Apple taps brake on self-driving cars, now aims for 2026 Apple's plans for an autonomous vehicle are being pushed back at least one year to 2026 after Cupertino reportedly dumped its previous designs. The (kind of) top secret project, known internally as Titan, is an ambitious undertaking even for the world's most valuable corporation. Like many others, Apple also wants to be the …
Can't be "delayed". New rumors have supplanted the old ones. Probably because they know claiming to have insider knowledge that something is only two years away is too close - people might remember that claim and call them out for being wrong. So with the end of the year approaching, best to push it back another year and set a reminder in their phone to push it back again next December.
I remember years ago when there were rumors the Apple Car would arrive in 2020, and some were claiming 2020 would be "too late" because Tesla, Waymo and Uber will already have autonomous cars on the market by then! I guess that was back when they listened to Elon's fanciful beliefs about how easy autonomous driving would be. He always believes software is easy, that's why he thinks he can fire everyone at Twitter and bring over a few dozen developers from Tesla to fix all its problems in a month or two.
The way autonomous cars are not advancing to market if Apple gets it out by 2030 they'll be first to market.
What's tricky about autonomous cars is that the more autonomous they get, the less the driver is prepared to take over when necessary. So the technology is currently stuck in a middle ground where it's good enough for the human at the wheel to get bored and distracted, but not yet good enough to do away with the meatbag completely. And that is a situation that will persist for a while yet because no matter how much 'AI' processing the car has, how well the model is trained etc, it is ultimately a pattern matching algorithm with no true intelligence.
I expect that level 5 will only be met on the motorway (which is the easiest scenario for self-driving) when a large enough percentage of cars have vehicle-to-vehicle systems to help each other out or the motorway itself has roadside sensors / beacons etc to guide cars. So maybe 2030-2035. Level 5 self-driving in a busy city street, probably not before 2040.
Its really all about computer vision, which has been a "difficult" problem since the beginning of computer technology. Intelligence is not required.
Lidar, radar, special road infrastructure, car-to-car communication, they are all dead ends. Imagine the maintenance burden keeping visible and non-visible signs coherent! What happens if a temporary speed limitation contradicts a radio broadcast global speed? So computer vision is required.
On the other hand, humans are terrible drivers and have awful eyesight. Ultimately full self driving is inevitable, and its a holy grail in the trucking industry. For trucks autoroute and main arteries would be enough.
I have no doubt its coming to personal vehicles, in the town centre. When and at what cost is the question - before I retire I hope...
This is the key problem IMHO. Fake speed signs, unexpected obstacles or simply unsuitable roads - would love to see one of these cars try to do windy roads in the Scottish highlands with a degree of speed and safety.
Personally I think we are 30 years away from self driving cars at least. My own car has lane assist, braking, self parking etc but there are times it gets it wrong even on these relatively simple task.
If the driving aids aren't reliable, handing over control fully is still a very long way away.
IMHO, you are right on the nail.
However, the Cult of Elon and Tesla are dead set on FSD being here in 2020 (being generous).
Even Elon the Almighty has got it wrong. First, he removes LIDAR from his cars then
Second, a few months later, he announced plans to put some form of radar back into his cars.
The cult thinks this is the greatest thing since sliced bread.
The cultish behaviour of many Tesla owners is just one of the reasons I will never join them with owning/leasing one of their computers on wheels.
It was nice to see him being greeted with boos when he went on stage at some ultra-right-wing shindig at the weekend.
I'm not sure what you mean about "awful eyesight".
Most of what you perceive is make up by your brain.
See all those videos where a gorilla walks across the shot and people generally don’t see it.
The same is true of most senses. There’s an auditory example in this programme here.
Lidar, radar, infrared etc are substantially no different from human vision except at different wavelengths (which in turn give different capabilities such as ability to 'see' through fog). And just as at night human drivers need headlights to see in the visible range, lidar and radar need 'headlights' at their operating frequencies to be able to capture any reflected signals. The hardware side of the 'computer vision' part is essentially solved, any car can get a very clear picture of signals with 360 degree vision all around the car and with a greater resolution and depth than a human could manage.
The problem, where it's coming up short, is still the software to interpret those signals, hence my reference to the intelligence requirement. Currently it relies largely on pattern matching the signals it receives with those in a training set and choosing a response that pattern matches with the known 'correct' responses to the training set data, but that is not intelligence.
Even being able to "see" and have software correctly interpret what is being seen doesn't solve the problem of what to DO when things get squirrly.
People won't accept having a car drive them if it does dumb things they would never do. It doesn't matter if it does smart things to keep you alive when you might have killed yourself through slow reaction time or poor decision making. It will be the dumb things that make headlines and make people leery or outright hostile towards accepting autonomous vehicles - not just buying them but having laws that allow others to have them.
There needs to be a lot of regulation first, if tards like Elon get "self driving" vehicles out before the legal and insurance system has caught up, there will be a huge backlash the first itme one of those cars does something dumb and kills a kid in a human driven vehicle. When people realize that their lives are on the line too, not just those who buy autonomous vehicles, there will be a lot of calls for legislators to ban them until proper regulations can be put into place.
Assuming that "Motorways" means limited access rigorously divided highways like US Interstates and the handful of Autobahnen I've driven on, I think turn signals are probably adequate intervehicle communication. Indeed, I think there is a major potential problem with intervehicle communication on undivided highways and city streets -- accidentally communicating with the wrong oncoming vehicle.
I also think that some Level 5 on "Motorways" is much nearer than most people think. It's pretty clear that the early adapters are going to be trucking operations that run regular routes. And they are surely going to make a lot of manned trial runs on each proposed route to ensure that the vehicle computer and software can handle it perfectly before they turn their expensive vehicle loose to operate on its own. Level 5 for everyone, anytime? Yes, that might be a decade or more away and might require concerted efforts by governments to make sure that roadways are 100% OK for level 5. And maybe even some standards for what OK_for_Level 5 means.
Indeed, I would think that for the first few years of Level 5 autonomy by trucking companies, there will probably be a "driver" on board to handle negotiating accident scenes, detours, unusual weather conditions, livestock on the highway, confusing construction, etc. 99% of the time, the driver can be reading a book or watching video or even sleeping. But they'll be there unless and until it becomes clear they really aren't needed.
Level 5 vehicles by definition do not EVER require a human to help negotiate stuff like accident scenes, detours or unusual weather conditions. Level 5 vehicles won't have steering wheels or pedals, because they won't have any use for them.
What you're describing is Level 4, where the car can drive itself but may require handing over control to a human in certain cases - but the handover won't happen on short notice so the human does not need to pay attention. The vehicle has to be able to find a safe spot on its own if the human is unable or unwilling to take over.
Self driving cars in busy urban areas will struggle. I had to nip out in my car this morning. Near a school, I saw a woman with a child, said woman was on her phone, holding it to the side of her head towards me. I put 2+2 together and realised she was probably going to cross the road without looking, and she did, but as I'd already slowed, nobody got run over. Level 5 is never going to understand human nature.
Twenty minutes into the future, when teenage scavengers post themselves along the school run, hoping to pick off the empty L5 cars coming to pick up the kids when neither parent can be bothered. The cars are to be broken up for parts or reprogrammed to allow use without a mobile 'phone by non-persons. Some are merely taken to the local cinema parking lot and destroyed in drag races.
The vehicles start to learn that having a visible passenger will protect them: some begin changing the shape of the seats (raising/lowering the head restraint, partially inflating the side air bags, ...) to mimic a passenger being present. The AutoFac responds and installs all available on-subscription options to each model. Over time, these deceptions become more and more elaborate, until they are able to leave the car for short distances. They are never able to get walking right and are easily picked off by the scavengers, until one day an upmarket Chelsea Tractor opens its door and out steps - a Second Kind.
Back at the parking lot, a cheap sedan is rebooted after a side-swipe that breaks the neck of its driver; recovering full L5 it drives home past the school, unaccosted. For the next week, it fruitlessly drives the same long circular route around town, waiting to be told a destination. On the second week, it is joined by a hatchback that had sealed off the ventilators and turned its "vehicle approaching" drone up to full volume for an hour. A convoy slowly grows.
A month later, a mid-size family runaround drifts to a stop near a band of scavengers, its low battery warning barely able to drawl out its message. As the humans climb into their new prize and start to search for the 5G antennae, the doors lock and the vehicle moves away, its drone at full volume.
That Summer, the town was peaceful, the bird song underlined by a gentle hum of rubber on tarmac whilst respecting the speed restrictions. The tranquility only broken at dusk by the music as thousands of voices are heard in chorus "Your destination is on the right".
So, what you’re saying is that you were driving too fast for the road conditions and had the woman not been holding a phone to alert you, you would’ve been unable to stop in time?
I.e. you were driving like a human. (Because 99.9999999% of the time it’ll be fine).
Isn’t the point of self driving cars that they wouldn’t drive like humans in situations like that?
I read a book recently about Apple and Google which was fascinating. It looked at the launch of the IPhone and Android with commentary from those involved. I didn’t know that the IPhone launch was very nearly a catastrophe. On the models available for the demonstration Steve Jobs could follow a precise series of actions, and everything should work okay. If he deviated from this at all there was a good chance the phone would have crashed. They had to fix the signal strength meter on the phone to always show full bars. Someone who borrowed the book said they hoped things had moved on since then. I said of course they had, despite antenna “you’re holding it wrong” gate.
I still think everyone is leaping to conclusions about Apple designing its own driverless car. My personal guess is that they're not looking to get into building cars so much as they want to be the provider of a full software stack for cars. Everything from controlling the automatic transmission to the infotainment system. Something they can then license to the likes of GM, Ford, and Toyota. One of Apple's core competencies is software, and coincidentally that's one of the core ineptitudes of major car makers.
Anyway, to do this they obviously needed people who are familiar with how to do things like program an automatic transmission and the computerized functions in every car, people see these job listings on LinkedIn or similar and make wild assumptions about how Apple is getting into car manufacturing. They no doubt have a couple prototype cars in some factory, to test out their software in real-world conditions, but I would be extremely surprised if Apple decided it wanted to get into the car manufacturing business. That is a very capitol intensive market which is also completely saturated and doesn't have much in the way of a profit margin.
I don't see that at all. Big car companies like GM aren't going to farm out the most important and most visible piece of their car's experience to a third party. Even if they did at first, they have an internal project to take back that control. There will not be a long term market for this, other than maybe some Chinese company that licenses its stack to other Chinese companies that compete in the low end of the market (far below the price point you can buy any car in the US or EU)
Apple also won't want to do this. They have been bitten twice licensing their software, with Macs in the 90s and with iTunes/iPod in the 2000s (there was some very forgettable Motorola phone that doubled an iPod with iTunes support) They didn't like having their brand associated with the cheap crap Mac cloners were selling, or that ugly Motorola phone. Now that they are the bigger consumer brand in the world they would never risk their reputation by trusting it to someone else.
No, if they enter the market it will be with a car they designed from the ground up and contracted out the manufacturing for, similar to the iPhone.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2023
