% is.

As an offset to this, I wonder how much the triggers for insurance claims cost in lost revenue? It wouldn't surprise me if you ended up gaining back well over 100% of the lost spending in other areas, especially when talking about death.

[LordBucket]

I wonder how much the triggers for insurance claims cost in lost revenue?

According to this overall us auto insurance loss ratio was 69.1% in 2011. For every $100 collected in insurance payments, the industry paid out $69.1 to claims. Though other sources seem to suggest that it's been 55% in previous years. If we plug in the 55 billion estimate edited into the post above yours...55 billion paid by consumers for auto insurance, and 38 billion paid out to auto insurance claims in 2011.