Even worse than that is they didn't make Siri, they bought it, then didn't develop it a whole ton further.
So who are they going to buy that makes self driving software?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple Autonomous Driving Training Program Confirms Self-Driving Software Platform
- Thread starter MacRumors
- Start date
- Sort by reaction score
Even worse than that is they didn't make Siri, they bought it, then didn't develop it a whole ton further.
So who are they going to buy that makes self driving software?
The problem is, "everyone else," especially people who hang out on internet tech forums, have no experience and know next to nothing about the subject.
What Apple is doing is largely speculation. Bits can be gleaned from job requisitions, interesting people hired from the automotive world, the increases in their R&D budget, public filings, and the fact that Bob Mansfield is heading up the project. To me, that speaks volumes.
Coming from the cellular wireless infrastructure world, to me, the fact that Apple, starting with nothing, could pull off developing the iPhone 10 years ago, with no experience in the cellular industry was an astonishing feat.
Partnering with Moto on the ROKR was a shrewd move by Jobs, and a key enabler for Apple. Not for the phone itself which was ho-hum and a flop (which Jobs likely knew far in advance), but getting Apple engineers exposed to a world where cellular communications information was next to impossible to come by at the time, where the main players then (Ericsson, Nokia, Siemens, Alcatel, Lucent, etc) were incredibly tight-lipped about the required technology and techniques (a lot kept out of the public domain by trade-secret, rather than patents which are public) to play in that arena.
Here's my understanding of UBI (universal basic income).
Increasing use of automation will displace more and more jobs. This will start with entry level, menial work like cleaners, assembly line workers and cashiers, then as AI improves, expect the fallout to spread to jobs like driving (with the advent of self-driving cars), even to jobs once thought untouchable, such as teaching and accounting.
Companies will benefit in the form of greater profits (lower costs, improved productivity), while the government will have to foot the costs to society (higher unemployment and all the accompanying ramifications). In short, the benefits are privatised, while the costs are socialised.
Lower supply of jobs meets rising population. This is a recipe for the collapse of society. The only feasible solution I can think of is that the government gives everyone a monthly allowance. This would be enough to cover their monthly expenses, and essentially free them up from the need to do "mundane, pointless work" which in reality has already been replaced by technology anyways. To look at it cynically, it's a handout to jobless people to prevent them from rioting on the streets and turning to crime, just packaged more nicely.
Freed from the need to work to make ends meet, this would free people up to focus on enriching their lives by doing things that truly matter, but which they may not have been able to do previously. For example, people could use their new found time to write a new song, or travel the world, or create something beneficial to society. Of course, they could continue to work as well.
The money would come from increased taxation on these corporations who are making all the money. If their profits are predicated on creating more problems for society, it stands to reason that they should help foot the bill of the problems caused by their actions as well.
I am of the opinion that if a task can be replaced by technology, it eventually will. A new world order is inevitable. The question is - who gets to write it.
Autonomous driving is completely open at this point. It's impossible to us to know what Apple's strategy is, but a strong and competitive autonomous driving system will be extremely valuable.
E.g., if the automaker spending billions on it now can get what they want from Apple and aren't having enough success creating it themselves, they will happily send those billions to Apple.
But my guess is that the plan is still for Apple branded autonomous cars. (I think they'll partner with an established automaker to create a car company, sorta like Saturn, with Apple having a lot of control over design, and marketing and the automaker running everything else, and a new dealership network).
But Apple's history is that it makes a product customers don't realize they need until Apple releases it. Most car makers outsource all or part of their their computer systems. They have to buy them from somewhere -- why not Apple? If Apple can do it better, make maintenance of systems easier and less expensive to maintain, more accurate, among other criteria sure I'm sure car makers would give a listen. And Apple has already dipped it's toe in car s/w with CarPlay.
And as for an actual Apple Car why would Apple want to get involved in a low margin, high capital intensive business with myriad of state and local laws on dealerships, when it can make as much $ programming car's innards alone. To me it looks like Apple will use Didi Chuxing as its vehicle for a proof of concept. And then sell to auto makers with that example.
That car will never make it to market or if it does it will not have a proper error log so you can see what happened to cause the wreck.
This is what I'm wondering, as well. The best ideas I can come up with are:
1) It's (at least partly) an exercise to develop better realtime computer-vision software, sensors and AI
2) They're going to hire the West Coast Customs crew, Xzibit, and start selling modified cars in bulk to select partners such as Didi.
The major automakers are investing billions, but we've seen their lack of direction and experience with software hurt them before in areas such as dashboard systems, which is why Apple stepped in with CarPlay and Google stepped in with Android Auto. Everything tells us that automakers absolutely suck at software. Historically, they've been happy to release insecure, awkward, and proprietary software which does nothing but draw the ire of their customers.
Apple are likewise already investing billions in computer-vision and AI. They need to get better if they want to match Google, and they know it. This project could align with their strategy in more than one way.
Since Apple is doing this, I'm confused why Lattner left to go work on such a similar project at Tesla.
Granted, Tesla HAD a working L2 program, but at this point Tesla is falling backwards with their own Tesla Vision. So both Apple and Tesla are starting from next to nothing. FWIW I have one of the new AP hardware Teslas. It's awful, and I understand why Tesla is getting sued over it. There's no indication that the info from AP1 is useful to AP2. And the head AP1 guy bolted.
I wonder if management at Apple is now so business-y, that Lattner preferred to do it for Musk. I further wonder if Apple will beat Tesla to L5. Tesla's financial risk if they can't get to L5 (capability, regardless of regulatory approval) is tremendous -- a buy back of every single vehicle.
Granted, Tesla HAD a working L2 program, but at this point Tesla is falling backwards with their own Tesla Vision. So both Apple and Tesla are starting from next to nothing. FWIW I have one of the new AP hardware Teslas. It's awful, and I understand why Tesla is getting sued over it. There's no indication that the info from AP1 is useful to AP2. And the head AP1 guy bolted.
I wonder if management at Apple is now so business-y, that Lattner preferred to do it for Musk. I further wonder if Apple will beat Tesla to L5. Tesla's financial risk if they can't get to L5 (capability, regardless of regulatory approval) is tremendous -- a buy back of every single vehicle.
There is a long way to go, I agree. But give me a break with the "I don't want to be killled by software" crazies. On the road, if you look at people driving, 10% or more of drivers are looking down at their phone or something else.
Yes, there will be errors, yes people will die, but the safety aspect WILL be the driving force behind this when the technology matures.
Not that anything you say here is wrong, but it does remind me about how it is perfectly okay to say that we really don't know what Apple is doing in automotive, and lot of what some claim to know is based mainly on poorly informed speculation. Apple is pretty darned good at keeping secrets. They never telegraph what they are doing until it's time for the big reveals or the product is so far into production that evidence turns up in the supply chain. I totally understand the sport of reading tea leaves, but after so many wrong guesses, I'd hope a little bit of healthy circumspection would come into it. Well, at least MR stopped referring to the Apple Car.
Random Apple user, circa 2020:
"Siri, drive me to Apple Motors, I need a new Mac Pro"
Siri:
"I'm sorry. The Mac Pro is still under development."
"Can I drive you to Starbucks instead?"
Apple user:
"Sure."
"Siri, drive me to Apple Motors, I need a new Mac Pro"
Siri:
"I'm sorry. The Mac Pro is still under development."
"Can I drive you to Starbucks instead?"
Apple user:
"Sure."
Additional data. I tried "I too, want to know how to purchase two flowers" with OK Google, and Hey Cortana.
Ok Google: "I too want to know how to protect you flowers":
First link: https://genius.com/Nicki-minaj-black-barbies-lyrics
Hey Cortana: "I too want to know how to purchase to flowers."
First link: http://adequateman.deadspin.com/a-complete-guide-to-buying-women-flowers-1724715160
The issue with all these voice systems is that they don't understand pauses and mispeaking. If stop mid sentence to think about what I want to say a human understands that I am not done. If I say the wrong name, and then quickly correct myself a person would understand that I didn't really mean the first name, I meant the second. To me, until I can screw up my command, and the assistant understands what I meant to say, these things are simply a novelty.
[doublepost=1492879752][/doublepost]
Want and need are very different. Pass a law that requires that in dash devices be updated for the life of the vehicle and suddenly no one will be backing SmartDeviceLink. They want control, but they need experience.
Last edited:
Most likely not. There's a lot of thought being put into these scenarios now, for level five autonomous vehicles liability will mostly go to the vehicle manufacturer or software developer. There is still work to be done on legislative and insurance industry but this could spell the end for automotive insurance altogether as a combination of vastly lower accident rates (human error is the number one cause of crashes) and large manufacturers choosing to self insure would remove the need.
edit: and true level fives won't have a "manual mode", even including steering wheel and pedals would just be a waste of space. Manual mode is reserved for current limited systems such as Tesla Autopilot which gives you plenty of warning before leaving a highway.
Did it ever occur to you that the rest of the human race ALSO has a stake in this?
The issue is not only "what option do *you* prefer?", it is also "what option is likely going to hurt or kill fewer third parties, whether in other cars, on bicycles, or on foot?"...
As for who Apple is making this software for, I can't believe the ignorance of the comments I'm reading here.
To me the future is utterly obvious. There are a whole lot of new car manufacturers in China with factories full of leading edge equipment. Right now they are learning their craft building for the domestic market.
They will never be allowed to expand much outside China under their own brands, for the same reason (only far more so) that Japanese car makers were forced to engage in "voluntary" export controls.
The solution is that these Chinese manufacturers will become contract manufacturers for Western companies who will sell the cars in the US under their own brands. Basically TSMC+Foxconn+Apple applied to cars. Apple will get there first because (as usual) Apple is looking at what matters six years from now, not what mattered six months ago. But they will probably be followed as soon as the model is clear.
Tesla are, I'm sorry to say, the RIM, the Moses, the Joseph Smith of the auto world. They showed the direction, but they won't get to the promised land. They have small pieces of the 21st century automobile paradigm (electric, autonomous) but they don't have the HUGE piece that matters --- a new production system.
I expect the mainline US manufacturers (GM, Ford) to be the Microsoft in this story. They'll screw around in desperation, try a few different (clearly stupid in retrospect) things, and finally land on a strategy that keeps them alive, but less relevant every year. South Korea and Japan will probably follow the Samsung path, using AndroidAutonomous or whatever Google calls what it sells them.
I don't know where the European and other such manufacturers (Tata in India, Proton in Malaysia, ...) fit in. Presumably they will try to compete by using Google's (or anyone else who's selling --- Ford? GM? BMW?)'s software, but can they compete against what's likely to be a fairly rapid transition that's changing the engine, the "autonomy" + other electronics, and the styling itself, of a car, all within the context of Apple that's probably selling these things at prices that will, at first, seem unbelievably low --- remember Chinese contract manufacturers? (Remember how floored people were by the price of the first iPad? And even though they complained, and continue to complain, about the price of iPhones today, no-one is able to provide that same level of hardware and performance at a cheaper price...)
It only took ten years from the iPhone till today, with RIM, Nokia, MS and a host of other important names at the time (Motorola, Sony) more or less totally changed by the experience, and with a whole new set of players (ARM, QC, MediaTek, Huawei, Foxconn, TSMC) in the driving seat. The car change-over will be the same order of magnitude. Slightly slower because people don't change cars as fast, and because the various players will try harder to use legal barriers and protectionism to hold onto their turf, but basically the same overall revolution.
Is it me or if autonomous driving seems easy to do in the end?
I mean, two years ago, I called it pure science-fiction, then I saw what Tesla was doing and thought these guys were 20 years ahead of the competition. And nowadays, it seems like every company is rolling out their own solution gradually, even if they never belonged in the car industry.
I mean, two years ago, I called it pure science-fiction, then I saw what Tesla was doing and thought these guys were 20 years ahead of the competition. And nowadays, it seems like every company is rolling out their own solution gradually, even if they never belonged in the car industry.
It's much better to be killed by a loved one than by an unfeeling machine. And it's better for the loved one, too, knowing that their distracted driving caused your death.
[doublepost=1492889483][/doublepost]
It isn't easy. 80% of it is easy enough. The last 5% (which most people focus on) will be very difficult. The final 1%, the Kobayashi Maru scenarios, will always be a problem.
Some human drivers have, through extreme skill or blind luck, survived a totally unpredictable accident. Drivers who have not survived an accident (even those that could have been easily avoided) aren't here to weigh in on whether a computer programmed for that situation would have led to a better outcome.
It has to do with poaching and knowledge transfer. Autonomous driving is a hot field that many companies want to be part of to be relevant in the future so they're poaching employees with large paychecks plus incentives from leaders in the industry like Tesla.
How on earth is this pretty exciting? Self-driving cars are about as exciting as CD auto-changer.
As for Apple? Yawn. Re-learn how you became a multi-billion valued company; then fire Tim Cook.
You keep on saying Apple is not reliable because some mistakes they've made.
Come on, is Google any better?! They've spent 20 years creating dozens of useless overhyped projects that horribly failed. But we (kinda) trust them about self-driving cars because this time, if there's a mistake, people won't be frustrated or angry: they're gonna die.
And customers' death is bad, bad pubblicity, so they'll try to avoid it.
Come on, is Google any better?! They've spent 20 years creating dozens of useless overhyped projects that horribly failed. But we (kinda) trust them about self-driving cars because this time, if there's a mistake, people won't be frustrated or angry: they're gonna die.
And customers' death is bad, bad pubblicity, so they'll try to avoid it.
Hmm, nothing I've ever seen by Logitech was what I'd regard as sturdy enough for real use. In a critical situation, steering wheels and brake pedals can be subjected to a lot of force, to the point that even real hardware may take a beating. It'd be kind of rough if the driver had to take control, only to have the steering wheel snap off in their hand. 
How. Many. Times?
Apple ARE NOT building a car, get that through your heads once and for all.
Apple ARE NOT building a car, get that through your heads once and for all.
You book on Kayak to take a Boeing operated by JetBlue to SFO.
You open Uber to get a Toyota from SFO to the tenderloin.
At the end of the day, you took Delta and grabbed an Uber. The manufacturers were irrelevant to your experience.
Following that model — the non-owner model — there's a ton of room for improvement from a system integration and passenger-centric viewpoint. Everything presently is geared towards drivers or owners & operators. Your entire life is on your phone. No reason you shouldn't be able to plug it in and have a customized passenger UX.
Cars will become Dell and Lenovo which just run the programs and OS of the target user. They're becoming commodities.