Comment on AI ‘Workslop’ Is Killing Productivity and Making Workers Miserable
quetzaldilla@lemmy.world 2 weeks agoI may be a tax professional today, but I’ve been passionate about computers since before computers had GUI.
Your computer most likely uses plugins that I wrote back when I was obsessed with programming, especially if you use Linux systems.
Basically, I was one of those who wrote the documentation you depend on when you are troubleshooting your AI code.
To do your vibe coding, you depend on tools that people like me who design and maintain them. And once we go, these tools will break and you will be helpless if others do not step up with the same level of passion and intellectual curiosity that is critical to problem solving processes.
**AI cannot simply build new tools because it lacks the intellectual curiosity that drives humans to understand their environment, identify problems, and work collaboratively with others. **
**AI is not intelligent. It is only capable of looking backwards, not forward. It is a performative regurgitation of information gathered and synthesized by others, and it does that very poorly. **
All LLMs are hallucinating and going psychotic even when there’s billions of dollars being thrown at the problem, because it is a problem that cannot be fixed because of the very nature of how LLMs are put together.
My advice to you is to reduce your AI usage and practice making things from scratch.
Not just coding, but other processes too like baking, cooking, repairing objects, writing, painting, etc.
Thinking and learning are literally the best parts of being human, and all AI will ever teach you is how to avoid thinking and it will steal all the opportunities for you to genuinely learn something new.
Because even after a twenty year hiatus, I can probably do whatever coding you are doing-- but I can do it without relying on AI.
ALoafOfBread@lemmy.ml 2 weeks ago
I am also learning to code and have been for years casually. In fact, I can program basic stuff in a couple languages.
I do not believe I am harming a single person by making scripts for myself and small team by using AI that allows me to produce better code while still being able to do my day job (which is not programming).
In fact, I learn a lot by using AI and having to do research. I learn more than I would on my own because being able to use LLMs to assist me has allowed me to take on projects that wouldn’t make sense for me to attempt if I had to spend weeks to get to the same outcome.
I am well aware of how LLMs work. I have read numerous articles, white papers, talked to ML engineers, etc. I know their limitations pretty well. I think they are much less capable than almost all laypeople I have spoken to about it.
All that is to say: LLMs have some legitimate use cases. They are wasteful, inaccurate, etc etc. But they have use cases. People are hyperbolic in both directions on them.
quetzaldilla@lemmy.world 2 weeks ago
When experts across multiple disciples, all across the world, are sounding the alarms about AI that is not being hyperbolic.
You are basically learning from something incapable of teaching you the fundamentals, and who will never challenge you and help your refine your own critical thinking and research skills like a mentor or a professor will.
When you use AI, all you are really doing is helping your employer make a profit lay the groundwork to justify firing you and your colleagues down the road.
You know what I made with my programming and coding skills?
Patches and updates so you could have more fun playing with your computer, making digital art apps more streamlined and easy to use, adding accessibility features for the disabled.
AI is just there to make you more “efficient” so you can make someone else a profit, all while the billionaire owners continue to rape the planet and everything in it.