It depends what the software is for. In video games, or some kind of desktop app? Sure.
But, for example in my job, I make a control system for a medical device. The validation procedure is extremely long and thorough. Longer than the development time in fact (not counting the verification testing that happens in parallel with development). Things don’t just “get in”. And any debug tools or codes only exist in custom debug versions of the software; not the version that will be released or is submitted for validation.
Automated cars are even more safety critical while at the same time being much more visible to the general public. A single cheat code could well put a car company out of business even if it didn’t result in any casualties.
There won’t be cheat codes.
You do realise what some people are willing to pay for cars right now don’t you?
For bragging rights to own the first self driving car?
You remember what people were willing to pay for the first iPhone?
How can you say something like a LaFerrari is “worth” what it costs? How about a P1? Or an Aventador?
There are almost a billion vehicles in the global fleet. Dominating the market in driverless technology will pay for the few millions (maybe tens of millions) that Google has plowed into this project. And once it’s widespread, the market might be a very long term one. We’re not going back to driving ourselves after 20 years of being robotically chauffeured.
You think human beings are better at driving through ice and snow? Being from the Northeast, I tell people that driving in ice and snow is very easy – until you have to turn or stop.
Cars can tell if their tires are slipping. It’s why anti-lock brakes work so well. A car can detect much more quickly it’s losing traction, and can start pumping the brakes way before you, the driver even realize what’s going on. Then, it can pump the brakes faster and more efficiently than you – timing the pattern to maximize friction and keeping the vehicle straight.
I can imagine a car chugging ahead in a blizzard carefully tracking the slippage of its tires, and being able to see better using various sensors. Talking to the cars ahead about driving conditions, and anticipating what may happen.
As for stoplights, they’re mainly for us humans because we can’t get through an intersection very easily with cross traffic. That traffic light where you passed the truck might never had been there if all cars self drove.
Imagine two cars approaching an intersection on a collision course. The cars contact each other, figure out they’re on a collision course, and one car will slow down while the other speeds up. You’ll miss each other with more than six or seven inches to spare as your car careens through the intersection at 75 mph!
Okay, I’m not sure whether I want to be in one of those cars. When I took my kids to Disney World, I told them the only “rides” I would be getting on were the escalators, and only the slower escalators at that. However, driverless cars could talk with each other, figure out where collisions may happen, and then adjust their speed to avoid them. And, unlike people, one car isn’t going to complain “Why do I have to be the one to slow down?”.
Look how inefficient cars are. You have four or five seats in a standard sedan. You have eight or nine in a SUV or minivan, but 90% of the time all those seats are empty. And, how often do you drive? Every day, but 90% of the time, your car is sitting there waiting to be used. You need a slab of concrete by your home to put your car when you’re home. That space is still there when you’re at work and sitting on another slab of concrete just for your use. The store has another slab of concrete just for you to park. Malls usually have massive amounts of spaces that are never used except for Christmas shopping.
Driverless cars could solve that. Instead of sitting there for 8 to 9 hours while you work, your car could be picking up and dropping off other passengers. While you sleep, your car could be running about and making you money. Instead of all those cars during rush hour with one person occupying five seats, driverless cars could do very efficient carpooling. Picking up and dropping off passengers while you go off to work. There would be 75% fewer cars on the road.
And driverless cars are coming around pretty fast. In 2003 DARPA had a contest where driverless cars had to go down a course. None made it more than a few hundred feet. They did a bit better in 2004. In 2005, quite a few managed the course. In 2007, DARPA had an inner city driving challenge. Almost all contenders made it through. In 2010, Google started on self driving cars. In 2013, they gave a blind employee a car for commuting to work. In 2014, they developed a driverless car without a steering wheel.
Unlike flying cars, driverless cars are coming. Flying cars were trying to solve a problem that didn’t really exist and their cost and complications simply didn’t outweigh the benefits. Driverless cars, on the other hand, will save tens of thousands of lives lost to traffic accidents. They’ll speed up transportation and make it much more efficient. Land use will change as vast parking lots could be made into businesses. No more parking garages. No one running around looking for on street parking which according to one study was responsible for 1/3 of the traffic in San Francisco.
I wouldn’t be surprised if the first self-driving cars are available in 6 to 10 years. I wouldn’t be surprised if 20 years from now, all cars must be self driving.
Agreed, and the DLCs (do I get credit for creating the acronym?) will probably be programmed to leave a good bit more space between themselves in those situations so that people don’t have heart attacks.
I expect DLCs will first be widely used on highways. If you’re on a long trip in good weather, there’s no reason you should have to be focused on driving for hours on end. Put on Otto the Autodriver and take a nap, read a book or chat with your family. It’s also a much simpler programming challenge than city traffic.
This won’t be the major niche for self-driving cars. We already have something that does that, called taxis. And while some people do use taxis, they’re far less popular than personal cars. People get their own car because they want to have it whenever they want, without waiting for someone else to vacate it. They want to be able to leave their stuff in the car securely. They want to be able to decide how much wear and tear to put on it, and possibly to make custom modifications to it (even if they don’t, they want the freedom to). All of these point away from sharing the car with strangers.
Driverless cars are an exciting concept. I wish they were already in use everywhere, and I’m sure they will be, eventually. I’m just curious about how they will be introduced into the market.
This makes sense; similar to Amazon’s initial business model.
Yes, there is already a waiting list for the first lunar vacations.
Nobody ever paid more for an iPhone than they did for their house.
Ferrari, McLaren, and Lamborghini make their money selling high end autos that are valuable, in part, because they are rare. Mass producing these cars so that everyone can afford one is not their goal.
I agree. There will be no going back. What I’m wondering about is how do we get from “Don’t have” to “Can’t live without.”
As great as it would be to be the first kid on the block to own a driverless car, the real advantages of this technology depend on there being an entire fleet in operation. Owning the only driverless car will reduce the number of drunk drivers by only one. The aerodynamic efficiency of closely grouping cars requires a group. Communication between cars to avoid collisions and to navigate intersections is not a solo exercise.
I expect that the early years of the driverless car industry will include a combination of all the above ideas. However it’s accomplished, a transportation system with all driverless vehicles will someday seem as commonplace as cell phones and the internet; and as life-changing.
This is true, and it may be significant in the development of driverless cars. Most inventions are intended to be new solutions to old problems. It’s important that the inventor understand what problem needs solving and the existing solution with which the invention will be competing.
One previous invention was predicted to be “bigger than the PC”, “bigger than the internet”, and “will make walking obsolete.” But marketing the Segway as a replacement for walking overlooked the obvious (in hind sight) limitation that most people’s walks don’t start and stop near a convenient place to store the machine. It turns out that the Segway was really competing more with the bicycle, which was bad for the developers because most cyclists think of exercise as a primary benefit of riding.
I think it’s going to take a lot longer.
There’s a huge market for robo-cars, especially as the baby boomers age; But there are also huge technical problems. Take for example, what happens when a robo-car encounters a motorized wheelchair and a duck.
(The car stopped in place, and refused to move when it saw a person in a motorized wheelchair, waving a broom, as she chased her pet duck which had run out onto the street.) Yes, it’s a humorous image—but it reveals a huge problem: what to do when the car gets confused.
Now, a wheelchair a broom and a duck moving erratically in the street makes you laugh…But what if it had been ,say, a child on a bicycle, using a rope to tow his friend on a skateboard, who is holding a dog on a leash? These things happen in real life.
Or how about this problem: programming the car to break the law. Who is going to take responsibility for that?
Here is how I violate traffic law every day, and I want my robo-car to do the same:
Where I live, the law says that a solid painted stripe in the middle of the road means “no passing”, and no crossing the line. I live in a quiet neighborhood, and the road leading out of the neighborhood to a busy street has this type of stripe painted at the intersection. But the person who lives in the house near the corner parks his truck in front of his house on the street, near that stripe. The truck is just wide enough that I have to swerve a bit to pass around it—and when I do, one of my wheels crosses the painted stripe ,several inches into the oncoming lane. This is a clear violation of the law—which my common sense knows to ignore–but how will a robo-car deal with it?Will it stop in the lane and refuse to move? Can a corporation knowingly program the car to violate the law?
These are bugs in the programming—but until they are fixed, there are going to be a lot of problems with self-driving cars. The first time you get stuck in your car which refuses to move…you’re going to be a very,very angry customer.
So, you spend 10 minutes thinking about potential pitfalls in the programming, and therefore conclude that companies with thousands of people working on these projects will take decades to understand and work around the same pitfalls?
Powers &8^]
I appreciate the lack of mocking, but no, I wasn’t kidding.
I mean, I realize we’re talking about fairly fantastic technology here, but as long as we’re discussing the way the technology works, we might as well be as accurate to the canon as possible.
Powers &8^]
Thousands of people have been working for decades at Microsoft–and they still have to release service packs to fix their many,many,many mistakes. When your computer crashes,you face the Blue Screen of Death, and it’s upsetting. … When your robo-car crashes, you’ll be even more upset.
So, when confronted with an utterly bizarre situation that the programmers couldn’t possibly have anticipated, the car still manages to make the correct decision. And you view this as a problem? What do you think the car should have done, run down the wheelchair? When there’s an erratically-moving obstacle in the road, you should stop and wait for it to pass. When the Google car encountered an erratically-moving obstacle in the road, it stopped and waited for it to pass. The technology works.
But see, the Google car already has lidar. Radar, sonar and infrared aren’t going to help with the computer vision problem any more than lidar. Water falling in front of a fine grained version of any of them is going to cause an enormous amount of data flood in. The computer vision problem is primarily one of establishing a figure/ground relationship in an image that changes constantly. The Google car accomplishes what it does accomplish through it’s already-made, very detailed, lidar map. It must be maintained to be useful, and isn’t useful when the machine is overwhelmed by noise such as rain, snow or a bunch of blowing leaves. The human brain deals with this problem fairly well with its limited sensing resources, and in real time. I’d imagine a good portion of persistence of vision comes from the fact the brain understands what it sees, as far as a human understands anything. As of now, the computer equivalent is absolutely science fiction.
As I’ve said in other threads, the Google car is a good fair weather autopilot in it’s limited map, which is somewhat an example of Google brute forcing reality with money. I’m happy to hear that Elon Musk agrees with me that a good autopilot is possible if you throw enough processing power at it, but a truly autonomous car* doesn’t seem possible with any permutation of current technology.
*My standard for this would be a car that can drive anywhere I can drive, and I can be morally excepted from taking a nap for the trip. Morally excepted would mean that I can’t receive tickets for the car holding up traffic, speeding or doing other such nonsense because it has become confused. Until then, I’m at least going to have to put down the banjo and drive once in awhile.
Imagine the dystopian possibilities. Do the authorities require your arrest, no matter how petty the offense? Instead of having to send an arrest team to your house, which has to be armed and wary of being shot (this is why they don’t often arrest for petty crimes, they just stick a warrant on your record and wait for you to get pulled over for a traffic stop), they can just have the driverless car take you right to jail.
Jailing people is expensive. They could possibly have a driverless car take you to work every morning and return you to jail in the evenings, with all of your salary going to the state.
Want to run from the law? Good luck driving a getaway car if it doesn’t have a steering wheel/has a remote kill switch.
A different kind of car chase perhaps, but still potentially good.
So the perps / heroes either have a manual drive car or a custom automated car that’s not on the network. But the cops can utilize every unoccupied car in the city in helping to bring the chase to a speedy / safe conclusion.
Granted, it removes half of the peril, but it would be pretty nice seeing how the heroes get around road blocks everywhere, trains of cars forcing them off the road or…I dunno…cars trying to attach themselves in some way and then be dead weight.
