• nullspace@lemmy.world
    link
    fedilink
    arrow-up
    55
    ·
    7 days ago

    Three major updates before the coffee’s done brewing brings to mind a team scrambling to fix a fuck up, fucking up even more, then scrambling to fix that on top of the original fuck up.

    • matlag@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      8
      arrow-down
      1
      ·
      6 days ago

      “[Dev1]Oooooh sh*t! Prod is down!! AI, we need to fix this ASAP!
      -[CodeAssistant] Please wait, I’m working on a solution.
      -[Dev1] No, wait! I haven’t given you any details yet!
      -[CodeAssistant] Proposed solution: delete entire codebase and database to restart from scratch! Do you want to proceed?
      -[Dev2] Mike, do you want to grab a drink after work?
      -[Dev1] Yes, sure! … Wait! NOOOOOOOOOOO!”

    • dejected_warp_core@lemmy.world
      link
      fedilink
      arrow-up
      7
      ·
      6 days ago

      This all kind of implies that the devs can push to prod whenever they want, which isn’t how everyone else is set up. Some of us want to cordon off all this chaos, just a little, with release windows and scheduled downtime if needed. Which is yet more LinkedIn Lunatic material.

    • ricecake@sh.itjust.works
      link
      fedilink
      arrow-up
      37
      ·
      7 days ago

      Bigger question: who the fuck is bottlenecked by their typing speed?

      You can’t vibe code your way out of needing to know what to build, how it fits together with what you have, or the things that actually make the job tricky.

      • chuckleslord@lemmy.world
        link
        fedilink
        arrow-up
        22
        ·
        7 days ago

        You forget that I can now describe the code I want faster than I can write it. Frankly it’s irrelevant if the code at the end doesn’t do what I want or that I don’t know what it does. The speed, baby. Never before could I push this much code this fast. 200+ new bugs a day, four times as fast as I used to be pushing features. The code base is worse than it’s ever been and it’s only going to get faster from here. This is the future, baby - the vibe coder

    • captainlezbian@lemmy.world
      link
      fedilink
      arrow-up
      8
      ·
      7 days ago

      I type slower than I talk, but partly because I’m usually thinking through what I’m typing and not as I’m talking

        • smeenz@lemmy.nz
          link
          fedilink
          English
          arrow-up
          2
          ·
          6 days ago

          Let chatgpt do your thinking for you, and don’t forget to let it know your plane landed safely!

    • vaultdweller013@sh.itjust.works
      link
      fedilink
      arrow-up
      5
      ·
      7 days ago

      For me they are roughly on par with each other when accounting for stutter and autocorrect/ me being stupid and mixing up spelling styles. Regardless voice recognition software almost always shits the bed with my accent eventually since it gets thicker the longer I talk, it doesn’t like my fucked up mix of American brogue, old arkansaner, and standard southern Californian even got a bit of German Minnesota mixed in.

    • Bluewing@lemmy.world
      link
      fedilink
      arrow-up
      4
      ·
      7 days ago

      I type waaaaayyyyyy slower than I talk. And I talk waaaayyyy slower than I think. And I would definitely destroy any computer trying to keep up with my random and meaningless thoughts.

    • ZILtoid1991@lemmy.world
      link
      fedilink
      arrow-up
      2
      ·
      7 days ago

      I sometimes can type faster than I normally speak, but that depends on the keyboard (needs something without a membrane sensing apparatus (not rubber domes) for it to work), and I sometimes can type faster than I could think clearly, which can be detrimental to the quality of output. But hey, at a company, they could use it as a metric.

  • ☂️-@lemmy.ml
    link
    fedilink
    arrow-up
    20
    ·
    6 days ago

    if you are shipping 3 major updates in a single day, something probably went super wrong.

    • matlag@sh.itjust.works
      link
      fedilink
      arrow-up
      8
      ·
      6 days ago

      No, don’t worry: all the tests are done by AI, and they’re all ok with the output. It may create some tensions with the users in production with behaviors the users don’t expect or understand, but we no longer call those bugs, it’s all because human users are inefficient and should be replaced by AI assistants.

    • notabot@piefed.social
      link
      fedilink
      English
      arrow-up
      5
      ·
      6 days ago

      Management decided that the issue tracking system should be set up by the intern, using an LLM to slop out the configuration. The good news is they don’t receive any bug reports…ever.

  • f3nyx@lemmy.ml
    link
    fedilink
    arrow-up
    19
    ·
    6 days ago

    this is what happens when you let someone who gets stuck in vim make a company

  • SaharaMaleikuhm@feddit.org
    link
    fedilink
    arrow-up
    39
    ·
    7 days ago

    Wearing a surgical mask to protect others is tyrannical, but wearing this fucking muzzle to maximize profits for your corpo overlords is freedom. Get it right!

    • HaraldvonBlauzahn@feddit.org
      link
      fedilink
      arrow-up
      7
      ·
      6 days ago

      Why so pessimistic? 1000000x developer sounds much better! /s

      Fun fact: The average Linux kernel developer adds probably less than 1100 lines of code per person-year to the kernel. It’s estimated 40,000,000 lines of code with over 40,000 person-years of work as input.

  • kossa@feddit.org
    link
    fedilink
    arrow-up
    29
    ·
    7 days ago

    shipped three major updates

    Aside from what’s obviously wrong with that from a technical point of view…as a user I loathe this shit. Like, you think I love it, if your shitty software is nagging me every three hours to update?

    • Jyek@sh.itjust.works
      link
      fedilink
      arrow-up
      8
      ·
      6 days ago

      This is for prompting an LLM. The “senior dev” in this case is not coding anything and neither are typing anything. The idea is that speaking to the LLM is faster than typing a conversation with it.

      • lps2@lemmy.ml
        link
        fedilink
        arrow-up
        7
        ·
        6 days ago

        And apparently junior devs are clamoring to eavesdrop on how senior devs are prompting the LLM and that’s a problem. I can see the problem of being in an open office plan and having a ton of folks talking over another but the solution is just cubicles and not this face hugger microphone

        • luciferofastora@feddit.org
          link
          fedilink
          arrow-up
          3
          arrow-down
          1
          ·
          6 days ago

          My office’s open floors have a thorough setup of noise dampening elements. The floor is some fancy carpet, the ceiling lined with some kind of foam elements, each desk has a backing made of that stuff too, and it works fairly well to keep the noise level manageable even when five people around me are in calls (which, given they’re second level support, happens fairly often). Our headset have ANC and some fancy built in voice isolation system, so it’s not even a bother for people on the other end.

          Sure, it’s not cheap, but it’s neither as isolating as cubicles nor some dystopian bullshit like this. There are solutions if you’re willing to spend on them (and after making a show of leading by example, our boss somehow got very willing to spend that money…)

      • qarbone@lemmy.world
        link
        fedilink
        English
        arrow-up
        2
        ·
        6 days ago

        My comment is in reference to the first sentence in the post saying “the keyboard era is over.”

  • WorldsDumbestMan@lemmy.today
    link
    fedilink
    English
    arrow-up
    8
    ·
    6 days ago

    Needs to be more furry, for that extra touch.

    Maybe pig-snout shapped, and with pig-ear headphones.

    Ths guy types like an LLM too.