Hacker News new | past | comments | ask | show | jobs | submit login
Driverless Cars Get Green Light in California (cnbc.com)
141 points by codegeek on Sept 25, 2012 | hide | past | favorite | 161 comments



Industries that will be disrupted: - Taxi, shuttles, buses and limo services - Truck drivers, Delivery, Food delivery (auxilary: train and air transportation as well) - Insurance industry - Fleet management (time to write an app for managing 1000+ automated self-driving vehicles) - Policing. We won't need the CHP much in the future

And that's just from ~20 seconds off the top of my head. I'm glad lawmakers are allowing the future to happen and I hope this puts pressure on other states to join in.


It's fun to speculate on what will happen when driverless cars become mainstream. Essentially, the question boils down to: "What would happen if every car owner in the world suddenly had a 24/7 personal driver?"

- 7 year old kids can now own and "drive" a car.

- Need a lift? Ask your dad/friend to send his car to pick you up.

- Forgot your laptop at home? Send your car back and ask your flatmate to put your laptop in the car.

- Leaving New York for some vacations in Miami? Take the plane and pick up your car whenever it arrives.

- Want to take a nap? Pull down the windows and windshield curtains.

- Traveling abroad for a few weeks and don't know what to do with your car? Set it to taxi mode and profit.


I think eventually, car ownership will go down significantly, since companies managing fleets of electric cars can do it far more efficiently than private ownership of long range cars. Uber is faster than walking to the car you parked half a mile away. There's always 'free parking' for your large house party in the city. You can also create small cargo only cars for small deliveries like that laptop situation. It's far more efficient from an energy/vehicle cost/liability stand point. Amazon prime 'local' might just replace stores from a pure efficiency standpoint. The number of total cars neccesary will decrease and the large amount of real estate devoted to cars will decrease significantly. Car maintenance costs will decrease with predictable schedules, and economies of scale that come with fleets.


Car sharing only works well for urbanites without children living with them. The rest of us keep a lot of stuff in our cars: strollers, child seats, spare clothes, shopping bags, diapers, pens, ski racks, toys, first-aid kits, etc. Even with computer driven cars I'll still be willing to pay a lot to avoid the hassle of loading and unloading every day.


It's one of the great cost distortions of private car ownership: Anybody who is parking a car in public, filled with surfboards and strollers or not, is actually storing private property (car and contents) on public space. And even worse, "regular" parking spots are often free, even though they minimize the number of cars that can be "stored" on a given surface area as opposed to garages. Thus, we incentivize wasting urban space that is inherently scarce. If this stays the same, using a more efficient, stackable, standardized cargo pod that is auto-loaded onto your self-driving rental car, will be more expensive for you because you now actually see the cost. This is one of the many hidden costs of driving that people will suddenly have to face with on demand automatic car rental.


What if you kept all your car stuff in some sort of modular compartments that could be easily loaded and unloaded? At the moment there isn't much incentive to do so, but that might change with this scenario.

What I'm getting at is while your argument seems sound, and may in fact prevail, when you change one piece of the puzzle sometimes other solutions become possible.


> What if you kept all your car stuff in some sort of modular compartments that could be easily loaded and unloaded?

Where am I going to put these containers when I don't have a rental car?

Lugging them back and forth to the house is not a practical option.


Maybe your driveway is replaced by a compartment container. Maybe (certain kinds of) cars will be able to automatically load the compartments. Or maybe you're completely right.


Maybe your driveway is replaced by a compartment container.

I propose we call this component the garage.


> Maybe your driveway is replaced by a compartment container.

Lots of folks don't have driveways, garages, and so on.

More to the point, they may not be at a "personal storage location" when they're dropping off the rental car.


s/(driveway|garage)/$WhateverSpaceYouCurrentlyParkYourCarInThatWillBeEmptyIfYouAreRentingInsteadOfOwning/


I can't leave my stuff in the street, especially since it isn't the same location every time I park. In other words, there's not necessarily a dedicated space that will be empty if I'm renting instead of owning, let alone one that I control

In fact, folks who don't have dedicated/owned spaces are probably the best audience for rentals.

Moreover, even if I'm willing to take said stuff into my residence, I may need said stuff on my way to the drop-off.


It's getting kinda pointless. I'm suggesting something that might happen in a future where self-driving cars have completely disrupted all of transportation. You're countering "there's no dedicated space" - well, just maybe, that can change? It's not even a full space, it's just a "dock" to load the containers.

In fact, even today, people who don't have a dedicated space (even if that space shifts around a bit, ie. is a "pool" space on the street) can't own a car: Where would they leave it? People who are renters today already have to deal with not leaving stuff in the car, so this discussion doesn't apply to them.

Long story short: Your original argument was that not owning your own car will not be be an option for you, because you leave a lot of stuff for your kids in your car. My argument is that, with a bit of fantasy, maybe this isn't an insurmountable obstacle.

> I may need said stuff on my way to the drop-off.

Why would you drive your self-driving vehicle to the drop-off?


One of the small but nagging issues which will need to be resolved is how to deal with "items left behind".

When more people begin to use public cars for personal use/hire people will begin leaving phones, strollers, jackets, wallets in these vehicles. Unless the cars are to have omnidirectional cameras or electronic trackers to monitor everything, people are going to lose things. This is going to be a pain point. It might be just becoming aware -as people do in taxicabs but at least honest cabbies will return items as its in their interest to maintain good reputation --a stranger is only bound by altruism.


I think it will be similar to leaving your wallet,jacket,etc on the subway.


There's something that I am genuinely wondering. Will car ownership really decrease?

Where I used to live in 2003, in China, taxis were really cheap, and public transportation was functioning OK. Yet, after some economic boom, people started to buy more and more cars. It was very provable that cars were a negative in every possible way. More expensive, PITA to maintain, inconvenient to park, incompatible with drinking alcohol or being tired, but still, people preferred to own and drive their own cars!

I know a self driving car is not exactly like a Taxi driver (or even like a private driver), but the parallel is close enough for me to wonder about the consequences.


I think there's another factor in play in a place like China. Many of these new car owners probably grew up with cars being an unattainable luxury - now, suddenly, they can actually just own one. I can't blame such people for jumping on the opportunity, practicality be damned.


For the average chinese person, is it cheaper to use taxis & long range busses & subways to go everywhere, or a personal car? Will the taxis come out to your residential area, at any time? Your still paying for taxi labor on top of a car.


> - Leaving New York for some vacations in Miami? Take the plane and pick up your car whenever it arrives.

Alternatively, in say 20 years or so, going to Miami which would take 21 hours (at 65-70 mph), could probably be done on an overnight sleeper-car (at probably something more like 120-130mph).

Even a cross country trip would only take a day or so at such a rate, and with a few strategically placed stops, you could have a nice week long cross country trip at a fraction of the normal cost.

Here's a few negative thoughts though:

These driverless cars might significantly increase road usage and wear since driving becomes less of a burden.

Suburban sprawl could grow out of control since it's no longer necessary to either have public transit options or be attentive to have a long commute.

Some people might even choose to buy automated RVs instead of homes, decreasing a sense of community and investment in place.


> Alternatively, in say 20 years or so, going to Miami which would take 21 hours (at 65-70 mph), could probably be done on an overnight sleeper-car (at probably something more like 120-130mph).

Believe it or not we already have this. It's an Amtrak train between DC and central Florida. You are a passenger aboard the train but your car comes with you! http://www.amtrak.com/auto-train


Let me guess, this costs quite a bit more than flying and renting a car for a week?

This is not a cheap stab at US rail, I looked into this on the similar service that the German railways offer - and it did not make financial sense over flight+rental.


Fully automated cars let's you have all sorts of crazy solutions in the 200+ year time-span.

Think 100% electric high speed trains that let cars 'board' at highway speeds. Evacuated underground tunnels for low drag long distance high speed trips. Hybrid Maglev cars that rocket down the highway on a cushion of air while being inductively charged. Even things as simple as highways without road signs of any kind.


"Evacuated underground tunnels for low drag long distance high speed trips."

That's interesting because while I've heard of and thought about the concept of a vacuum tube train that could go ridiculously fast, I didn't think of a merely low-pressure tunnel that ordinary cars (with pressurized compartments, anyway) could use in flexible manner. You wouldn't go 15,000 mph, but you could do several hundred pretty easily.


Increasing energy costs work against all that.


> Leaving New York for some vacations in Miami? Take the plane and pick up your car whenever it arrives

my car? Like I own the car? Why would you own a car? Who takes care of the maintenance?

Just kidding. But seriously, if self-driving cars were a thing, I can think of very few reasons why any normal people would own cars.


Yes, exactly!

If the cars could drive themselves, they would probably be more like taxis without the drivers. We no longer need parking lots because the cars are constantly in use, that will free up a lot of space and mirror what already happens in dense urban areas (which depend on taxis, public transit, parking is very $$$). The cars will wear out in a couple of years from constant use (or maybe not if electric?), but we won't need as many of them. Rush hour commutes are a problem, but road capacity is limited anyways and peak pricing + mass transit + car pooling can lead to a workable solution.

Eventually, since all cars are automated, safe highway speeds could be increased dramatically (reduce space between cars, more predictability, reroute and reduce speeds to prevent full stop bottlenecks). Accordingly, you won't even be able to enter the highway in manual mode.


I agree, it could definitely be a situation where owning a car is like owning a plane. Sure, some people do it as hobbyists, but why would you when there are extremely efficient (computer operated) and inexpensive (no taxi driver, 100% uptime except for filling up/charging) taxi services that can take you around. Especially as more people use them (thus reducing wait times) I could see taxis being usable/profitable even in the suburbs.


I wonder, since cars also have a large social signalling aspect, at least in many societies.

I guess that can still be maintained, with various levels of luxury in taxis/rentals/however it works.

But then where will all that social signalling money go? I guess since it's mostly financed money anyway, that capital will have to seek other venues? Will consumer debt drop due to this? Or will something else pick up the slack?


I think simply owning a car will be such a signal, similar to how owning a plane today is a mark of the super-wealthy. There will be "automotive Boeings" that produce cars for transportation companies, boutique shops that produce them for the wealthy, and not much in between.


Yes, this is already the case in China and I suspect in many places around the world. In Shenzhen, taxis are very cheap and available within 100m around the clock pretty much everywhere. Even if you drive on average 1 hour per day which is very unusual in a high density city, owning a car would probably be more expensive then always taking the taxi. Yet, a lot of people still purchase cars (mostly luxury brands).


I suspect this is the case almost everywhere in the world. A car's purchase price + gas + maintenance + taxes easily surpasses what you'd spend on taxis, even in a 5 year period. The keys to car ownership are convenience and the emotional aspects.


Taxis don't go everywhere.


Hauling, vacations, kids, bike rack...the list goes on.

Many, many cars get customized for one reason or another. You can't do that with a taxi.


You're right that people now use their vehicles for a variety of tasks, but I think that's actually part of the problem.

My dad has a pontoon boat that he has to twice yearly haul around. It might make more sense to get something specialized for those two days than drive something sufficiently powerful and fuel inefficient year round.


I need the bike rack every other day. Others have dirt bikes, surfboards, skis. This is pretty normal.


Assuming 6315 waking hours per year and 540 hours a year spent in a car, I would say a 9% increase in productivity for the average American.

Numbers from:

http://answers.google.com/answers/threadview?id=392456

http://hypertextbook.com/facts/2002/PatrickLi.shtml


Averages are nice, but I'll be interested to see what happens to the workforce that depends upon them entirely, as mentioned above: truckies, taxis, FedEx, etc.

I say, leave it for a year to lull the companies into a false sense of security, then tax the hell out of companies using them for a few years to pay for some re-education for the old employees.


Why on earth should the responsibility (of those that will become redundant) fall on the early adopters of a new technology?

If I had become a fax machine expert would it be fair to charge a tax on those that began to use email to support me?

People should take responsibility for themselves and their own actions.


Easy on the pitchforks, man.

Maybe it's because I'm Australian and have got used to the free ride of healthcare and education, but I'd gladly take a minor tax increase if it means a significant portion of the population gets into an entirely new career quicker, instead of remaining unemployed and being MORE of a strain on the economy.

Career is a strong word, though. The sorts of people we're talking about usually have no other option, and education can change that.


You should be used to this kind of talk by now. It's the modus operandi of socialism and the Democratic party. Who ever heard of personal responsibility?


Let them eat cake?


Let them bear the consequences (unforeseen or not) of their decisions.


How dare that person get cancer, become unemployed because of unforseen shifts in the job market, etc.

Social safety nets exist for a reason. You may opt out by moving somewhere else if you would prefer not to participate in a society with said safety nets.

I say this as someone in a $200K/year+ household. People who preach personal responsibility are oblivious to the realities of the world.


Well every one knows that becoming severely ill is a possibility for anyone, that's what insurance is for. Can't afford insurance? Give up unhealthy habits to lower your rates, get a better job, look after family so that they will support you if you need help.

Get made redundant due to unforeseen shifts in the job market? Educate yourself, get a better job, work harder. Innovate, invent, think, mow your neighbors lawns. Use your brain and don't embark on a dead end career.

I agree that some social safety nets are needed for a civilized society (No one that is ill should go without out treatment if they don't have any money etc) but at the same time there are limits and I think paying for people who lose their jobs because of new technology is way past that line. We need to encourage people to be smarter, not dumber.

"People who preach personal responsibility are oblivious to the realities of the world." I live in South Africa and government handouts do nothing but keep the poor oppressed and uneducated. They are used as a tool for control by populist politicians. People will have more children just for another R250/month from the government instead of creating value. Socialism is a huge threat to everyone everywhere. It keeps the populous dumb and gives governments way too much power. Please don't preach about the realities of the world from your $200K/year+ household pedestal. Socialism is obviously working out great for you.


>I live in South Africa and government handouts do nothing but keep the poor oppressed and uneducated.

I live in Australia, one of the best examples of how government handouts can ruin a race, and I'm saying this as someone coming from a low-income family with poor education and working+educating myself to a $100k/y job before I'm 30.

>I think paying for people who lose their jobs because of new technology is way past that line.

I think completely supporting them would be a bad idea, but as said in my original post, assistance in re-education should be partially paid for. I am always of the opinion that education should be at least partially subsidised - full subsidy encourages a poorer quality of education.

Example: You're 26 and have just spent $80k and 8 years of toil studying furiously for a degree, and the market for your chosen field evaporates. Sure, you've got the rest of your life ahead of you, but now you're significantly in debt because of something you had no control over.

This is a difficult analogy to compare with taxi drivers, but I hope you see the point - and over the next 10-40 years we will see this happen more and more as machine intelligence replaces more and more jobs.


I don't have much to add to tetomb. While certain safety nets can be successfully argued to make sense in a large, wealthy, productive, and strongly-led society, that doesn't imply all safety nets make sense in all societies. It also doesn't imply that even the desirable nets can be had feasibly at a good quality of service. For deductive reasons why they don't, see Carlyle, Mises, et al. (If you prefer inductive observation, you can also look at a wide variety of examples across history of failed and failing attempts.)

Lest you think I have no experience with the US situation and system, I've never lived in a 6-figure income/year household, I have a family member who was unemployed for over a year, who broke her arm only months after losing her job and medical insurance (all the years of medical insurance payments wasted, she having never been injured or seriously ill during that time--but that's a risk one takes when deciding (when one has a choice) to pay for insurance), finally getting a job that paid less than half of what she used to make--she's just trying to make it to retirement because all the money the government has forced her to "save" is effectively untouchable until then. I have a cousin who recently got knee cancer (but he'll be fine because it hasn't spread, his dad's an attorney, and a lot of cancers are curable), and I have another family member with a mental illness who without medical treatment could not function. He has paid the (often brutal) consequences of going off the medication enough times, fortunately he's been stable now for some years. (And mental illness is far worse than cancer, from the underfunded and incomplete medical understanding and treatments to government systems for aid to the social stigmas that make making friends, getting family support, getting a job, and a whole bunch of other things incredibly difficult.) I'm not oblivious to realities of the world. I see the causes deserving of aid as well as the crappiness of institutionalized aid systems.


> Traveling abroad for a few weeks and don't know what to do with your car? Set it to taxi mode and profit.

CarBnB



7 year old kids can now own a guided missile.

7 year old kids can be kidnapped in hacked vehicles.

Statewide gridlock due to DDOS.

Just looking at the flip side.


Think about those for a minute.

I rather doubt a self driving car in the hands of a 7 year old could possibly be more dangerous than a manual car in the hands of a 7 year old. It's fairly trivial to keep kids from using an iPhone inappropriately. A simple password would easily stop it, or we can keep using physical keys.

7 year old kids can be kidnapped in non hacked vehicles. It's considerably easier to track down a stolen car from 2012 with OnStar than one from 1992 without.

Relavant HN article from today: http://www.bbc.co.uk/news/magazine-19660765. I don't see why DDOS would be the attack of choice for vehicles, but say it were to happen. DDOS attacks happen. They can be defended against (see http://blog.cloudflare.com/65gbps-ddos-no-problem).

For the foreseeable future cars will have manual overrides, and I can't see any attack or series of attacks, no matter how horrifying, resulting in 32,885 deaths over the course of a year.

Certainly there are many scenarios and details to consider, but the typical objections seem to be red herrings, strawmen, or just totally bogus.


> Statewide gridlock due to DDOS.

Modern cars are 100% dependent on computers, yet this doesn't happen. Why would it begin happening in the future? Note that the Google car is designed to be autonomous, it has no dependencies on external systems - there is nothing to DDOS. Yes, you can blind/obstruct it's sensors in a variety of ways, but it's much easier to just aim a laserpointer to a driver's eyes today.


"Traveling abroad for a few weeks and don't know what to do with your car? Set it to taxi mode and profit."

- come back to a completely trashed car that you just paid $50k for.


We have insurance for that. And I imagine people's emotional investment in their cars will be less when they don't drive them themselves.


The auto industry itself; a family will far more easily be able to get by on a single car: "Take Little Joey and Little Suzie to work, drop dad off at the office, go pick mom up at home, then at 3:30, pick the kids up at school, drop Suzie off at soccer practice, and have dad home in time for dinner at 6."

For that matter, shared, scheduled cars on a neighbourhood scale would be pretty great.


I've made this point many times on driverless car threads. The auto industry sells cars on the basis of miles driven, not number of families. As the passenger miles will increase with driverless cars, the car makers have absolutely nothing to be concerned about, and much to be delighted about driverless cars.


> The regulations would allow vehicles to operate autonomously, but a licensed driver would still need to sit behind the wheel to serve as a backup operator in case of emergency.

That alone limits the potential "disruptiveness" quite a bit.


It reminds me of the "Red Flag" laws, passed in the early days of the automobile, that basically made cars useless, so you'd have to defeat their very purpose in order to obey:

http://en.wikipedia.org/wiki/Red_flag_laws


In the United States, the state of Vermont passed a similar flurry of Red Flag Laws in 1894. The most infamous of the Red Flag Laws was enacted in Pennsylvania circa 1896, when Quaker legislators unanimously passed a bill through both houses of the state legislature, which would require all motorists piloting their "horseless carriages", upon chance encounters with cattle or livestock to (1) immediately stop the vehicle, (2) "immediately and as rapidly as possible... disassemble the automobile," and (3) "conceal the various components out of sight, behind nearby bushes" until equestrian or livestock is sufficiently pacified.


Is that real? O_o


I strongly suspect the "Quaker legislators" part is untrue, because I'm pretty sure Quaker dominance of Pennsylvania government had ended long before 1896. (I couldn't find a reference on this point though, so I might be wrong.)

I don't know if the rest of the quote is accurate.


Law makers are uncertain about the technology right now. When they're the ones sitting behind the wheel in one of these cars, reading a book while it drives, then we'll probably see new legislation removing the necessity of the driver.


Do law makers drive cars still? (not counting for fun)


For now. That will change eventually.


That's true, but it's probably a temporary measure (a decade or so?). And even with a backup driver, automated cars have the potential of vastly decreasing traffic jams.


What kind of license do they need? Would a regularly licensed teenager now be able to "drive" taxis and other commercial vehicles?


I'm very annoyed by this requirement. It doesn't hold up to the slightest scrutiny as being necessary yet completely hinders most opportunities for true disruptiveness.


You don't think the novelty and risk associated with a fully automated system that's still in the experimental stages warrants any immediate caution?

Commercial planes have had autopilot for years, but they still require at least the pilot or co-pilot to sit behind the yoke.

When driverless cars have had a few million or billion hours without statistically significant incidents, I think we'll see the laws adapt to not even requiring a human fail-safe.


Actually, for commercial airliners its "worse". They require both a pilot and co-pilot behind the yoke, even in autopilot.

If an airliner is flown with less than two crew in the cockpit, even for a short period of time, the remaining pilot/co-pilot must wear a mask when flying above 25,000ft.


Airliners used to have a third crew member, the Flight Engineer, but that job has been automated into oblivion.


A friend mentioned that non-scheduled (no passengers or commercial freight) flights in large aircraft (eg. 737) can still operate with a single-pilot and a rated flight engineer only. The pilot had to have her mask on (or clipped to the helmet) for the entire flight. This option is rarely done commercially, obviously.


I think the slightest scrutiny shows its complete necessity.

No licensed driver, car encounters situation it's not programmed for.

Now what?


Car pulls over, unlicensed driver calls a cab.


Unfortunately there is a very, very, large set of possible situations a car may find itself in that requires 'user intervention'. Especially without much context-aware systems, or AI, the best error handler a driver-less car has is "stop, inform driver". Is there a curb to pull over to? What if the system is unable to determine if the path is clear because something like a branch is in the way?


Well, my line of argument is that any driver in a self-driving car is going to be completely inattentive (think available-to-public, not a test-phase car). For any self-driving car to be considered safe enough for public use, the car will need to be proven absolutely safe in "emergency" situations (i.e., auto-shutoff, etc.).

So, if the car is proven to be absolutely safe, then why do we need a driver? To pull it off the road after it shuts off or stops?


Take professor Thrun's course on building self driving robots at Udacity, and then you'll see why this requirement is necessary at this stage of the development of this technology.


The whole purpose of the new autonomous vehicle law is to allow development of the systems in real-world conditions, before they can get to the next level of driving without a driver behind the wheel as a backup.


I seriously doubt this will disrupt any of those industries, except possibly shuttle buses that cover very restricted routes. It will be years before these get past the handful of people that will likely rush out to buy one (Woz, for example, if they put the tech into either a Prius or Hummer)

The technology pundits are ignoring the human factor, giving up control, trusting your kid's lives to a computer, the same technology that crashes on the desktop, and on our smartphones. What happens when the first car has a BOSD and the backup human driver isn't paying any attention and the accident happens within seconds of the crash? These things will likely kill a bunch of people and will go the same way that the C-5 or the Segway did.

Maybe I'm too cynical, but until we have HA computers that exist outside of a datacenter, I doubt a proof-of-concept by Google will go anywhere in this or the next decade, let alone disrupt anything. The idealism and flying cars optimism is well placed and I'd love to see it happen, but the future nearly always takes much longer to happen than people predict...


>trusting your kid's lives to a computer, the same technology that crashes on the desktop, and on our smartphones

This is all math here. If a computer is statistically better at driving than a human you can ask the same question as "Would you give up control to decrease the chance of an accident? Or would you risk your kid's life just to feel in control?"

>What happens when the first car has a BOSD and the backup human driver isn't paying any attention and the accident happens within seconds of the crash?

People will probably die. The the same thing that happens when a human crash a car which is already happening everyday.


While I agree, the problem lies, as in most cases, with our education system in this country. The general public has proven time and time again to have a very tenuous grasp of mathematics.


"I doubt a proof-of-concept by Google ..."

You know they actually have a physical car? That car was able to drive itself from Mountain View to Tahoe.

As of August 2012, it has logged 300,000 miles.

http://www.theatlantic.com/technology/archive/2012/08/google...


If they actually work in real-world conditions and if they can be made to be reasonable in cost, I can see the conversion happening in a very short time frame. A few years.

But given the conditionals of my first sentence, the conversion may not start in earnest for a decade or more.


> but until we have HA computers that exist outside of a datacenter

We do, in many places - your car is one such place. Another is every single hospital in the civilized world. You know, there's a continuum of world between HA datacenters and your smartphone.


> The technology pundits are ignoring the human factor, giving up control, trusting your kid's lives to a computer

This already happens every time you hit the brakes in a modern vehicle.


Not to mention the huge revenue shortfall when parking meters and public parking lots are no longer used. You tell your car, orbit the building for a while I'll be about 40 minutes.


The car doesn't need to orbit. It can go off and park itself somewhere else, and come back as you are about to leave. Think of it as an automated valet.

This does mean we don't need the huge car parks around places of interest as everyone can be dropped off and picked up at the front door.

(And when the car is going to self park, it can use speeds and routes that avoid adding congestion to the people trying to get somewhere.)


If we all start doing this we will end up with huge traffic problems. Luckily, the wasted car mileage and fuel consumption will probably make it an unattractive option. One thing you could do though is ask your car to drop you off and go find a parking.


Do you really think it'll cost less (in gas money) to orbit rather than to just park? Or do you have a Tesla? (http://www.engadget.com/2012/09/24/tesla-supercharger/) :)


Yes, gas powered cars will be largely dead in 30 years. It will be all electric bot cars.

The cars won't orbit though, they will go park themselves or pick up someone else for an errand.


I'd rather you park it, if you don't mind.


I expect that the way Google would design it is that you tell your car 'come back in X minutes' and the car would do the calculation to optimize for cost and time on target, with a user preference setting which has higher priority to the user.


That's a waste of gas. Parking will still be needed as much as ever.

The exception to this will be autonomous taxis, of course.


I've had way too many experiences driving in a city when I've just needed to run inside somewhere for 2 minutes and really wished someone was in the car with me to drive around while I'm inside rather than spending 20 minutes to find a parking spot. An "orbit" feature would be pretty useful for situations like that, and would even save gas occasionally, considering how long you can spend driving around looking for parking.

(Too bad it'll probably be a while before robot cars can legally operate without a human in the driver's seat and paying attention, though)


My car can idle and drive slowly, for an hour on less than 1/2 a gallon of gas (have to watch it more closely the next time I'm stuck in traffic trying to get across the Bay bridge). That is $2 in California today. It costs $3 - $5 for an hour of time on some downtown parking meters, can cost $5 per 20 minutes in the parking garages. Trust me, gas is cheaper, and with a hybrid doing the driving, I could see the 'hover' mode being like 50 cents of electricity.


It's only cheaper because the cost of the space you're taking up on the road and the cost of the traffic you're causing is not accounted for. The solution to this is a VMT tax that accounts for that cost.


Perhaps, San Francisco policy is to use parking prices to discourage vehicle use. They haven't gone the London 'congestion tax' route yet, perhaps they will.

There is a difference between what "is" and how the system responds to what "is". In this case I don't doubt for a moment that drivers of self driving cars would use this technique to economic advantage, and yes if it got too bad it would invoke a response (much like the $5 parking meters with credit card readers in them was a response to cars in general).


And don't forget that with fully automated cars that can at least roll a few meters without starting a combustion engine (e.g. hybrids) you can park the cars without having a "road" between every two rows of cars, reducing space. Similar to this, just not with shelves but the cars themselves moving:http://www.rollster.de/picture/upload/Image/8780a17bd3128fe2... Also, garages would not need to be standing height, only lit on demand etc.


While it won't go away, the auto insurance agency will certainly change. I wonder if liability will shift from driver (passenger?) to manufacturer/Google.

I for one can't wait until we see car fatalities drop to the hundreds per year if not 0.


Please can you explain how each of those will be affected in a big way? I can't think of ways that most would be affected in a real tangible way. I use taxis because I don't own a car (edit: ignore this one, I wasn't thinking straight, hah), my deliveries (groceries, items from online) are manned because there is management of the delivery needed, not because someone needs to drive. I can see how the insurance industry will be affected in a large way though.


I believe he was implying that the humans would be impacted. Taxis that drive themselves don't get tickets, don't forget to turn on the meter, and don't need to be paid part of the fare. Basically any people moving services are made more economical by removing human labor, insurance risk, and traffic liability.

It remains to be seen if those savings can actually be realized of course.


There is no need for the delivery to be manned. The vehicle can pull up, you enter a code and it provides the goods vending machine style. (Note Amazon already has these in several cities.)


I have no intention of being home every time I get a delivery.


At the moment the delivery drivers have all these work rules and health and safety regulations (eg 9-5, limited hours per day/week, holidays off etc). It makes sense to schedule deliveries around their convenience trying to get as many deliveries per driver hour possible.

But robotic drivers don't care. You can dick them around as much as you want (tell them you'll be home at 6, then tell them 7, then tell them 2am). It doesn't bug them having to wait around for you.

Consequently the vast majority of the time deliveries will happen around your availability. That is probably good enough 99% of the time. As to what happens when you are out of town, the delivery can just wait till you get back.


Challenges it will face: - insurance regulation, moral dilemmas (decision to make when confronted with living hazards), drinking and riding, more time for advertising, and a rash of start-ups pitches that mash up concepts like Uber & Kozmo because they can.

Still seems worth it.


Here's one for you: City real estate dips in price. Suburb real estate rises, especially at the fringe of the suburbs. A self-driving car with a LTE wireless hotspot would make a previously absurd commute manageable.

Heck, you could sleep during your drive in to work.


Anything related to a gas station. No need to go with the car to fill up.


It'll be interesting to see whether and how much protectionist BS makes it into the final regulations. I'm all for very strict safety regulations on these vehicles. The first accident caused by one could set the industry back decades, Challenger-style, unless the vehicles are so demonstrably safer that legislators can resist the inevitable panic of their irrational constituents. But I hope any such regulations are driven by the actual need to make the vehicles safe, not by the desire on the part of existing stakeholders to make the vehicles' introduction as painful and difficult as possible.


Existing stakeholders will protect their turf and rely as you say on the 'panic of irrational constituents.' And it would be irrational given the amount of human error already occurring--people falling asleep at the wheel, texting, drunk driving, carelessness, etc. Also, the market will take care of this without State regulations--who would want to purchase a perceived unsafe vehicle?


A large part of the problem is that much of the total risk of driving is borne by people other than the car's occupants. People in other cars, pedestrians on the sidewalk, etc.


The one inaccuracy you have here is stating the market will take care of it. It won't because the majority will know these cars are better and safer.

The 'irrational constituents' are always a disproportionately powerful minority. That's one difference between free market and politics.

Though, even a 'minority' (< 50%) of people considering these cars are safe enough to own would likely be enough to sustain the industry.


The Challenger accident was a very public disaster for NASA. When McDonnell Douglas felt the effects of management complacency resulting in DC-10 crashes, the public perception that the jet was unsafe doomed the company. If a driver-less car has a "minor" accident due to a technical glitch, it wouldn't achieve the same level of notoriety. The Ford/Bridgestone defective tire incident is a classic example of management complacency causing accidents, but because it was a series of accidents without graphic video for the news story, it didn't achieve the same level of public awareness. http://en.wikipedia.org/wiki/Firestone_and_Ford_tire_controv...

Both Ford and Bridgestone companies were able to "compartmentalize" the fallout and stay in business.


The 'unintended acceleration' recalls [1] provide a recent example of how things can very easily go the other way (though of course Toyota's still in business).

[1] http://en.wikipedia.org/wiki/2009–2010_Toyota_vehicle_recall...


> The first accident caused by one could set the industry back decades

Unless they deploy to multiple regions, maybe? I imagine some cultures will be much more forgiving of and reasonable about the occasional accident than American culture. If even a single city in the world successfully deployed autoautomobiles, the impact would be so swift, obvious, and desirable that most other societies would follow suit in spite of the loud, irrationally fearful minority.

I hope. :)


Oh, I know! Challenger basically made private spaceship manufacturers and pilots have to hide their faces in public, so vilified were they!


I don't think there was much of a private space industry then. What I'm referring to is the shift in public perception and the delay that imposed on NASA.


I don't see this being disruptive at all. I think they will be adopted in industrial applications, such as self driving dump trucks on mine sites, where humans are not riding in the vehicle, and where there will be significant cost savings.

There is also the human factor. People aren't going to give up their livelihood quietly, and people won't be willing to trust their life to a computer.

Also, who's going to control these vehicles, government? Does that mean that the government has the power to redirect a car to anywhere they please, depending on the passenger, who has been identified by facial recognition?

Trains/trams/buses are proven, already exist, and are a much better solution for public transport.


Does the driver have to be awake? Sleeping and driving can become the new texting and driving.


I predict that we will have to legislate Driverless as the default state of all vehicles within the coming decade. why? because at greater than 10% penetration of automated vehicles, those people still in 'manual' mode will become more and more reckless.


Whether or not you believe in the climate crisis or global warming, I strongly recommend all proponents of driverless cars to read The Energy Glut[1] in order to get a good idea of one potential scenario of how America will end up if every man and his dog can use a car to get anywhere. If anything, the studies which are referenced in the book are rather hard to ignore.

As a quick example of the sort of thinking in this book: more cars on the road means that there will be fewer people willing to risk being outdoors anywhere near a road (i.e. most residential areas in this day and age), resulting in higher levels of obesity. As more people become obese, fewer people are willing to walk or cycle ever-decreasing distances, so they get a car -- increasing the number of cars on the road. After that, we enter a rather vicious circle. The average weight of the world population will increase rather quickly, and we will end up like the people in Wall-E. I'd rather ride my bike, but thanks for the offer.

[1] http://zedbooks.co.uk/paperback/the-energy-glut


Then let them be obese. It is not my responsibility to make sure other people don't get fat. We shouldn't stifle innovation (if that's what you're getting at) because we're scared that people will get.. even more fat.

And isn't obesity more about diet than it is exercise anyway?


Hah, that's a fair approach. From what I can understand, obesity is more about the lack of opportunities to exercise rather than the amount of calories eaten (see http://dc.streetsblog.org/2012/01/17/maps-show-striking-corr... for an example, although obviously correlation != causation). It seemed to be the case that as the number of cars on the road increased, calorie intake actually went down.


If the cars were safer and more aware of pedestrians, like what we have seen from google car demos, I would be much more confident crossing the street, etc.


That depends, I would be more than happy if they were limited to 20mph. If they were limited to something more like 40mph, I would just make kids stay the hell away from roads: http://humantransport.org/sidewalks/SpeedKills.htm

That said, I would be interested in seeing how driverless cars perform at stopping in the case of sudden and unexpected obstacles compared to human drivers?


This is awesome and revolutionary. No more looking for parking spots- just send the car to the nearest garage or back home. Also the cars could coordinate to minimize traffic. So many possibilities!


Not for a while: "The regulations would allow vehicles to operate autonomously, but a licensed driver would still need to sit behind the wheel to serve as a backup operator in case of emergency."

Given the current state of the technology, this makes sense. And in the future, we may yet have fully driverless cars, instead of merely autonomous ones.

Edit: The way google has been testing these cars, AFAIK, is through a special "experimental vehicle" permit.


Indeed -- They'll have to be on the road for years before they are allowed to operate without a person inside.

I have a few concerns about the short term ramifications.

1) The person behind the wheel can't do work, read, eat, or apply make-up if their primary responsibility is to second-guess the computer. So the most tantalizing benefit of being driven wherever you want is going to have to wait until some future day.

2) A vehicle that obeys the speed limit and makes a full stop at stop signs is going to infuriate some drivers. They should probably avoid matching the speed of adjacent cars when possible, to prevent blocking aggressive drivers that want to pass. These cars are probably going to avoid the fast lane.

3) I can see plenty of scenarios where an autonomous vehicle acts safely but still gets in an accident. Slowing to avoid a collision could result in being rear-ended, or a human driver might merge into the car because he failed to check his blind spot. So there will be accidents. I assume the cars will have to be rigged with video cameras and telemetry to prove that the car's programming was not at fault.


In Australia (and presumably the rest of the world?), it's never the fault of the driver who was rear ended. Everyone else is expected to maintain a safe distance in case the driver needs to perform an emergency stop.

Actually I remember hearing somewhere that Japan uses a 50/50 split for all accidents, without working out who's to blame, to encourage everyone to be more responsible. They'll definitely have to tweak the cars driving styles per region. I'm now actually pessimistic that this will ever leave the USA any time soon, given how long it seems for even simple things to make it out.


Defensive driving includes not speed-matching people in your blind-spot, and being mindful of merging cars, going so far as to move into an adjacent lane if the merging lane disappears. The cars need to consider this anyway.

More thoughts on speed-matching the car in your blind-spot: there are many, many things that can go wrong, even if a car has omnidirectional visual information. If a tire blows, you have to compensate. I wonder how much sway-room a computer needs to retake-control/move the car off the road.


They should probably avoid matching the speed of adjacent cars when possible, to prevent blocking aggressive drivers that want to pass

As a sibling comment said, this is unsafe. The Google car actively avoids being in another car's blind spot as much as it can.


My concern is mostly for edge cases, like stretches of two-lane highway where both lanes are going the same speed because of congestion. Is the car going to fluctuate its speed often, to try to isolate itself from neighboring cars? Or would it reduce its speed even further below the limit so that the left lane is constantly passing?

Keeping the left lane moving faster is probably wisest, but it's going to infuriate the drivers behind the autonomous vehicle, even though they're in the slow lane.

How far would a car go to avoid being rear-ended? Would a tailgater be able to cause the car to accelerate? At what point would the car stop trying to avoid an accident because it's safer to get rear ended than continue avoiding? When you put these things on the road, human drivers are going to start messing with them.


What do the regulations allow that's not allowed today? Google has already driven 300k+ miles autonomously, with a driver behind the wheel as a backup operator.

Also, what's the benefit to the owner/operator? Are you allowed to pay less attention to the road so you can be doing something else 99% of the time? Are you allowed to talk on the phone? If neither of these is true, then the legislation might put the cars' developers on more sure legal footing, but would appear to negate the primary utility of an autonomous vehicle.


You just know this will lead to Google Street View getting updated a little more often as the cars get to navigate themselves around constantly.

Oh right, except for this: "The regulations would allow vehicles to operate autonomously, but a licensed driver would still need to sit behind the wheel to serve as a backup operator in case of emergency."


"In 2009, Google obtained 3,500 miles of Street View images from driverless cars with minor human intervention.[citation needed]"

http://en.wikipedia.org/wiki/Google_driverless_car


I'd love to see other cars in their testing group too. Don't get me wrong, I harbor no ill will towards Prius or their drivers, I just don't fit in them.

Public transit with these installed would be fabulous as well. You'd still need someone behind the wheel as backup, but it would be great to have a potentially tireless public transit system.


How big are you? The latest Prius is essentially a full-size sedan in terms of the front seat. What kind of car do you prefer?

It's hard to imagine that this tech won't be available for all classes of cars once it is sufficiently well tested.


6'6" and built like an NFL lineman. Last time I got in one, I couldn't sit up straight, my knees were resting on the dash and I couldn't close the door. I'd prefer something in the full size category but would rather it was a van, SUV or truck so I didn't have to squat down so low to get in. I do a lot of outdoors stuff that takes me on some rough roads and I usually have need of hauling and some cargo space.


Ok, driverless cars are cool, because they have this robot touch which is appealing to computer scientists. Furthermore, we will get there eventually, as they will drive saver and (without a driver) will decrease the costs for transport and logistics industries.

But does it affect the average human? It just strikes me how many anecdotes I read in connection with selfdriving cars ala' "calling the car to pick me up from work".

Assuming that most HN users are from the US I can just conclude that public transport is really bad and people do not use their bike there. E.g. most people here (Europe) just do not need the service of a car in their daily life. So in the end it's a logic step, but I do not think that it makes sense to actually buy a self-driving car for yourself.


Isn't it a requirement for "driverless" cars to still have someone who can take over control of the car during emergencies? I mean, the Google cars that were driving had someone sitting in the passenger seat, I thought.

I don't think we'll see truly robotic cars anytime soon.


I welcome our robot overlords. I do wonder, however, who's going to do quality assurance on these cars. The first startup launching MVP driverless cars with a horrible accident record will set back the industry for years.


I think that's the whole point of setting out requirements that autonomous vehicles must meet before going into production/sale. Just the same as current car manufacturers have to meet certain safety requirements, in the future manufacturers may have to prove some level of reliability, robustness, and exception handling.


Wow. Technology is really changing the world now. What's next? Flying automated cars? I cant wait for the next invention! The government/inventors should better make these driverless cars very functional and safe to use since we all know that computers cannot think right away unlike human when it comes to urgent incident situations. http://sommer-solen.dk/cctv-have-en-afskraekken...


This is the big reason why high-speed rail in CA is a terrible idea and destined to become an expensive white elephant. Driverless cars will start to become available just about the time that the rail line is complete, after the inevitable lawsuits, engineering problems, and construction delays. Why would I buy a train ticket to LA and then deal with getting from the train station to my final destination when my car can drive me directly door to door?


So, as times change ... where will I be allowed to go to drive, myself..? For people that enjoy driving; because, as has already been suggested by this thread, these autonomous cars will probably be very boring on the road.

I suppose race-track memberships and racetracks will become more common --- or perhaps they'll be like a horseman's club - a hobby for select members.

Maybe the USA could even get a Nürburgring ... or one per region.


Sure. We still allow people to fly planes, don't we?

I expect that once a drivers license is not a requirement to exist in much of the US that the standards for being allowed to drive will be tightened significantly though.


Does anyone else find it amazing that we have driverless cars but yet, most subways and trains still require drivers/operators?


The metro in Copenhagen is driverless. http://en.wikipedia.org/wiki/Copenhagen_Metro


It's a union thing. In NYC, the L service is completely automated, but the union won't let trains run without two crew members. It was apparently hard enough to get the union to allow the G line (the least popular line that only runs half-size trains every 20 minutes) to run one-person operation on the weekends!

(Though if you don't need drivers, why do you care about the union? I guess the problem is that it will take maybe 50 years to convert the other lines to automatic operation. And you still need someone on the train to answer rider questions and handle medical/police issues.)


BART was meant to be fully automated. The "Fremont Flyer", which ended up in the parking lot, put an end to those plans.


List of driverless trains http://en.wikipedia.org/wiki/List_of_driverless_trains#Compl...

Although these are short-range, urban trains.


(http://en.wikipedia.org/wiki/List_of_driverless_trains)

Docklands light railway, in London, is driverless.


Unlike cars, trains can't stop on a dime if there's an obstruction on tracks. Nor can they swerve to avoid it.


Is this an argument that a human driver is required? To me, it sounds like if there's nothing you can do, the computer's job is much easier, not harder.


Unlike humans, computers never get tired from working long hours, can see around corners using various sensors and process the information fast, and can act immediately on instructions from a remote operator.


The SkyTrain in Vancouver has been driverless since it was built in 1986. I don't think any such LTR systems have had drivers since then.


I keep imagining the technology used with recreational vehicles. Spend a day as a tourist in Chicago. Hop in the RV and go to sleep. Wake up in Memphis. Visit Elvis, enjoy waffles & fried chicken, then hit the hay and wake up the next day in Orlando. No down time.


What would be potential startups in this space? Besides the obvious "fleet of driverless cars that drive people places."

It will be revolutionary, and Google is only going to get a plurality of the profit that comes from it.


Curbside delivery. Build a truck that has a bunch of different little cubbies. Load it up and send it around. App on the customer's phone beeps when the truck arrives and gives them a locker number and an unlock code.


"Parkless" cars. You own it, it drops you off and picks you up at work, and spends the rest of the time at your house.

Personal local delivery. Your car drives to a lot next to a grocery/department/whatever store. A grocery clerk walks to your car, presses a button on a smartphone app, and loads up your car with your online purchases then your car drives itself home.

Same as above except your car/truck is rented out for overnight local deliveries / short-distance freight hauling while you sleep.

Use of sensor streams from driverless cars for surveillance, mapping, etc.

Installing high-speed municipal wireless data or cellular infrastructure on fleets of roving cars without the necessity of securing right-of-way for fixed structures or ground data lines (e.g. using microwave or open air laser comms for high-bandwidth links to aerial drone stations which then eventually link down to base stations or just form a completely wireless mesh across an entire country/continent).


Out of curiosity, what happens when someone spray paints the cameras on a car?

Because the car will stop if someone jumps in front of it right?

So it's a sitting target for a gang to vandalize it.


How is that different than slashing someone's tires?

Where do you live, Rio de Janeiro? I don't think there's a city in the U.S. where "random disabling car vandalization for fun by gangs" is actually a thing.


Imagine the possibilities.


Imagine your next Über driver.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: