• REDACTED@infosec.pub
    link
    fedilink
    English
    arrow-up
    14
    ·
    16 hours ago

    Having to do the most amount of bug fixes for the app (that does not run the AI itself) is not the flex you think it is

  • bentcheesee@lemmy.zip
    link
    fedilink
    arrow-up
    44
    ·
    1 day ago

    Incredibly misleading and/or stupid graphs are so funny to me. Because you ship out the most updates, doesn’t mean it’s the best, it means youre fixing and/or generating more bugs and issues.

    Yeah, I updated my minecraft mod 20 times in a week, it doesn’t mean it’s a stellar mod, it’s less than mediocre at best. It was primarily fixing bugs and a crash. Meanwhile the Create mod updates about once every three weeks or so on average, but that’s because they properly playtest and bugfix and patch and do all that before they send out an actual update.

  • percent@infosec.pub
    link
    fedilink
    arrow-up
    5
    ·
    17 hours ago

    To me, that just looks like “move fast and break things” taken to an extreme level. Quantity ≠ quality. Honestly, I think it kinda makes the other companies look better, in terms of software quality.

    (I’m going solely by the chart though; I have not installed any of these apps)

  • bitjunkie@lemmy.world
    link
    fedilink
    arrow-up
    70
    ·
    1 day ago

    “My codebase is way better because it has 300x as many lines of code” - that fucking moron, probably

  • GissaMittJobb@lemmy.ml
    link
    fedilink
    arrow-up
    83
    arrow-down
    1
    ·
    1 day ago

    On the contrary, the rate of mobile app updates being high is more of a red flag of an app development team not having the situation under control, being forced to panic-ship fixes.

    • cub Gucci@lemmy.today
      link
      fedilink
      arrow-up
      4
      ·
      1 day ago

      Why? I genuinely think that daily delivery in my field (b2b specialized software) would be a very good practice. Why in mobile apps it’s not the truth?

      • GissaMittJobb@lemmy.ml
        link
        fedilink
        arrow-up
        7
        ·
        24 hours ago

        It’s a bit different with mass market mobile applications because of the supply chain constraints - most notably the Apple reviewing process. Your next app release may for whatever reason they feel like unexpectedly take an additional week, so do ensure that your QA is in order before releasing.

        Another significant factor is the lack of control you have over the software once released - any bugs you ship may potentially be out there for a long, long time.

        Web applications don’t have these constraints and can as such be deployed an infinite amount of times per day. The same goes for backend services, deploy to your hearts content.

        This basically means that most larger mobile applications have adopted approximately weekly release cadences, and that we’ve had to get very good at using feature flagging to control our software in the wild, and avoid large impact of shipped bugs.

  • yarr@feddit.nl
    link
    fedilink
    English
    arrow-up
    31
    ·
    1 day ago

    I think it’s because grok ships the most bugs, so they have to ship the most patches.

  • Rose@slrpnk.net
    link
    fedilink
    arrow-up
    108
    ·
    2 days ago

    Oh wow, Elon figured out how we’ll finally get AGI. The key thing is to publish an automatic mobile client update every single hour of the day! That was the secret productivity metric that every single other company was missing. Thanks, big brain business boy!

    • jonne@infosec.pub
      link
      fedilink
      arrow-up
      125
      ·
      2 days ago

      Not to mention, the app is just the thing that calls the API to the server that runs the actual models, it’s not a reflection on how quickly you’re improving those models. In fact, there should be little reason to push a new app update once you’ve built it.

      • frog@feddit.uk
        link
        fedilink
        arrow-up
        30
        ·
        2 days ago

        The reason is so Elon Musk can make his stupid chart implying that more production releases means better in someway.

        • jonne@infosec.pub
          link
          fedilink
          arrow-up
          31
          ·
          edit-2
          2 days ago

          He’s really done a great job at dispelling this image he had of being some kind of genius in the last couple of years.

          • anomnom@sh.itjust.works
            link
            fedilink
            English
            arrow-up
            18
            ·
            2 days ago

            He’s been doing that the whole time, he just finally got to an area more people online are experts in.

            • jonne@infosec.pub
              link
              fedilink
              arrow-up
              12
              ·
              2 days ago

              Yeah, I personally noticed when there was this whole Thai cave saga, but I’m sure there were signs for decades before that.

  • wulrus@lemmy.world
    link
    fedilink
    arrow-up
    4
    ·
    1 day ago

    That statement would make a lot more sense with a benchmark graph based on standardised tasks.