It's great to hear that companies like Tesla, Mercedes and Audi (and many others!) have been committed in to changing the future of driverless cars and the ever growing progress of technology in this field.
Do you think driverless cars will become the next common means of transport in the near future? Your views are much appreciated :)
Sad to hear, however I believe accidents like these are bound to happen and unfortunately events like these are the ones which will eventually help improve the safety of these cars.This is the same reason how Boeing and Airbus have had to improve features of their aircrafts over time to make sure they keep on correting their faults and ensure the safety of their aircrafts and passengers.
The model is is only a 2 out of 4 in autonomous driving however though (4 being fully automated).
It still is stated that it does require supervision.
Tesla are being unfairly maligned for this (although it is questionable to call the feature 'autopilot' given the audience is the generally moronic public).
As far as I'm concerned the user was using this in an inappropriate manner, and unfortunately paid for it in the same way people who stick their dicks into vacuum cleaners do.
well, thats not an autonomous vehicle (yet!) so you are bound to have problems!
The death toll on fully manned vehicles is higher than 1 as far as I know, I'm not sure what you're pointing out.
Not that the Tesla is even driverless.
Better not fly, because except for take off and landing, most of the flying is in auto pilot.
Why, will they be cheaper or better?
No they will be more expensive, scary to be in, maybe even feel creepy and they will be slow. They will have trouble competing in traffic fairly as human drivers pick on them by jamming in front of them all the time. Will they activate auto-horns? Human drivers will make fun of their beeps, enjoying themselves as proof they beat (and upset) the computer behind them.
The auto companies can make crappier, nastier cars and compete on the quality of the driverless systems instead. They won't update new tech, instead it will commoditise and shorten the life of their products even more.
What will the world look like after we have produced 10 Billion driverless cars? And what will we do with all the old ones?
¯\_(ツ)_/¯
Pffft.
Why, will horseless carriages be cheaper or better than horse-drawn?
No they will be more expensive, scary to be in, maybe even feel creepy and they will be slow. They will have trouble competing in traffic fairly as horse-drawn carriage drivers pick on them by jamming in front of them all the time. Will they have mechanical voice-boxes? Horse-drawn carriage drivers will make fun of their beeps, enjoying themselves as proof they beat (and upset) the machine behind them.
The auto companies can make inferior, cramped cabins and compete on the quality of the combustion systems instead. They won't re-upholster seats, instead will commoditise and shorten the life of their products even more.
What will the world look like after we have produced 10 000 horseless carriages? And what will we do with all the old ones?
I like how you apply previous thought errors to a very similar forthcoming issue. Keep up the good work :)
"As mechanical muscle pushed horses out of the economy, mechanical minds will do the same to humans"
Wow so many up-votes for expensive re-run technology that costs $$$$
On OzB!
What on earth is cool about a driverless car when it jerks you around every time a taxi, hoon, courier, or truck jams in front of it all the time?
And why let companies that have avoided solving the transport problem with such dire consequences (pollution, unrecyclable waste, requiring government subsidies, etc.) keep trying?
When you're old or ill and medically unfit to drive, you'll be grateful that they exist.
Why not get a driver? or Uber?
A taxi ride across town costs over $100 in Sydney. One way.
Uber soon enough WILL be driverless.
@syousef: All the more reason why the solution is public transport so all of us can get around with minimum fuss. Why not improve that instead? Driverless is not cheap or without risk. Why keep building cars? Must repeat history?
@zerovelocity: Yeah I just love going out to the riverland in public transport. I love how it just pulls up at my house while I connect my camper trailer and all the passengers wait patiently for an hour as I load everything in.
So good.
@zerovelocity: Because there wont be any Graffiti, torn seats, junkies, drunks, racists or other malcontents in my future driverless car!
@krisosha: There are Billions of cars on this planet, for non-commuter tasks surely there is one you can afford already.
And good luck getting an autonomous car, or insurance for it that will allow you to tow a trailer, any trailer, for quite some time. And even then, they will probably still run out of power or fail before you get there with that kind of load.
@syousef: "A taxi ride across town costs over $100 in Sydney. One way. Uber soon enough WILL be driverless."
And by then it will be:
$100 for Uber and it's lawyers
$10 for the gubmint to monitor how things work out.
… and still I'll beat you there. Yet, driven as a King you will not be.
It is unlikely you'll be able to work or even update your status as the yoghurt carton-like shell, its light plastics and whining electric drive will shake you to death on most Sydney road. And the Sydney drivers jamming in front of you all the time will cause it to emergency brake and swerve commit all kinds of insanity
Too many dystopian movies. I'm surprised you don't think the cars will hunt us.
With a line like "It is unlikely you'll be able to work or even update your status as the yoghurt carton-like shell, its light plastics and whining electric drive will shake you to death on most Sydney road. " I hope you don't expect to be taken seriously.
@zerovelocity: I'm not even talking about driverless cars, your statement implys that building cars in general is bad, and we should all use public transport.
@syousef: Sure I do.
The difference between dystopia and elbonia is just a matter of perspective:
I've seen Google's driverless cars on flat, smooth roads. The average Sydney pot-hole or gutter camber will have unexpected effects, like forcing them into a crowded gutter, or left turn down a one-way street despite their CPUs protestations.
Besides, noone's ever told me they like using their mobile on a bus. And you don't see anyone using laptops on buses because they'd be donating the contents of their stomachs down the aisle.
None of this stops today's dazed technology consumers thinking about the driverless car experience, like it'll be some bubble of utopian, zero gravity in which it is a breeze to catch up on their email, social or most important tasks.
The reality is totally different. A driverless car in low traffic may seem sorta ok at first glance. But difficult, fast moving commuter traffic will make its occupants physically sick, creeped out, even terrified. And our traffic is nothing in comparison to any in SE Asia. The only thing I can compare it to is being forced to wear an Occulus Rift in a flight simulator and some weird kind of VR piped through your eyeballs.
Mark my words, if these things happen in Sydney, you will see bubble cars bouncing around with the windows getting explosively coated from the inside, though they will probably be so plastered in online advertising you won't see it until someone green clambers out…
I've seen Google's driverless cars on flat, smooth roads. The average Sydney pot-hole or gutter camber will have unexpected effects, like forcing them into a crowded gutter, or left turn down a one-way street despite their CPUs protestations.
You don't think they have potholes in America and Europe where driverless cars have been tested?
The reality is totally different. A driverless car in low traffic may seem sorta ok at first glance. But difficult, fast moving commuter traffic will make its occupants physically sick, creeped out, even terrified.
And you think the first thing people will do with driverless cars is turn them into rollercoasters?
Mark my words, if these things happen in Sydney, you will see bubble cars bouncing around with the windows getting explosively coated from the inside, though they will probably be so plastered in online advertising you won't see it until someone green clambers out…
I'm convinced you must be taking the mickey.
Road traffic speed is very slow for a computer to analyze. If you think it runs out of time when traffic gets a little fast your very much mistaken.
Road traffic speed is very slow for a computer to analyze. If you think it runs out of time when traffic gets a little fast your very much mistaken.
Well there's a straw man if ever I saw one. Where did I say computers are too slow to analyse fast traffic??? I didn't.
@zerovelocity
Baseless presumptions. You can't stop progress no matter how much you want to.
The fact is that they remove a huge amount of room for human error, insurance premiums will drop and they will have an impact on hoons and poor drivers.
Not baseless at all.
We get it wrong all the time. Car software is total crap, car companies don't write or design the electronics that support it properly. It is not implemented transparently, is often left as-is with flaws so long as warranty problems are not caused. Examples: VAG and Mitsubishi emissions scandals. And yes, I could if I had the time could insert a link with your choice of manufacturer here. They are all bad at software, and do not make a great effort to support it. If you still don't believe me, take a look at the code. Why else can so many modern cars be hacked into? You can see youtube vids of Audis being stolen using mobile phones, other cars being driven without keys, or by remote hackers over bluetooth and even Wi-Fi where the car's support it.
You just think that computers can drive better than us. And of course, given time they could.
BUT you make very fundamental presumptions/mistakes about humans:
And on top of all that, they will be s-l-o-w: No gubmint or lawyer will agree to driverless cars going faster than traffic, or exceeding the speed limit to avoid a side impact, or even getting anywhere near the speed limit in case the margin for input error has been affected.
If you don't believe me, get your car's speedometer tested, at the factory they read up to 10% slow. What magic will the computer use to avoid going too slow?
You also made a presumption about insurance. What insurance company will think computers with flaws controlling dangerous weapons in unexpected situations will be less risky?
I don't want to stop progress. On the contrary, its time we dumped cars and started progressing to better modes of transport, esp those involving controlled conditions that limit the number of extraordinary and unexpected encounters and stop generating so much damned pollution to produce, use and maintain.
You want it in engineering terms? Public transport is a different use case.
And don't use terms like "gubmint" if you want to be taken seriously.
@syousef: No, not at all. But where is the use case for billions of new autonomous cars on public roads, retiring the old fleet, let alone one that massively costs, frustrates and eventually kills us all?
@zerovelocity: good points but I know for a fact that insurance is already preparing to release products to meet this era at heavily reduced premiums. Remember that this space is commoditised as far as insurance goes.
Regarding software et al, you're probably right but the security and programming is coming leaps and bounds with major interest in security as a whole so you'll likely see lots of investment in getting it right.
Either way, it's exciting.
I'm not sure about slowness of the vehicles but to be fair, speed limits in Australia are already very conservative and add congestion to that….
@imurgod: Not sure which way you think its exciting or positive. Millions of professional drivers will be out of work, and the roads will be more congested and dangerous than ever as even more use gets encouraged.
No insurance company wants to get left behind in a new market. Uber will insure itself, unless insurance companies offer it cheaper than Uber considers the risk to actually be worth. Why wouldn't insurance companies encourage consumption? They play the long game and can put premiums up at the first hint of trouble. Also, if Uber insures itself like the rental companies and Banks do, they will have to explain revenue and market losses to their shareholders.
Software programming/skills/quality is not a matter of an upgrade. It is a matter of culture, creativity, process maturity and quality control . Those things are not improving one iota as companies race to the bottom in the race to release quick/beat the competition to market.
No, not at all. But where is the use case for billions of new autonomous cars on public roads, retiring the old fleet, let alone one that massively costs, frustrates and eventually kills us all?
Cars don't get replaced all at the same time. There is gradual turnover every year. Our transport is going to have to get better if we're not going to kill the planet. That doesn't mean public transport will fit every scenario.
@syousef: Absolutely. But why are we so keen to make new cars when we have so many already? Is the only solution to scrap and replace every one, simply to fit a new power source, or automation?
And if it is, why do not agree on doing it differently this time, so the problems we caused when building the existing fleet don't happen again. Humans have brains that can solve problems, we don't have to create more.
In a world where cup holders are considered an innovation, we must somehow look at what we've done in the creation of our Frankenstein. We now need to change the way we go about doing it. Just automating the Thing so he will come home on his own is not going to fix the problems he causes when he goes walkabout.
Millions of professional drivers will be out of work,
You sound like a Luddite. We already have millions of jobs that have disappeared, ditch diggers, Zeppelin builders, comptometrists, etc. It's a good thing. Driving is no way to make a living, its drudgery. Free humans up for better things. Oh, and read the Ford press release that came out recently, they are planning to mass produce driverless cars within 5 years.
@zerovelocity: professional drivers out of work, sure. That's sad but such is the way of automation. So many other industries face the same path. The fact is an automated fleet will be cheaper on so many fronts.
Insurance works on risk and this is viewed as a lesser risk by removing human error. It's already been accepted by the insurance industry. The industry actually works the opposite to how you think. They all want to jump on this first and be the name that us at the forefront of it. Cyber liability is the same for example. Lots of risk but premiums paid outweigh it.
Uber currently doesn't self insure and I can't see how they will with their current model but critical mass may change that.
The fact is that in the next few decades, building, finance, medicine, retail, etc. will all be heavily affected by technology. You should see what can already be done. It's mind blowing.
@Gershom: "We already have millions of jobs that have disappeared, ditch diggers, Zeppelin builders, comptometrists, etc. It's a good thing"
Uh no. People actually need jobs. Giving their jobs to expensive machines may not be good, esp. if people do a better job and changing to automated costs a lot, and/or:
These corporations will claim everything in a fight to stay ahead. Ford are barely playing catch up.
The question is, which government can be convinced to let computer-operated vehicles on their people first. The question from that is which person will be killed first. It may be either of us, our friends, family members, or another. Having all these unskilled people out of work might free them from drudgery, but it will also take a living, even a certain amount of life-purpose away from a lot of people.
And apart from the frustration we will all have to put up with once these things hit the road, the points I proffer above are just the costs I can think of right now.
when Teslas fully autonomous mode goes live there is a plan for ride sharing/ taxi type service for people who don't own cars. if you are on holiday somewhere you can toggle in your tesla app for your car to drive itself to subscribers of the ride share service and ferry them to a destination. you get revenue for allowing this to happen, a couple of cars can service an extended family , zipping backward and forwards between them, taking people to and from events.
I don't understand why you think these cars will be slow, they have more info about the road available to them and can readjust and recalculate changing conditions a few hundred times a second.
there is nothing to suggest that companies will use autonomous software systems to outmode existing vehicles and leave consumers wanting another model .
e.g. Musk will eventually switch on full autonomy for three Tesla vehicles,two of which are on the market already and one of which has seen several product revisions as its several years old.
less cars on the road. higher speed public transport that can tell how many people are waiting , adjust routes, make extra stops based on demand - these things will drop the total number of vehicles in play very drastically.
you could have some extra roads that are dedicated for autonomous freight trucks, no humans on the road, trucks travelling at 150kmph or similar and then changing back to normal speed protocols when they exit onto normal roads
They will be slow.
Because the lawyers will have to approve their operation in any condition they encounter. They simply place constraints on their operational parameters. Each gubmint does this too, writing compliance standards before registering cars for use. When they get caght out doing it badly, they just change the rules.
But alas, each accident will not be a forensic investigation. Sure they will download the cars telemetry, logs and everything the User did prior. But unless every accident is overtly due to something common, they will spend years missing the root causes, the programming bugs, poor sensor operation, etc. So all the first gen and 2nd gen and 3rd gen etc cars will be programmed to go slow to give them the best chance of avoiding collisions. And mark my words, all these will need to be thrown away. Have an early Tesla Roadster? You can't even have driver aids retrofitted. The generations will come thick and fast, and the new ones will all get you there quicker than last year's model.
Any driverless car has to be programmed to drive in such a way that it:
Can deal with extreme heat, cold, condensation, ice, rain, extreme dust, dirt, mud, rocks in flight, debris. Not just on the road but in and around the sensors it relies on.
Means the uninitiated, imagine your grandparents for example: Result: Go slow.
None will beat a human in traffic until someone hacks their car and reduces or removes the safety controls. Not much will prevent that on all the early generation cars (Tesla and Prius were all hacked). And the incentive will be there to implement improvements/upgrades/updates so they can be made faster and have increased value. Seriously, few will have security that would stop a 9 year old armed with a tablet.
If we need to build special roads for driverless cars, why not build more efficient, dedicated transport systems that we have been balking at for so long? New roads come at a massive premium and barely delay the inevitable.
Don't get me wrong, I love Elon's approach. Car sharing does really help, and he is doing something at every level he can. But the car industry does not welcome Tesla, and is hard at work strategising with its lobbyists to retain market penetration, retain gubmint assistance and continue to produce opaque, poorly designed, substandard products with little more than designed obsolescence (Business as usual). If Elon ran the industry we might get somewhere, but it wouldn't avoid the problem of 100s of billions of cars we no longer need.
And by the time everyone agrees we have to do something about that problem, we'll have another nty billion more, and no resources left. Even if we turn all the trees, parks, houses and old cars into new roads in the process.
Ford and Holden Rivalry.
I'm waiting for Super V8 Driverless cars!
Coming to a racetrack near you (:
Driverless V8 cars on Bathurst come at me…
Wonder who will win?
Google Roborace.
Lol, I can see it now… skynet gains sentience and the driverless cars chase spectators =P
All in all…. humans lose
It will be the single Nissan Altima L33 since Nissan will reuse their engineers from their high tech Nissan GTR supercar. Yippee, Nissan to dominate Holden and Ford for another decade till they get excluded from winning.
Podium 2nd place goes to Volvo S60 since its accident avoidance system ensures the car always finishes. Yay to the end of spectacular V8 crashes.
The V8 Supercars television series would become marketing to sell more autonomous cars. Buyers will have the idea they're in the fastest, loudest and most uncomfortable commuter and still get passed by cyclists, vans and 50cc scooters. Its a win-win for V8 fans and the safety of cyclists and scooters.
Will there be a 'hoon' mode for all the bogans that will buy one though?
there kind of already is.. the ludicrous mode on the Teslas.. or did you mean something where the speakers make awful noise pollution simulating revving combustion engines and exhaust backfire ?
@funkymonkeyman
Won't change much for supercars. It'll still be a boring race that nobody but bogans care about.
Not sure if driverless cars will be the long term future. It all hinges on the first few legal cases after the first few deaths.
Agree, it is probably too early to call that driverless cars will be the future. However I do see potential and it's awesome to think that something like this already exists. Albeit there is still a long way to go
I think just as Volvo has said, if you aren't committed to take the punishment for fines, injury, or death then you aren't ready to roll out driverless cars.
However, before we get driverless cars, we will get driverless transportation.
Firstly it will be cable cars, then trains, then golf carts in golf courses/theme parks.
It'll probably wash over to small 'n slow taxis.
I doubt we will get full-blown taxis, buses, and trucks legally driving/mainstream any time soon.
Maybe not even in our lifetime.
Having driverless aircraft is much easier simply because of less "traffic" in the sky… but due to the danger of malfunction causing skyscraper collisions, we probably won't have them without pilots.
Then again, the year 2000 has gone by, where is my jetpack and flying car !!
Driverless trains have existed for several years now. I've ridden on a few myself.
@Michael15286:
Yeah I worded the sentence wrong.
The first driverless things we will get (ie no pilot) will be things on rails/lines.
And we've already had those around, but nevertheless, they are the first.
The second in line are small/slow taxis and golf carts.
That's what our current laws and technology can handle right now.
Thirdly, we probably will get "people movers" in the form of taxis, buses, and trucks.
However, not any time soon.
These require more complex solutions as they drive in faster speeds with so many variables.
And we don't have any laws governing this form of transportation yet.
Until both technology, laws (and society) advance further… its merely a dream.
And there is a strong possibility that we may not yet observe this advancement in our lifetime,
…but rather the lifetime after.
Fourthly, we won't ever get "driverless" planes or spacecraft.
It may transport people. But there's a good chance one of the people would be a pilot.
That way in case of emergency, there is a backup solution.
The chance of disaster is too great to risk it otherwise.
Big truck companies will be among the first to adopt autonomous vehicles in any real way. They will be the first consumers to be accepting of the downside risks given that they would see substantial rewards, such as no driver costs, theoretically cheaper operational and maintenance costs due to the computer driving more efficiently than humans, and better utilisation and management of vehicles (no need to schedule breaks for tired drivers).
Given our political situation where it seems no one is willing to make big decisions, trains will likely be among the last form of transport to become fully automated, as no pollie will be willing to take on these risks that industry will.
For the moment, driverless cars are a sideshow, trucks will pave the way and increase public acceptance of the new technology.
Fourthly, we won't ever get "driverless" planes or spacecraft.
It may transport people. But there's a good chance one of the people would be a pilot.
Spacefaring involves highly computational manoeuvres interspersed by extended periods of inactivity. These are both much more suited to a computer.
Not likely, more to transportation then just driving, given that the majority of trucks being sold still dont even have a drivers airbag as standard (kenworth im looking at you)
Our road network as well will make automotous driving limited at best, given the bulk of our road network isnt even actual road.
Though i think the main thing that will stop it all in its tracks is accountability.
Trains/Aircraft are the 2 most likely forms that can be fully automated in the short term, they both have trained operators for fallover/accountability, not for the full operation of the machinery, but even then thats been marred with problems, ie the death of a worker on the automated train network in north west WA
Wait aren't uber already operating driver less cars somewhere in the states ?
What do you expect from the future? Your dog driving a car? lol
I think they are inevitable. Fast Forward a few hundred years and I'd expect freeways and certain roads would only allow driverless cars and be designed for them. It would be far safer that human only control, similar to airplanes.
Thats very much true, car insurance and responsibility of driverless cars will be a big change. Who is to blame when a driverless car has an accident ? The car, the manuf who did the software or the driver. We all know that s/w is never guaranteed for anything so this will be a big change for s/w.
I'd rather catch the bus then have to go around in a driverless car.
You mean you would rather catch a driverless bus? What makes you think that mass transportation will be excluded from this trend?
yup,Tesla already have their bus in the works and its driverless from the get-go
So you'd like to catch a bus, then go around in a driverless car? Wouldn't it just be easier to get in the driverless car first?
I'd rather take the stairs than ride in a driverless lift.
Pretty likely.
First examples by 2020 will be followed by a mass rush to make sure all new models are 'autonomous capable' within 1-2 years (would you buy one that wasn't?) The likes of Uber, Google, Apple, Ford, GM, etc. will shift into Robotaxis in a big way and by 2025 the changeover will be well underway.
The only question then is when they start banning manual drivers from the roads. After all, we KNOW how unreliable and dangerous they are.
By 2030 all kinds of driver as career, be it trucks, buses, taxi drivers, etc. will be on borrowed time - and society will be being reorganised around autonomous transport and delivery.
On august 16 skynet became self aware…..
yes, but they have renamed it TarNet
That's not Skynet. That's clippy. Heaven save us all.
I think they'll always be human assisted (but will drive themselves) as when the shit hits the fan, it's good to have a human brain in control that isn't preprogrammed.
Evidence tends to be the other way; and that's in aircraft where the driver is seriously trained.
Two problems - first if the SHTF and the human is dumped into the loop, they tend not to be able to do much. Second, humans stuff up on a regular basis. You are most likely to be in an aircraft crash when taxiing, when the pilot is in control.
As I say, I think the question is when they start banning human driven vehicles from the roads, probably starting from the CBD and working out.
One day "you will go to the moon"
It's only a matter of time. I can't wait for our heads of state to also be computerised.
We can do that already. Just replace them with this piece of code.
DECLARE @action text
SELECT @action = SELECT MIN(common_sense) FROM actions
EXEC @action
Filthy stored proc. yuck :P
Why not, she's not doing anything
I'm with you on this one OP, definitely believe that autonomous vehicles are inevitable, the real question is when it will happen.
I was initially expecting the technology to require another 20-30 years minimum to become implemented, however it seems more and more likely that we'll start to see the mass production and adoption of these vehicles in 5-10 years.
The benefits are numerous, the first and most important being that autonomous vehicles are going to be the only real way of bringing the road toll to zero. Autonomous vehicles don't drink and drive, road rage, fall asleep at the wheel, get distracted by their phones etc. Remove the human factor and most car accidents will be a thing of the past.
Other benefits may include higher speed limits, no need for licenses (people unable to drive could now use cars without assistance) the ability to use time in the car more productively, the ability to have trucks run long distances without taking breaks, and having your car pick you up and drop you off then parking itself or driving home (which may eliminate the need for many parking spaces, especially in inner city areas).
It may be possible that one day, driverless taxis will be so cheap to use that it may no longer be necessary to own your own car. You just call a driverless taxi whenever you need one and it'll take you where you need to go for a fraction of the price of today's taxis. If this happens, there would be a huge drop in traffic volumes, and traffic jams could become a thing of the past too.
The main downsides are the loss of livelihoods for truckies and taxi drivers, and problems with safety associated with software glitches, hacking etc. I'm sure car companies will put a lot of effort into improving safety between now and when we start seeing these vehicles on the road. Interestingly, a lot of people I talk to don't like the idea of autonomous vehicles because they 'like driving' and don't want this 'fun' activity taken away from them. Hopefully these people will understand one day that giving up driving in order to save lives (by lowering the road toll) is a worthy sacrifice.
Definitely agree with you too. The possibilities are endless with this technology! It really isn't surprising that many people are dissing this concept of autonomous cars, considering its in the first stages production. However, we should all know that ideas like these take time to perfect and I'm sure there will be a time in the future where we would question why we didn't do this earlier haha. Just like how the Wright Brothers successfully introduced the concept of aircrafts, I'm sure it is only a matter of time before we start taking these cars for granted :)
The trouble people have is the time it will take to perfect this technology.
Arguably it has taken 100 years to perfect cars with a combustion-engine. That is hardly innovation. In that time, the Wright brothers work was expanded on so much we got to the Moon and Mars using carbon fuels. And we got to the outer reaches of our own galaxy using solar, nuclear and other forms of propulsion.
Ecologically, cars are our worst enemy. Nothing has caused more damage to the Earth than cars, after us. What does that say for the industry that made them when other fuels got us to the outer reaches of the solar system and with less investment?
The problem isn't really the technology, it is our use of it. Why would you trust that industry, or the governments that support it to go on?
Not only can they not manage the tech they have, they have to manage the transition. Electronic pilots may not seem very bad as a concept, nor when introduced alone, in isolation. But when introduced with other drivers, their behaviours will change. Perhaps it will require new roads and cause massive frustrations for other drivers (who vote). It will be a mess.
Why? It's bad enough being in a car with most other humans driving. Once the computers are doing it, human drivers will learn how well driverless cars see and stop. Traffic will become more cut-throat as these drivers cut in and out more with less margin for error. Collisions with driverless and driver will increase. More serious accidents on motorways will occur as the dumb driverless cars block traffic at a maximum speed of speed-limit less >10% margin for error.
And on top of that, who in their right mind would place their loved one's lives in the hands of a hackable, malfunctional, computer designed by a corporation with dubious claims about its reliability and ability with all its bug-laden Artifical intelligence or machine learning?
It's going to be bad enough when the couriers and truckies are all terminated and replaced by robotic pilots incorporating AI. Time to invest in Cyberdyne?
I see where your argument is coming from and it is quite valid. However, don't you think it is way to early for us to criticise the functions and reliability of these cars? I mean it has barley been a couple of years since they were successfully introduced and it is still got a lot to improve on considering the significance and difficulty of its role. I know it is definitely not the safest thing to travel in right now, however once we give it 5-10 or even more years to learn and adapt to the nature and reality of driving, I'd say it has a very good chance to better the skills and reactions of human drivers.
I still agree that complete AI control is not the best idea right now and human supervision and guidance is still required to prevent foreseeable accidents. However I see great potential in this field and it shouldn't be considered a failure at this stage of its development. It's kind of like how a child progresses in life to an adult; we need to give it some time before we start judging it's competence in this field as it has got a lot to learn and understand about the realistic nature of driving.
I too am concerned about the safety of this idea like you at this stage, however I believe this fear will eventually be overcome once we give it some time to perfect it's role :)
And we got to the outer reaches of our own galaxy using solar, nuclear and other forms of propulsion.
We did? o_0
Apt username. ¬_¬
[@stuhtb]: "However, don't you think it is way to early for us to criticise the functions and reliability of these cars?"
No, not at all. Car companies are barely good at incremental development, cost gubmints billions in support $$$ and do not produce long-lasting, safe or clean products.
We need to take better decisions about the adoption of future tech when it has such an enormous impact. The easy, incremental changes are what we need to look beyond. Humans must learn to work together better to avoid problems, let alone global disasters like cars.
@Scrooge McDuck: Indeed! My bad, that should have been Solar System. All this talk about auto-pilots in natural environment just blows my mind ;-)
All your points are valid but those are more of challenges than disadvantages. Technology has always played the role of a double edged sword. If we look back in time, every big invention had similar challenges especially in early stages. For instance, the early adoption of electricity was much more shocking than one can imagine. Back in those days (as seen on TV) :) , even basic safety standards such as insulation, fuse were missing, a household would try to connect all electrical appliances to single point disregarding input / output specs completely. There was lack of knowledge on how it works, no standard infrastructure for cabling, voltage supplied by various providers was inconsistent and yet today we simply take electricity for granted, isn't it?
Driverless technology has a similar potential as well as challenges. Today there is no infrastructure i.e. there are no network of smart roads or peer vehicles communicating back / giving feedback to each other's cars (but maybe just a matter of time. i believe when cars were launched, there was no infrastructure back then i.e. no roads, signals and maybe the roads were shared by horses, other livestocks, humans and cars. so it wouldn't be drastically different if we see driverless and regular cars on the same road during transitioning stage. There are no standards mandated today but as the technology matures enough, I'm sure there would be various standards which would emerge maybe even at global level. If you believe driving safely means adherence to road rules (protocols), I can assure you that 2 machines / software can follow protocols more efficiently & effectively than humans (at least in most average cases).
You raised a very interesting point on how we we transition into it.
I am not an automobile expert but looking at specs of any decent car these days, one can't deny that there are already numerous electronic modules and relatively small amount of software / firmware already working for you behind the scenes. As a matter of fact some on-board diagnostics are even exposed to user if you hook into the OBD port and tinker around. Even quite a few smart features exist in isolation such as various sensors warning you for objects in close vicinity, smart parking assist, cruise control, smooth braking, reverse camera. Thus I feel, most of the bits and pieces are in place even now. The very first step in transitioning towards new technology should be to get all individual sub-modules connected within the car and facilitate communication between those sub-sytems to start with. Eventually, if we can extend this to say smart roads, communication between peer vehicles, the results can be promising. Of course, there will be threats and security issues but again those are more of challenges and we will learn how to overcome them in due course. This is more of my visualization, not saying it would follow this path but just a possibility and more natural transition path.
Ecologically, cars are our worst enemy. Nothing has caused more damage to the Earth than cars, after us.
I think this is the current situation we are in today but primarily because all / most cars are owned privately and used as luxury/ personal means of transport. Even with cabs, they are hired as a form of private means of transport. However, as @TheElderLegend pointed out if we can build driverless cabs or driverless public transport, it can be more efficient. Public transport currently is categorized as a bus, a train or a ferry. All of them are bulky in size which results in more fuel consumption regardless of number of passengers actually on-board and they also require considerable infrastructure investment, maintenance costs. However, if the public transport is redesigned in hybrid fashion (i.e. small in size but still used as public transport (for instance shared cab), it may yield more throughput in general. Thus cars are not our enemies, how we use them is upto us.
If the automobile industry gets it right, the technology can really be a big plus and has a lot of potential to offer.
I would like to summarize all this with an old adage 'This too shall pass' :) so just hang-in there and embrace what future has to offer.
@ameyas7: We all technology can solve most problems.
However it is the cost, the risk and the lost opportunities that we spend developing it. In this case, if you consider how much time it will take to make it safe,
OBD ports were not added to provide exposure to the user. Car companies were forced by the US government to install them to prevent consumers getting ripped off by dealers locking customers into servicing by being the only ones that could connect to the car's diagnostics.
In Oz, we only have them because they (mostly) didn't bother to take them out for production in the Oz market. (Some have).
Where is the evidence that the cars, the infrastructure and the tech can be implemented before its too late?
Make no mistake, car companies do not innovate to sell a new model. They add features the marketing dept likes and thinks won't up the price at all. It's a race to the bottom tech-wise. Innovation occurs when they have to meet new regulations around ANCAP, CO and NOx, esp. in the US. That is why all cars have airbags and crumple zones but still have internal combustion engines. It is why they came up with high-efficiency diesels that become very problematic/costly and get scrapped in half the time of a petrol engined car. If the car companies were capable of what you hope, we'd have moved to LPG in the 70s and full electric 5-10 years ago. Functional hybrids were proven on Lead Acid batteries in 1994, so this is no stretch of the wonders of tech you place so much faith in.
But trusting them with your life? Not even NASA can keep its people safe under computer control. You think they are going to test well enough to cope with every situation that nature throws at it? Dirty sensors, rattles and malfunctions won't make autonomous cars unexpectedly halt in front of semi-trailer trucks on a regular basis?
Making computers deal with the chaos nature throws at them is not a technical challenge. It is madness, and problems will always occur, and the consumers will suffer awful fates instead of financial loss. And we will take the easy road that ends in a dead-end, instead not solving our transport challenges whilst we still have the time.
@zerovelocity:
of course there will be cost associated with any R&D, both monetary as well as opportunity costs. however, technology grows exponentially. today's smartphones are comparable to a super computer back in 1960/70s. imagine the time, space and other costs associated back in those days but yet we have more powerful, reliable and cheaper gadgets capable of achieving almost similar results just within 3 decades. work on autonomous mobiles has been going on since long time ago and it has been tested for considerable amount of time now. its not yet launched officially for public but if we believe in exponential growth, it won't be too long before it becomes accessible publicly.
Make no mistake, car companies do not innovate to sell a new model. They add features the marketing dept likes and thinks won't up the price at all. It's a race to the bottom tech-wise. Innovation occurs when they have to meet new regulations around ANCAP, CO and NOx, esp. in the US. That is why all cars have airbags and crumple zones but still have internal combustion engines. It is why they came up with high-efficiency diesels that become very problematic/costly and get scrapped in half the time of a petrol engined car. If the car companies were capable of what you hope, we'd have moved to LPG in the 70s and full electric 5-10 years ago. Functional hybrids were proven on Lead Acid batteries in 1994, so this is no stretch of the wonders of tech you place so much faith in.
well the good news is autonomous mobile is a marriage between automobile & software / embedded technology. apparently, the techies and geeks are more excited with this space than the folks from automobile side of it. thats the main reason why tech companies such as google, uber are keen to dive into this space. hence i feel that the pace in this area isn't just bounded by automobile industry alone. of course, they need expertise from automobile companies but i feel the push is more from tech side of it. i could be wrong as i have never worked with an automobile company or don't have any idea about the work culture in that space but i believe it would be more of factory culture in general. the real breakthrough innovation only happens at the very top say F1 or thin R&D departments and then all other departments simply implement the designs / assembly line (again thats just my perception of automobile industry). having said that, automobile industry has come a long way as well.
But trusting them with your life? Not even NASA can keep its people safe under computer control. You think they are going to test well enough to cope with every situation that nature throws at it? Dirty sensors, rattles and malfunctions won't make autonomous cars unexpectedly halt in front of semi-trailer trucks on a regular basis?
Again if we go back in time, when cars were launched initially, there were no seatbelts, airbags on day 1 but eventually every technology matures and if you believe in exponential growth, it won't be too long before this technology would settles down. Imagine when they installed traffic lights on the road and maybe they were not 100% reliable initially but still the technology survived. I don't think it will be tested enough to make it 100% full-proof but that goes back to your original Q on how will we transition into it, is very important. as i mentioned earlier, people would feel more comfortable if the first step is to simply connect various subsystems in the car and start leveraging from there. for instance, if you detect any object in X metres vicinity, apply brakes gently. at times i wonder why the automobile industry came up with air bags before inventing something that could detect nearby objects & stop the vehicle to prevent collision in the first place. yes airbags play crucial role but personally i feel preventing would reduce the impact by huge margin than just trying to reduce it after the collision. maybe it was challenging to detect nearby obstacles back then but is definitely not the case today. also as part of transition you would have some human driver who would override the auto-pilot mode until the technology matures enough. i believe thats what most automobile manufacturers would do in the initial phase. if you are concerned about technology failing down in the middle of the road, today's cars are not 100% safe in that regard either but we still enjoy the ride than getting paranoid about such breakdowns. accidents do happen and will continue to happen (sad but true) but as the technology matures enough, the tolls would go down drastically.
Making computers deal with the chaos nature throws at them is not a technical challenge. It is madness, and problems will always occur.
Computers never deal with anything directly ;) all algorithms are written by humans and as they say "to err is human" :) but that doesn't stop us developing new applications, learning new technologies. Machine learning is no longer a fiction. It is still not available for public use but again remember the exponential growth? :) the main reason why i feel driverless technology would be successful is because if you draw parallels to computer space, roads are like your network links, cars are like data packets and road rules are like protocols so if everyone adheres to them strictly, chances of chaos arising are lot lesser than chances of chaos arising when you allow 2 human beings to interpret the rules in their own way and rely on each other's judgement to share the resources/ space. sometimes the other person might be just new in town and is not even aware of rules completely. even driving style (aggressive vs passive) of 2 drivers are not the same. some people might be just ignorant or impatient to follow rules. human beings have their own limitations too.
to summarize, yes there will be challenges and it will take time for this technology to mature but how we transition into this technology can make or break things. but overall, personally i feel this will open doors to lot of opportunities not just in transport of human beings but even other areas like even disaster response where it is too dangerous to send rescue team directly.
@ameyas7: You seem to think I'm saying it can't be done, when in many ways, it already is.
Anyway I'm not arguing that they can't do it. I am however saying they can't do it easily, without considerable costs (that we will all need to bear), and in a way that is not tightly integrated with rules that in nature cannot be guaranteed. In other words, well. And worse, when failures happen happen the results will be unexpected, extraordinary and insurance and fault will become a legal minefield where costs spread to other parties than the parties that benefit from the technology use.
Your missive does a great job of making a religion out of technology and Moore's law, but the fact is that not all technology effort use and abuse is good. And little they do will resist;
I'm just questioning this 'advance', because consumer driven technology does not mean that it is good, or that it can't be badly deployed, badly implemented, over-used, abused, etc. It's success just means enough people like the idea of it. After all, there are huge holes all over the world full of products people discard after one use and others full of products that were never bought.
We can only dig so many holes, and we can only make so many automobiles. Which should we spend the next 20 years making? Ones that drive themselves or ones that are cheaper, can be maintained, recycled and that use energy more effectively?
THERE IS AS YET INSUFFICIENT DATA FOR A MEANINGFUL ANSWER.
(a cookie to anybody who recognizes what I'm referring to :)
How do we claim our cookies? Do we have to wait for the heat death of the universe?
I have shipped out cookies using China SpacePost, and have given tracking numbers to those who won.
Cookies provided courtesy of GearBest.
Shipping estimates are about 50,000 light years, or when the Earth has been invaded by Borg cubes, which ever is sooner.
Shipping estimates are about 50,000 light years,
So via the centre of the galaxy? ¬_¬
im actually thinking it will go more the route of cars similar to that of the film irobot , in the sense that there will be automation in specific situations such as highways/motorways but still require manual operation in most other areas.
so pretty much what tesla is doing right now , just better.
Uber will start using them this month in Pittsburgh. Will be an interesting trial to see how well it works. Keep in mind this is Pittsburgh in summer. I don't think these self-driving cars have any chance in the winter with snow.
As mentioned before, Uber were going to trial self-driving cars in Pittsburgh. That has begun with 2 engineers sitting in all of these cars, 1 in the drivers seat and another in passenger to monitor. Engadget Article
Seems the automated driving works most of the time but things like a double parked truck and 4 way stop signs required manual intervention. For anyone who has driven in the US, 4 way stop signs can be daunting to figure out who goes (basically first come, first serve). Also a challenge will be the winter in Pittsburgh which is a mess and moreover leaves cars quite dirty when you mix in the dirt, salt, snow, other crap on the road. I imagine you'd have to keep the sensors constantly clean.
We'll see how well they can program around these things.
nocure on 21/08/2016 - 20:57 Comment score below threshold.
