Ask HN: Anyone else feels that their job quality has declined severely recently?

I work at one of the bigger companies and Recently I have become completely demotivated by AI advancements, manager behaviour and exec apathy. We are not in a terrible economy yet, but workers are already suffering.

I have kept up with AI advancements, and used cluade code since Day1, even stayed with it when they added limits, I never advocated my company to use it, I always asked the reimbursement from my learning budget, and when asked I always say use opencode with openrouter. And we also have own CLI now. But its impossible to keep up with this whole thing and even if when we do, whats the point? I cant seem to convince anyone that we are useful, we are working harder and yet getting no rewards. Checking our internal promotion stats, the #s are down by margin that makes me cry, and most promotions are in non-SDE fields too. Its seems we are being phased out and execs have a plan. There is something weird going on.

I dont particularly care about _this_ job, but I care about our field. Everywhere managers are asking us to work more and execs are still not happy with anything, They dont seem to understand AI is the reason users dont use our products. They are actively hating on our products but no one seems to care.

5 points | by falloutx 15 hours ago

4 comments

  • Nextgrid 14 hours ago
    > cant seem to convince anyone that we are useful

    I wonder if the issue there isn’t AI but that a lot of jobs (including in tech) are bullshit and it turns out it doesn’t matter how well it’s being done.

    AI absolutely fails where quality is a requirement… but if it’s not then AI appears to be a good stand-in for a human. This was the case pre-AI too, plenty of mediocre developers were able to coast because the outcome of their work didn’t actually matter.

    > most promotions are in non-SDE fields too

    They’re just better at playing politics and maybe you should start too.

    [-]
    • falloutx 11 hours ago
      Thats definitely possible that a lot of dev jobs were just surface level and I would argue that if 90% of the feature dev is halted tmrw, you wouldn't even notice it for atleast couple years. This is why IBM and Oracle can still coast without having any new ideas. Lets even imagine 50% of the devs were never good enough, but that would still be less fluff any other field except labor.

      It always feels like we get held accountable for every little lapse, but higher up PMs, execs never get any consequences for their actions. They could run the company to the ground and still get promoted. If the top is incompetent, you cant really expect the bottom to be carrying all the weight.

  • rankiwiki 13 hours ago
    Feels less like engineers became useless and more like incentives broke. AI raised expectations faster than orgs updated rewards, so extra effort just gets normalized instead of recognized.
  • ludicrousdispla 14 hours ago
    From what you describe it sounds like software development is starting to get categorized as 'labor', and that a lot of companies will be having system problems over the next few years.
  • austin-cheney 14 hours ago
    I was a web developer from about 2006-2023 and all I remember was decline and insecurity. For most of that the backend was Java and the Java people were deathly afraid of JavaScript, not always but certainly more than 90%. The JavaScript people were afraid of everything. Insecurity was everywhere.

    I have since switched to enterprise API management and it has been great. I blame two things for the greatness: 1. Everybody has to obtain a certification, 2. It’s more operationally focused.

    Everything related to web development always felt like a race to the bottom. Almost nobody seemed competent to do the work and the goal was always delivery in the most minimal capability imaginable.

    There was a bright spot though. When I was the A/B test engineer for this major dot com life was great. I just built out experiments that defaced the production web site. There were many times the super hacky experiment code was less defective and executed faster than the real code that eventually rolled out to production.