• 0 Posts
  • 5 Comments
Joined 4 days ago
cake
Cake day: November 18th, 2026

help-circle
  • The nature of work has changed forever. There are code projects that are being banged out in weeks that would previously take years or could never be done.

    Yesterday, I just posted in another comment a few example of actual useful projects that were just created.

    Here are some examples of tangible results produced by Claude. Prior to LLMs these projects were impossible or would’ve taken years and years for some hobbyists to even begin making them. The latter two had failed attempts by people to do it without the aid of LLMs. Suddenly with Claude, this code got knocked out in a matter of weeks. These are working projects that people are actually using. Not the slop code that average people like to laugh at and ridicule.
    https://aioscilloscope.com/
    https://github.com/bve/OpenOSC
    https://github.com/kralonur/kws-x1-standalone-rs

    I’m not a programmer by profession, but find it interesting the mantra from professional programmers carried an air of elitism with the way so many preached perpetual autodidactism. It was fun to be smugly superior others. Now the bar got pushed to levels beyond the foremost experts alive (e.g. Riemann hypothesis). Now everybody gets to be the noob programmer. There’s tireless machine than runs circles everyone. You’ll have to constantly brush up your skills. Not as fun being on the other side of the table having someone else tell you that.

    The professional world has split in two. There are those who are forging ahead with the new tool and those who are resistant to the new world.

    Within the niche of gadgets made by small companies out of China, people have been making custom firmware so easily. Of course it still requires professional expertise of electronics and microcontrollers. Embedded engineers basically. This type of firmware was rare before because of the amount work it took. Now custom firmware is popping everywhere. And the engineers are able to go from nothing to 99% polished products in a short amount of time.

    I don’t think people are grasping what this means. This alone could spur another revolution in making gadgets.




  • Yep. Older generations of my family did piecework back in the day too. Initially I hadn’t thought about that. It came to me a few months into doing it. We’ve circled back to where we were a few generations ago.

    Piecework is about fixed rate, but this is essentially what it is because we’re measured by the rate of submissions. Some projects are hourly pay. Others paying per unit of production.

    Some companies are quite bad. They won’t pay for submissions that get a low grade. If it takes too long then the submission is paid lower or not at all.


  • I just got banned from one of these platforms. It was hell working that job. I really need money though.

    Everyone is constantly paranoid about getting banned. They require invasive logging to be installed. Everything is monitored. It’s micromanagment on steroids.

    One moment you’re looking at a picture of a thing, intensely scrutinizing it, hoping you’re working fast enough too. Suddenly the page auto redirects to a ban message. All your access is revoked. I spent a year fearing that every single day and it finally hit me. People wake up every morning to check and see if they’re not removed.

    My mind turns to mush grinding out data labelling/validation all day long. The pressure to meet rate and not mess up is intense. Your work is graded and you receive a score on each submission. It’s a lot of pressures.

    They are rolling out full time positions. So expect to see that in the coming year. I personally could not do 40 hours of this a week. This a consensus among workers too. There was chatter about roughly 4 to 6 hours a day of this kind of work at the most. Some did say they log the full amount of time we were alotted per week. That’s 10 or 12 hour days. Could you scrutinize photos for 12 hours a day? You have do it on an average of let’s say 2 or 3 minutes per photo. It usually require quick write up about it. It has to be perfect so as to not induce errors in the AI. You will receive a grade on each submission which is done by another worker like you who is under the same pressures.

    I’ve seen some of work people submit. Quite frankly the impoverished and less educated regions submit junk data. There’s no other way to say it. There’s more errors. The clarity of writing is worse too. Language models are pedantic about phrasing so fluency matters.

    They appear to be favoring workers from western regions lately. That’s only my gut feeling though. The pay can be quite high if you’re a professional in just about any field. Not only STEM. I personally did not see them pull tactics to reduce pay. Mine actually went up. I have heard the stories though. They pay third world countries a fraction of what they pay western countries for the same work. Then again, the quality kind of reflects that.

    Some who have done the job of reviewing work have expressed negative feelings about it. The job is to review other’s ability to label data. But they’re determining someone’s livelihood. Other’s don’t care. Some relish in the power they get from it.

    ETA: I forgot to mention. My computer would randomly lag while working. I assume it was whatever invasive surveillance their software was doing.