Dan Dodge, the founder and former chief executive officer of QNX, the operating system developer that BlackBerry acquired in 2010, joined Apple earlier this year, the people said. He is part of a team headed by Bob Mansfield, who, since taking over leadership of the cars initiative – dubbed Project Titan – has heralded a shift in strategy, according to a person familiar with the plan.
The initiative is now prioritizing the development of an autonomous driving system, though it’s not abandoning efforts to design its own vehicle. That leaves options open should the company eventually decide to partner with or acquire an established car maker, rather than build a car itself. An Apple spokesman declined to comment.
This whole thing of Apple designing, building, and selling a car still seems so extreme to me – it feels like jumping the shark, really – but at the same time, it could just as well be the genius move that prolongs Apple’s winning streak for decades to come. I have far too little insight into the car industry to say anything meaningful here, but it does fascinate me that a technology company like Apple is presumably entering the car market.
The car industry has never been profitable. Every car maker relies on some form of direct or indirect subsidy.
I have friend who is an aerospace engineer with a PhD in computer vision. She told me that truly autonomous cars are at least 30-40 years away.
There already seems to be a major problem of the marketers getting too far head of the technology and promoting immature emergency safety systems as automated driving. I read a recent test where the autonomous lane change system tried to drive the car directly into a concrete barrier wall at 100 km/h.
unclefester,
I’d honestly like to hear why she thinks that. We’ve already made great strides. Of course failures can and will happen, but personally I wouldn’t be surprised if automated cars are statistically safer than humans by the end of the decade.
IMO many of the so called ‘autonomous’ features such as ‘hands free’ and automated lane changing will either be banned or highly regulated in the near future. They have been widely and irresponsibly promoted as automated driving features instead of driver aids or emergency safety systems. This has made drivers complacent.
Tesla is by far the worst offender and actually promotes dangerous and inattentive driving with misleading promotion of the autonomous capabilities of their vehicles. A Tesla recently drove at high speed into the side of a white truck killing the Tesla driver. The Tesla vision system was unable to detect the white truck against a bright sunlit background. The Tesla ‘driver’ apparently wasn’t making any attempt to control his car. [He may have been watching a video at the time.]
.
‘Autonomous’ cars only work on smooth roads with well defined lane markings. They have difficulty reading traffic signs and can’t understand hand signals.
To work properly autonomous cars need an aviation-type safety system. That requires uniquely coded radio transponders to constantly share speed and position data with each other road users (including cyclists and pedestrians), machine readable (electronic) signage and lane markings and some form of central computerised traffic control system. That essentially means one standard global system will have to be adopted by all vehicle makers and road authorities.
In other words the manufacturers will probably have to start again with a clean sheet.
Me, I’ll be impressed when autonomous cars can handle snow-and-ice covered roads where you can’t even see where the sides of the road itself goes, let alone lane-markings, and can predict and handle patches of ice and packed snow like a human can.
I do not see these cars doing the above any time soon.
WereCatf,
To be fair, bad/missing markings and bad weather are some of the worse scenarios for human drivers as well, it’s not unusual to see cars in the gutters on unplowed highways. An AI could be programmed to handle millions of scenarios better than an average human, including those bad weather scenarios I suspect. To me the question isn’t whether we can do it, but whether we’re willing to commit towards the funding and resources necessary to do it.
A computer might potentially have an advantage with better proximity sensors/peripheral vision/reaction time compared to a human.
Edited 2016-07-29 04:46 UTC
I didn’t say it’s impossible to ever do, I said it won’t be happening any time soon.
Human drivers do better without road markings because they drive more carefully.
http://www.bbc.co.uk/news/uk-35480736
In some situations. In many parts of the northern US, not seeing lane markers is pretty dangerous in in-climate weather. Especially in mountainous territories.
I am sure it is a case of “depends on the situation”. But most roads on the UK are those sort of roads.
lucas_maximus,
The link says it’s the slower speed that makes it safer on small roads less than 30mph, many local residential streets aren’t painted here either.
But doing that in a high traffic commercial area would result in mayhem
Anything more complex than a simple low speed 2 lane road and I think painted lane designations become necessary, otherwise nobody would actually know where they’re supposed to be.
I agree with Bill Shooter of Bul’s post that when snow covers the markings it becomes a dangerous free-for-all, especially if there’s enough snow to hide the road boundaries themselves. Driving like that for any prolonged period is nerve-wracking.
Edited 2016-07-29 14:56 UTC
Those are more usual situations in the UK btw.
I would guess that the sensors and reaction time will be better on self driving cars than humans.
That’s part of the problem. If you react too quickly in icy conditions, for example, you’re going to skid and have a wild ride indeed. If you’re lucky, that’s all that will happen.
Speed isn’t everything. It’s about the appropriate reaction based on prediction of what will happen. Computers are darn good at reacting when the conditions are preset and precisely match what has been programmed, but machine learning isn’t nearly to the point that prediction of road conditions based on past conditions will be anywhere close to accurate yet. That’s not to mention objects, construction, fallen power lines…
darknexus,
That’s true, but computers have another advantage in that they can effectively learn in parallel. A human driver must gain experience slowly in person over time. With computers we can send out hundreds of drivers for one or two years to accurately capture real driving conditions and build a neural network.
I realize I’m simplifying a bit here, and it would take a while before the technology is both practical and affordable, but at least in principal the resulting AI would have hundreds of years of driving experience and might arguably be more qualified than any single human being.
Edited 2016-07-29 20:00 UTC
I thought Google’s car could understand hand signals?
https://www.engadget.com/2016/07/05/google-autonomous-cars-cyclist-f…
Birds do fly at high altitudes
Geese have been recorded at well over 10,000ft when flying between Iceland and Scotland.
As for the Apple car.
If they are working on a car, they will eventually break cover and buy a smallish car maker.
If they don’t then they would have t oget someone else to make if for them like Foxconn does for the iPhone.
Building cars is far more complicated than a bit of electronic kit.
The Documentary that the BBC did last year about how a Mini is made shows that very well.
They will need someone to build the thing for them (whatever that thing is).
I’d be looking at small/medium makers or subsidiaries of larger companies that might be looking to off load a plant or two. That way you might get wind of an Apple takeover in the offing.
There again, it might get canned tommorow.
I don’t think they will go down the fabrication route. I think they will make a design they can licence that other manufactures can build for them.
Jumping into the car industry and making a diesel doesnt make sense. Apple would need to develop an autonomous electric car to be a viable option in the market.
The worst thing about it is computing (and in particular Apple) already has enough car analogies without having an actual car for comparison…
Yeah, even Chrysler is looking for another auto manufacture to build cars for it. It would be a bit of a dream for Auto makers, to work like Apple where the manufacturing is done elsewhere by low wage non unionized employees and the bulk of the profits go to the designers.
I’d be surprised to see Apple actually buy someone like Chrysler, but so many before them have made the same mistake: Daimler, Blackstone, Fiat.
The first personal computers became available in the mid to late 70’s, roughly 40 years ago. Do you really think that a person at that time – even with a PHD level engineering degree – could have accurately predicted the date of arrival of mobile computing devices? The very notion of a technology forecast 40 years from now is patently absurd.
Elon Musk has said he expects fully autonomous capable vehicles from Tesla in 3 years, and the regulation to allow those vehicles onto roads without a human driver in under 10 years.
I am tempted to think Musk is optimistic but I do think we’ll see the first driver-less vehicles on the roads in under 10 years. Now that said these might be commercial vehicles following well defined, well marked routes, but they will certainly be out there.
Why? Because it’s always a good bet to follow the market. Today there is a huge amount of money pouring into autonomous driving innovation and there is a huge amount of financial incentive for early entrants with an effective solution.
Yes.. Have you never seen Star Trek or read Dick Tracy?
Even laymen predicted that.
I can’t agree more. Cell phones are cool, but what about flying cars?? and Mars Colonies?? human-like AI??
Yeah, 40 years is a lot in technology blah blah blah, but to me it’s not so much really. In fact, people with a PHD level engineering degree in the 70s, like Gordon Moore, not only predicted “the future” also defined a pretty accurate growing rate… so there’s no surprise at all in today’s technology!!!
So, I don’t think current technology is so “amazing” as stupid people say on TV. Tinder is Ok, but We expected Skynet. Really.
Dick Tracy debuted in the 1931 when valve (tube) radios were still a novelty. Dick had a two-way wrist radio 85 years ago.
Edited 2016-07-31 07:23 UTC
Elon Musk is half charlatan, half snake oil salesman and half subsidy-seeking parasite. He should never be taken seriously.
The first successful class action lawsuit for wrongful death will end the drivers-as-beta-testers approach to autonomous cars and a legislative crackdown.
Yep, and even if the technology does get better we’ll have an even harder road ahead of us to get it accepted. Damn Musk for his false marketing and the damage he will do.
Maybe they think it can be profitable to sell autonomous driving systems to car makers.
Anyway, Apple is sitting on such insane piles of cash, they can afford to burn some chasing future trends.
What I ind more interesting: the QNX founder? Does this means changes to the core of the OS? A different kernel for the car system?
I think it means that whatever system the put in place, it will be well hardened. none of this CANBUS hacking we hear about.
Love or hate apple, their stand on iOS encryption is good for the user and IMHO points towards the way of the future. There would be far too much egg on face for Apple to stomach to release something that was so hackable with potentially disasterous consequences.
Yes, they will need some real time behavior and the experience to write a _real_ µKernel. Not this Mach crap.
Vehicle manufacturers are totally ruthless cost cutters. They aren’t going to allow any supplier to make huge profit margins.
The (traditional) automotive industry is considered slow with respect to innovations. But this for a good reason.
In the past, the major companies did millions of kilometers test driving with _test_drivers_ not clients.
But now Google, Tesla and Co. try to use the IT development methods into car business: Test at the customer. This will not work and I hope they will see this quickly before more people die.
“The car industry has never been profitable.”
You could say the same about the computer industry – Except Apple.
CaptainN-,
The impression I have is that apple’s consumer electronics has been carrying it’s computer division for a while. It’s been a few years since they’ve given up on the server market. Even though they get a lot of praise, it seems they’re facing the same crunch with PCs as everyone else. The fact that they have not been investing their billions in the computer market tells me that even they don’t have much hope for it.
So that would be comparable to predicting in 1986 where computer technology would be in 2016.
Marvin Minsky – the father of AI – predicted in 1970 that computer would be able to outperform humans in any physical or mental activity by 1975. Forty-six years later computer still can’t demonstrate basic understanding and robots have very limited dexterity.
Edited 2016-07-31 07:16 UTC
I work on software tools related to autonomous cars, and I can say that 30-40 years is a very, very far guess. I would think the technology is at most 10 years away, and our most dire challenges are all cultural and legal/legislative.
FlyingJester,
What do you do? It’s a long shot, but any jobs near suffolk county NY?
I’m a tools engineer. I’m one a few developers who works on an internal application related to autonomous vehicles. I’m afraid I can’t get too much more specific about what I actually do than that.
No jobs in NY that I know of, the farthest east office of the company I’m with is in Chicago.
Edited 2016-07-30 01:58 UTC
Except it wasn’t an autonomous drive system. Assist systems aren’t designed to drive by themselves and I don’t know of any marketing that doesn’t point that out.
I’m not convinced apple wants to make a traditional car at all. They are not going to out-do BMW, Porsche, Toyota, etc.. They won’t even out-do Ford on a traditional car.
No one wants an apple product sitting outdoors getting dirty. No one is expecting it to work exactly the same for 20+ years. That’s just not their talent.
I could see them making some sort of fancy golf cart or even a smart-car type of companion product, perhaps a 3-wheeler with solar power. But I doubt we’ll see Apple-branded traditional cars on the road competing with the big names.
Apple only goes into the product space if they can truly distinguish themselves. Nice colors, shaped plastic, or model-year features won’t win them any market share here. Jon Ivy can’t/won’t make a car that looks much better than anyone else. He can’t push manufacturing standards in that field.
Also, Steve Jobs sped around in Porsche’s. Tim Cook looks like he drives a Mercury Sable.
Edited 2016-07-29 13:39 UTC
Where is my f’ing Hovercar?
https://m.soundcloud.com/fangbaby/hovercar