I tried coding with AI, I became lazy and stupid

I tried coding with AI, I became lazy and stupid# Around April 2025, my boss at $dayjob insisted we try AI tools for coding. It wasn't toxic pressure or anything like "20% of your code needs to be AI", just a concern from him that we could miss on something. I understand why he asked that and I don't blame him. We are in difficult economic period even for software, and we have salaries to pay. If AI can increase productivity or our margins, it should be at least put on the table of negotiations. I am not happy about this coming, but I get it. My personal stance of AI I have personal reasons to dislike LLMs. My partner lost their writing job due to chatGPT convincing their manager writers were now useless. A lot of artist friends struggle because of LLMs. We recently had an intern who lost her translator role due to LLMs. And even outside my personal experience, LLMs are based on stolen content, don't respect consent, waste huge amount of electricity and water, and are overall a new weapon for the capitalists in the class warfare. The other reason why I folded comes from a toxic relationship I built with my job when I became a developer. I detailed in a previous blog post how choosing this career came with very high stakes which triggered a shift in my brain that hasn't left me since: When I started web development seven years ago, I was in survival mode after years of low paying wages and unemployment. It had to work, and for it to work, I had to always learn more, read and listen about web development all the time, monitor the field, socialize as much as possible with my peers. This way, I would not get disposable and lose my job. I would build a network. I would be safe. — I, In a blog post from 2022 10 years and 3 burnouts later, one can tell this mindset, even if it worked out for a while, wasn't sane or desirable. I had managed to put aside this fear of being disposable, but LLMs triggered it back big time. What if AI vendors were right? What if a future company I apply to requires you to use it? Am I going to lose my job? I'm almost 40, what will I do? So I tried AI. First at my day job, because I wanted answers. But outside fixing TypeScript types errors, generate inaccessible template code, or review my code for errors, I couldn't find a life changing use out of it that all AI influencers talk about. I asked my colleagues about their own experiments, and lots of them came to the same conclusion: it doesn't seem to help me help our clients achieve their goals. When July came I was starting the image processing part of my new CMS that powers this website. Still stressed I couldn't get a real shot at coding with an LLM, and very tired by different personal events that fogged my brain, I decided it was the right task to try it seriously and get answers. After setting up everything in VS Code, opening the AI panel, giving access to the codebase and detailing my needs in a prompt, the LLM produced around 200 lines of code. Mostly functions using dependencies to convert, resize, process images. It wasn't perfect but after a few changes, the task was done and it had taken around 30 minutes, far less than if I had made it by hand. I was impressed. It really felt like I had superpowers! But then I had the idea to audit the code the LLM just produced, like I did at my $dayjob for a Vue application. Feeling that uploading files could be a source of security issues, I asked the same LLM to focus on this specific topic. It found several dangers: directory traversal attacks, file size limits, system file overwrite, etc. I had no idea the initial code was this unsafe. I had reviewed the code, but without enough experience in backend development, how could I identify issues I didn't know existed? And why, if it knew about all those dangers, did the LLM produced unsafe code in the first place? When I tried to fix the security issues, I quickly realized how this whole thing was a trap. Since I didn't wrote it, I didn't have a good bird's eye view of the code and what it did. I couldn't make changes quickly, which started to frustrated me. The easiest route was asking the LLM to do the fixes for me, so I did. More code was changed and added. It worked, but again I could not tell if it was good or not. That's when I stopped the experiment. I was shocked by how easily I had slipped into this slacker way of programming. The LLM had given me shitty code, made me ignorant about my own code base, and too lazy to try to fix it myself. And at the same time, the whole experience felt smooth, frictionless, empowering. On the moment I felt smarter, more productive, in control. But it was all an illusion. I knew about this as we had studies showing LLM use makes us dumb, and that self-reported productivity gains are false. But experiencing it for myself was a totally different feeling. It gave me a whole different perspective and answered my initial question: will I get replaced by AI soon? The answer is no. I don't think AI will take my job anytime soon because it's smarter and more productive than I am. I also don't think AI will make me 10 times more productive. If I lose my job due to AI, it will be because I used it so much it made me lazy and stupid to the point another human has to replace me and I become unemployable. I shouldn't invest time in AI. I should invest more time studying new things that interest me. That's probably the only way to keep doing this job and, you know, be safe.