Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
AI Apple

Apple Enlists Veteran Software Executive To Help Fix AI and Siri (yahoo.com) 29

An anonymous reader quotes a report from Bloomberg: Apple executive Kim Vorrath, a company veteran known for fixing troubled products and bringing major projects to market, has a new job: whipping artificial intelligence and Siri into shape. Vorrath, a vice president in charge of program management, was moved to Apple's artificial intelligence and machine learning division this week, according to people with knowledge of the matter. She'll be a top deputy to AI chief John Giannandrea, said the people, who asked not to be identified because the change hasn't been announced publicly. The move helps bolster a team that's racing to make Apple a leader in AI -- an area where it's fallen behind technology peers. [...]

Vorrath, who has spent 36 years at Apple, is known for managing the development of tough software projects. She's also put procedures in place that can catch and fix bugs. Vorrath joins the new team from Apple's hardware engineering division, where she helped launch the Vision Pro headset. Over the years, Vorrath has had a hand in several of Apple's biggest endeavors. In the mid-2000s, she was chosen to lead project management for the original iPhone software group and get the iconic device ready for consumers. Until 2019, she oversaw project management for the iPhone, iPad and Mac operating systems, before taking on the Vision Pro software. Haley Allen will replace Vorrath overseeing program management for visionOS, the headset's operating system, according to the people.

Prior to joining Giannandrea's organization, Vorrath had spent several weeks advising Kelsey Peterson, the group's previous head of program management. Peterson will now report to Vorrath -- as will two other AI executives, Cindy Lin and Marc Schonbrun. Giannandrea, who joined Apple from Google in 2018, disclosed the changes in a memo sent to staffers. The move signals that AI is now more important than the Vision Pro, which launched in February 2024, and is seen as the biggest challenge within the company, according to a longtime Apple executive who asked not to be identified. Vorrath has a knack for organizing engineering groups and creating an effective workflow with new processes, the executive said. It has been clear for some time now that Giannandrea needs additional help managing an AI group with growing prominence, according to the executive. Vorrath is poised to bring Apple's product development culture to the AI work, the person said.

Apple Enlists Veteran Software Executive To Help Fix AI and Siri

Comments Filter:
  • by Moryath ( 553296 ) on Friday January 24, 2025 @06:17PM (#65116541)

    Apple: "Hey, here's our new Siri AI. Ask it anything!"

    Customers tired of every company shoving "AI" bullshit and snake oil in their faces: "Ask it to fuck the hell off."

    • by martin-boundary ( 547041 ) on Friday January 24, 2025 @06:55PM (#65116665)
      The whole pattern is predictable, and has been for years.

      When you abstract away the starry-eyed wankery about what the technology is capable of doing and treat it as a black box with inputs and outputs, the conclusion is crystal clear.

      There's a black box that is trying to do complex things with human natural language input. It produces output with long range internal dependencies, and there is no useful way to measure and quantify the quality, other than using more human natural language specifications.

      Human natural languages are exceedingly bad at communication, they are imprecise and highly context specific. Even highly educated people attempting to communicate with other highly educated people fail at the task more often than not, as anyone who's tried to direct underlings and interns and talk with customers already knows.

      There's a reason highly formalized DSLs exist for programming and legal contracts and bureaucracy ad art etc. They are attempts at replacing poor natural language communication with something marginally better, so as to reduce the impedance mismatch between people. Even so, those attempts require experts who spend years studying and refining their mastery of these specialized languages, to the point where an ordinary person equipped only with natural language is unable to join the conversation any more.

      And here we have AI startups suggesting a better world where an ordinary person can use human natural language to direct a black box to do exactly what they mean. And if the black box fails the input communication, then more natural language communication on the output side will somehow correctly bring about the original meaning.

      • by Moryath ( 553296 )

        I am reminded of a constant truism of computing:

        Computers are great at doing what they were programmed and told to do, over and over again. As logical machinery, that's their purpose.

        PEOPLE are absolute shit at understanding how the computer is programmed, and at understanding what the fuck they actually told the computer to do.

        AI will continue this pattern, because people are the problem.

        "On two occasions I have been asked [by members of Parliament], 'Pray, Mr. Babbage, if you put into the machine wro

        • by Jeremi ( 14640 )

          PEOPLE are absolute shit at understanding how the computer is programmed, and at understanding what the fuck they actually told the computer to do.

          The above is true.

          AI will continue this pattern, because people are the problem.

          The above is maybe not entirely true -- on more than one occasion, I've gotten sloppy and asked ChatGPT a poorly-worded, ambiguous question that was really only a vague approximation of the question I should have asked (i.e. the one whose answer was what I actually wanted to know) -- and ChatGPT responded with the answer I actually wanted, even though I hadn't even asked the right question.

          Of course that trick only works if there were a lot of other people asking the same question badly in

          • Resoning models like o1, o3, DeepSeek R1 and Gemini 2.0 Flash Thinking Experimental 01-21, some of them just 3 days old, are using exploratory thinking before giving an answer. They don't simply blurt out an answer, but try first to understand what you want, and plan their response. So a model could handle situations that are not well covered in the training set.
      • LLMs learn at the same time all human languages, all programming languages and known DSLs. Besides that they train with images. They are not pure language models.
    • Her name sounds like a dungeon's boss monster.

      So, she has that going for her as well.

  • I still cannot navigate the hot mess that came with iOS 18
  • The problem with "AI" is not engineering or manufacturing, it is scientific. Engineering and manufacturing, you can fix. Scientific issues require decades and really smart people and can often _not_ be fixed or need a completely different approach.

  • Isn't it a general vibe with "AI" these days that it's all regurgitated content and utter low quality crap? It appears that Apple is the first company who's publicly acknowledged with this hire that gen AI is broken and can't fix itself. You need some with AI - Actual Intelligence - to back it up.

  • Nobody at Apple is irreplaceable. When any employee leaves, they are forgotten about within a week. A lot of people like to pretend that the people matter, but Apple is just a cash machine.

    • by Jeremi ( 14640 )

      Nobody at Apple is irreplaceable. When any employee leaves, they are forgotten about within a week. A lot of people like to pretend that the people matter, but Apple is just a cash machine

      No doubt, but I don't think there are any large corporations where the above isn't true. It's not their fault, exactly, it's just that any organization with significantly more than 150 people [cracked.com] can't fit into the human intellect, so it inevitably gets formalized and administered algorithmically instead.

    • Nobody at Apple is irreplaceable. When any employee leaves, they are forgotten about within a week. A lot of people like to pretend that the people matter, but Apple is just a cash machine.

      Part of the goal of growing a company, is to ensure no one person is irreplaceable.

      That goal, tends to become easier and easier when there isn’t just one of you, but an entire department of you.

  • by presearch ( 214913 ) on Friday January 24, 2025 @09:30PM (#65116959)

    He said
    "You've got to start with the customer experience and work backwards to the technology.
    You can't start with the technology and try to figure out where you're going to try and sell it."

    Apple may have done this better than some, but their implementation still smells of reactive desperation.

    • by tlhIngan ( 30335 )

      He said
      "You've got to start with the customer experience and work backwards to the technology.
      You can't start with the technology and try to figure out where you're going to try and sell it."

      Apple may have done this better than some, but their implementation still smells of reactive desperation.

      Nope, he did. Remember Siri? It was "pretty cool" in the demo, but it pretty much flopped out of the gate. And it happened on Jobs' watch.

      Just putting it out there, given Siri is also the current topic of discussion.

      • by Jeremi ( 14640 )

        Siri works fine -- as a voice-activated user interface, it's simple enough to reliably do the few tasks that it's good at, and it doesn't try to do the things it's not good at. It long ago reached the stage where "it just works", i.e. you can use it and get useful results, without having to think about how to manage the technology involved. My hope is that in their rush to add "AI" to it, Apple doesn't take away its reliability or day-to-day utility.

    • He said "You've got to start with the customer experience and work backwards to the technology. You can't start with the technology and try to figure out where you're going to try and sell it."

      Apple may have done this better than some, but their implementation still smells of reactive desperation.

      Apple reeks of desperation because sales have driven by a fashion statement justification for years. Especially to justify the obscene price. That fashion, isn’t making a statement like it was before.

      And that ignorance, is why artists starve. Jobs got lucky with his artistic vision. Statistically most don’t.

  • When I have to communicate with an "AI" implementation, I often have an impression of dealing with a mentally ill person. All answers seems to be correct and proper, until it says something so bizarre that it gets absolutely clear - this thing is abysmally stupid.
    • All answers seems to be correct and proper, until it says something so bizarre that it gets absolutely clear - this thing is abysmally stupid.

      Why do you believe your description of AI is somehow hard to accept in society as intelligence?

      We literally elect leaders with that same “feature”. We even gave them a title. Politician.

  • Just what the F!@K is it suppose to accomplish ???
    • Just what the F!@K is it suppose to accomplish ???

      To Make Greed Enslave Again.

      If society doesn’t like that, then it better get ready to put Greeds head on a fucking pike. Because that’s what it’s gonna take.

    • Who wants/needs AI?
      The Company.

      What is it supposed to accomplish?
      To create a dependency on it.

      or as I prefer to say "turn you into a drooling idiot".

      Not being fascetious here, those are the right answers to your questions.
      You're welcome :-)
  • by Fly Swatter ( 30498 ) on Saturday January 25, 2025 @12:24PM (#65117829) Homepage
    Any AI system that must always give an answer instead of admitting it doesn't know is a broken system.

Software production is assumed to be a line function, but it is run like a staff function. -- Paul Licker

Working...