By on June 13, 2017

Tim Cook, Image: iphonedigital/flickr

Apple has been perpetually flip-flopping in terms of developing autonomous vehicles. In 2014, the company was rumored to have begun work on an autonomous electric car, codenamed “Project Titan,” with hundreds of employees devoted solely to its development.

Management issues and logistical problems impaired its progress, leading Apple to abandon the project. Since then, Bob Mansfield has fronted a renewed effort to focus on building an autonomous driving system rather than a complete car. At least, that was everyone’s best guess, as the company has been semi-secretive about its mission since day one.

That changed on Tuesday, when CEO Tim Cook confirmed that Apple does indeed have a self-driving development program. The chief executive even went so far as to call it “the mother of all A.I. projects.” That’s quite the claim to make, considering making the tech work on a car is half the battle and Apple has no practical experience building an autonomous vehicle. 

“We’re focusing on autonomous systems,” Cook said in an interview on Bloomberg Television earlier this month. “It’s a core technology that we view as very important.”

He detailed his view of the future for the automotive industry as a combination of self-driving cars, electric vehicles, and ride-sharing. Apple plans to focus on autonomy, though some of Cook’s talk seemed to hint that the company hadn’t abandoned the notion of building an EV entirely. “It’s a marvelous experience not to stop at the filling station or the gas station,” Cook said.

However, the CEO’s penchant for not having to stop at the gas station — or the filing station, if you prefer — doesn’t indicate the company has any genuine plans to build a physical car. Cook just didn’t outright say that option was no longer on the table.

“We’ll see where it takes us,” he said. “We’re not really saying from a product point of view what we will do.”

It’s definitely too early for Apple to commit to anything. Despite having begun running its systems on at least one Lexus SUV, it has only recently gained an autonomous vehicle test permit to use in California. The company lags noticeably behind major automakers like Ford and General Motors, but is working on catching up.

A self-driving simulation group, which includes virtual reality expert Doug Bowman, has designed simulators that use VR to test Apple’s self-driving software. The company also invested one billion dollars into Didi Chuxing, the Chinese Uber, in 2016. Of course, none of this adds up to Apple being the first to achieve the fabled Level 5 Autonomy. It just proves it’s still a player and doesn’t want to be left out of the party.

“Clearly, one purpose of autonomous systems is self-driving cars — there are others. And we sort of see it as the mother all A.I. projects,” Cook told Bloomberg. “It’s probably one of the most difficult AI projects, actually, to work on. So autonomy is something that is incredibly exciting for us.”

[Image: iphonedigital/Flickr (CC BY-SA 2.0)]

Get the latest TTAC e-Newsletter!

12 Comments on “Apple CEO Dubs Self-driving Car Program ‘the Mother of All A.I. Projects’ ​...”

  • avatar

    That makes sense in that the goal isn’t self driving cars.
    The goal is AI, with self driving cars just being the first application of the technology.

    • 0 avatar

      AI is already used in Siri and also… Amazon Echo and Microsoft Cortana. Also in search engines too. Very basic but it is actually a form of AI. That’s right, AI is already with us – it’s just too crude for us to view it as AI at this moment.

      But being able to drive means being able to see – that will definitely push AI forward. So autonomous cars are a good incentive to accelerate AI. Lots of money there – lots of Baby Boomers want to get around without having to drive in their old age… I just hope AI is done properly…

      • 0 avatar

        ” just hope AI is done properly…”

        Or else we end up with SkyNet. :)

        Also want to mention IBM’s Watson, famous for beating Kasparov at Chess and winning at Jeopardy. Google ‘s Deep Mind recently beat the world champion go player, a game more complex than chess. These are accomplishments of the abstract and the intangible. Driving brings us to the physical world, which is a whole lot harder.

        Harder still would be creativity, which can be a final frontier of sorts.

        • 0 avatar

          Yes, but chess has a rigid set of rules, and the pieces are constrained to move in certain ways. It’s easy for the computer to run simulations of future moves and to figure out what the best immediate decision is.

          In traffic? Who can predict what any vehicle around you might do?

          I see AI working when all the vehicles on the road are talking to each other. Now if you’re trying to ride your motorcycle out there amongst those self-driving cars . . . good luck!

    • 0 avatar

      “The goal is AI, with self driving cars just being the first application of the technology.”

      Which is kind of, like: “The goal is walking uphill. Everest is just the first application….”

      The sheer number of useful products operating in environments that are infinitely simpler and more constrained, than the full brunt of the worst of the world’s traffic, is so long that starting by focusing on the latter, is just silly. Build a lawnmower that’s safe enough not to run over Fido, while being able to cut grass but leave flowers alone, first. Then an autonomous forklift, sufficiently safe for use around people without specialized training and workers comp protection as fallback. Or heck, even simpler: An automated luggage sorter/organizer at airports, that doesn’t constantly require human override…..

      The sole and only reason to go for the moonshot while still living in comparative caves, is that what one is attempting to sell, is not a functional product. But rather enough hype about some future supposed product, to channel freshprint from the Fed and starry eyed grads from “good schools,” in ones own direction. The former, by way of virtually free stock option compensation, being a key enabler of the latter.

      The whole show is a textbook example of Hayekian malinvestment/Austroan Business Cycle Theory: Excessively loose monetary policy, giving the illusion that more resources is available to the economy, than is really the case. So people invest in excessively abstract production processes, with payoff too remote and uncertain for the current level of actual economic wealth. And, like always, it will end with most of the massive efforts being poured into these misguided ventures, having been largely wasted. At huge real world cost, as many of “the best and brightest” will have spent their most productive years building cardhouses, rather than something more useful and sustainable for our current level of economic development.

      • 0 avatar

        “The sole and only reason to go for the moonshot while still living in comparative caves, is that what one is attempting to sell, is not a functional product”

        I agree. Autonomous cars is a sales angle and far from ready. We need vehicle to infrastructure first, and limit self driving cars to those roads.

  • avatar

    Apple has been rudderless since 2011.

    • 0 avatar
      tod stiles

      So that’s what I’ve been doing wrong these past six years. I’m get’n rid of my rudder first thing tomorrow so i can be like Apple.

      • 0 avatar

        Yeah the rudderless Apple boat keeps crashing into Cash Island.

        Every time an Apple Car rumor pops up I still believe they are developing a system (hardware/software combination) that will be installed into a current vehicle. Think BMW with Apple iDrive. While manufacturing a car might be the end goal here, I really doubt that will be step 1. Remember iTunes came out on Motorola phone before Apple figured out how to make the iPhone based on what they learned with the iPod.

  • avatar
    Shortest Circuit

    What is his address? I want to send him about 3 books on AI safety research, mostly just to show how ineffective our current software QC methods are when dealing with AI. You know, the “see a crack, patch it up” approach, where there are more edge cases than business rules? So you protected the occupants from 100 very well defined bad outcomes? An AI will find method #101 very-very fast.

  • avatar

    Spot on. Our “current software QC methods” don’t even catch all the simple logic and UI errors in modern software and web apps. For AI testing, we’ll need psychologists and psychiatrists doing the testing.

  • avatar

    I believe I saw Opie at the filling station today.

Read all comments

Back to TopLeave a Reply

You must be logged in to post a comment.

Recent Comments

  • Carlson Fan: “I could see quite a few places where you’d might want an electric model.” I’ll bet my...
  • Mike-NB2: Agreed. If Toyota were to bring this to North America it might mean I have to rethink my anti-Toyota bias...
  • ToolGuy: [Oh, then there is the gasoline-powered auger that another ranger and I may have used to install lantern...
  • ToolGuy: “in parks with noise restrictions” Last time I went camping (state park, reserved spot,...
  • ToolGuy: @EBFlex, I get lost with all the ATV/UTV terminology, but a couple more examples hopefully closer to the...

New Car Research

Get a Free Dealer Quote

Who We Are

  • Adam Tonge
  • Bozi Tatarevic
  • Corey Lewis
  • Jo Borras
  • Mark Baruth
  • Ronnie Schreiber