Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Google’s New Robot Car Raises Hopes, Reality Will Dash Them Soon (singularityhub.com)
25 points by kkleiner on Oct 12, 2010 | hide | past | favorite | 22 comments


Meanwhile, thousand of lives will be lost every year because we do not embrace this technology with speed.

It's insane that society does not embrace the technology, even with all the authoritarian risks.(autonomous cars will probably make tyranny more efficient, but at that point we're doomed anyway)

I hope somebody is dedicated to the task of speeding societal acceptance.

Maybe somebody should create organizations like "Parents for Autonomous Driving" and mobilize public opinion in favor of autonomous technology.


Nay-saying is easy. Making progress is hard work. Kudos to Google for pushing on this front. The author speculates about the state after a decade. By that time, this article will be forgotten, but google's work might still be remembered.

a) It's acceptably safe to sit on something like hundreds of mini explosions every minute. (hint: car) b) We actually have some atom-splitting power plants running c) Planes have auto-pilot modes and fly-by-wire. d) Trains can run without drivers.

By the time this technology gets into regular use, we probably won't even be aware that it is being used and think about it as some sophisticated form of cuise control.


Very well written argument, I was so excited when I saw the video it takes a few days to sink down to the practical reality that is our sue at the drop of a hat society. Was driving a hire merc recently for 6 days through Germany and was amazed (I use the tube over a car normally) at some of even the low spec advances like gps that knew roadworks, rear camera when reversing, colusion detectors, auto lights when dark even during the day, auto windscreen wipers set to perfect calibration. Like NASA research googles work could really benefit society and save lives even years or decades before a truly automated car


Can anyone think of ways to speed the sociolegal acceptance of this technology?

At the end of the day, I have a feeling it won't really come down to lives saved even if that's how the media spins it. At the end of the day, it has to help pad the bottom line of one of the car companies and they will then have the incentive to lobby Congress and get the right regulations inserted into the right bills.


Can anyone think of ways to speed the sociolegal acceptance of this technology?

Perhaps it could first get used by people who are visually impaired, or for some other reason are not able to currently get a driver's license.

Given the importance that being able to use a car to get places has on living a normal life (in most of the U.S.), one could then frame it as a human rights issue.

Once some people are using such cars, it's a much smaller step to allow other people to also use self-driving cars.


Once we have fully automated cars that obviously don't need the manual override features to be used, the spirit of the Americans with Disabilities Act says that it is unfair to deny those with disabilities that would prevent them from obtaining a driver's license the right to operate the automatic mode of such cars.


Age. There's a growing segment of the population who does not trust itself to drive (or is not allowed, if medical exams are involved in periodic driving license renewals).

But I'm sure they would like the freedom of movement.


I believe Google are potentially missing a trick. If they are developing an automated car that requires little to no human interference then surely the design would be completely different if all cars on the road were automated.

The biggest hurdle appears to be human error. 'What about drunk drivers who drive on to your side of the road etc'. My question is, what if all cars are automated? Zero drunk drivers, minimal risk of encountering unpredictable driving behaviour, much easier and safer to implement.

It will get to a point, during our lifetime, where humans controlling a car are limited to private race tracks. Speeding, drink drivers, poor driving ability, insurance premiums, all of these will push automation to the forefront in no time at all.


In what sense are Google "missing a trick" here? Before there's any prospect of getting all cars automated, car automation has to be good enough that that's a credible suggestion.

Google has (presumably) good lobbyists and plenty of money, but persuading legislators to require all motor vehicles to be automated before demonstrating safe automated motor vehicles would be a mighty tall order.


car automation has to be good enough that that's a credible suggestion.

I completely agree. I suppose what I was trying to get at is that the direction Google seem to be taking this is one where it will end up as an amazing feature on top end luxury cars whereas the technology could be very different if evey car on the road had the same technology and each was constantly aware of every other vehicle in its proximity thus reducing the risk of human drivers interfering with a predetermined path.

The latter sounds like it would require additional or even very different technology to implement.

That's my two cents, keep in mind I know more about the internal combustion engine than I do about programmed automated technology!


The transition will likely take dozens of years so the first automated cars will have to be able to handle human error.


I doubt it will take that long.

30 years ago the vast majority of cars didn't have airbags, ABS or even ECU's and technology develops exponentially faster now than it did in the 80's so in my opinion we will see the first mass production fully automated road legal vehicle within 10 years.


Even if the technology was perfected and allowed within 8 years there would be a massive generational gap. I doubt few people older than 40-ish would even consider switching it on. As a fallback safety system it makes a lot of sense though. It should be possible for the car to detect when the driver is making poor choice due to sleep depravation, inebriation, etc. Eventually it would detect when the driver has completely lost control and park itself.


Eventually driving on your own will be illegal for the exact same reasons driving drunk is right now, on a relative basis.


There may also be some pressure on the busier roads to force the use of the automated mode, once it can reliably space cars closer than can safely be done with manual driving, and once a large enough fraction of the cars on the road support automated driving. (Presumably we will see a gradual decrease in following distance in automated mode, with lots of careful investigation of near misses along the way.)

I've found myself wondering if Google will solve the capacity problems from New Jersey into Manhattan faster than the new rail tunnel(s) that may or may not be getting built; the existing tunnels are pretty full, but it looks like Google has figured out how to dramatically increase the capacity of the existing road tunnels.

But the other question is how parking will work out. Certainly adding lots of parking spaces in Manhattan will be expensive. What will happen if there is an argument that New Jersey income tax revenues can be maximized by maximizing the number of New Jersey residents who can commute to Manhattan, and that the way to accomplish this with the lowest infrastructure cost to taxpayers is to allow the cars to drive themselves back to New Jersey to park for the day?


Google shouldn't sell these cars to people. They should provide a sort of automated taxi service. You request a specific destination at a specific time and the car picks you up and takes you there. Then the parking question could be avoided entirely. Having thousands of cars sitting doing nothing all day is a waste.


An automated taxi service is certainly one of the things we ought to get out of this navigation technology (but you don't want to call it a taxi service, because you don't want to be subjected to the limits and fees placed on taxis by municipalities), but apparently if you look at commuter rail systems, you find that people tend to commute from suburbs to downtown in the morning, and back out in the evening. Commuter rail systems usually want to have separate overnight layover facilities on the outbound ends of routes, vs mid-day layover facilities in the downtown areas. You can force the cars to rack up empty miles, but it's probably more energy efficient to have separate mid-day and overnight parking.

The other thing the automated taxis should offer as an option is cars with multiple compartments, where if you wanted to get from one compartment to another you'd have to get out of the car. People who are impatient but don't care too much what their trip costs could request that they get one compartment of the car and it get routed so that they aren't delayed by other people in sharing the car, and then people who want cheap transportation and don't care so much how long it takes could book the other compartments. You could also have a hybrid option where the impatient would allow up to 5 minutes (or any number of minutes of the rider's choosing) delay to allow others to be picked up/dropped off, and there could be transfer points along major roads, and people wanting cheap travel could express a willingness to deal with N transfers. That could go a long way towards increasing the number of people in the average vehicle.


Would this be a bad thing, if you can get to everywhere you need to go? I certainly wouldn't mind if that were the case.


So do we really need an automated car at all. Our roads were designed for humans as are the signs and lights. I think people are investing way too much into the prospect of an autonomous car and no one is really asking for it.

Google and its researchers are smart. So these concepts are exactly that concepts. But what gets put into a car is completely watered down. Before any autonomous car rolls out the factory I'm sure we'll see a semi-autonomous car.

We don't need autonomous drivers we need better drivers. A car that learns how to handle adverse conditions and can assist the driver is probably more valuable than a car that can do it alone. I believe this is where the real payoff is. Mercedes-Benz already has S-matic in certain cars that can determine the speed of cars in the front and behind and brake accordingly. We've had ABS breaking for decades to assist in hard braking, traction control to assist in tight turns. So where is the car that can keep me on the lane in severe fog, help me drive over black ice, through a blizzard or torrential rain. That's the type of automation people want.

If I needed car that could drive without me I'd call a cab.


I'm sure that the many elderly and disabled individuals who are currently unable to drive themselves would be very interested in a fully autonomous car. If they're not asking for it, it may only be because they're not aware it's something possible to ask for.

In many areas of the US, not being able to drive is a pretty major impediment to quality of life. This technology could transform many peoples daily lives.


Elderly?! Hah, it practically takes a court order for the elderly to stop driving. In Florida, it's quite common for a pedestrian to be hit and killed by the elderly. Then the public outrage commences for a few months. Laws are crafted and then die in the state senate. Why? because the elderly don't want someone restricting how they drive and they constitute the largest (and getting larger) voting block.

The disabled market I could see going for this, but if Dean Kamen can't even sell a 5000 stair climbing robot wheelchair to this market how is a $50-60k car going to fair? Besides, if someone is so disabled that they can't currently get somewhere they probably aren't in a position to get in the car in the first place. I'm specifically talking about the people that require ambulance service.

And people always miss out on the human element of these technologies. You know why the FAA don't allow UAVs in controlled airspace? Because the pilot unions won't allow it. Currently, these cars are only allowed on the road if a licensed driver is in the car. Well the people that need them can't get a license. Think that's going to change? Not if the taxi, limousine, or trunk drivers union have a say about it.

Google may have lobbyists, but they won't stand a chance against the AFL, CIO, or Teamsters.

So my hypothesis still stands. You'll see assistive technologies long before you'll see a mass-produced automated car.


Everyone want improvement, it's the change they fear




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

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

Search: