I used to roll my eyes at the AI hype. Not because I didn't understand the technology (I'd built my own ML models before), but because everyone seemed to be jumping on the bandwagon without really thinking it through. When my previous employer suddenly announced "Everyone needs to use AI now," I felt like I was being asked to follow a trend rather than solve real problems.
It turns out I was both right and wrong.
Early Challenges
My first encounters with AI at work were frustrating. The expectation was immediate adoption, and while they did build tools and guides, most of them didn't apply to our team's work. Nobody really put effort into teaching — they just expected us to adopt. We were working on legacy systems so company-specific that you couldn't learn anything transferable. The projects had rigid constraints that made experimentation nearly impossible.
Meanwhile, I was using ChatGPT and Claude for simple stuff like recipes and general knowledge questions, but nothing structured. I couldn't see how asking an AI "What's the capital of France?" would help me write better code or architect better systems.
The disconnect bothered me. I felt like everyone was using AI blindly, without the deeper knowledge needed to use it productively. So I stayed skeptical.
Taking Control
The breakthrough came when I stopped waiting for my employer to figure it out and took ownership myself. I paid for my own AI tools and started experimenting on personal projects where I had full control.
I chose Laravel for my first real AI coding experiment. Not because it was trendy, but because I knew it inside and out. I've been using Laravel since version 3, so I could easily judge whether the AI-generated code made sense. If something went wrong, I wouldn't struggle to debug it.
Laravel Boost had all the AI tooling configured and optimized, so I could just jump in and start benefiting immediately.
The Results
Using Claude Code and Cursor, I rebuilt a Laravel SaaS project that had previously taken me months to develop. I did it in a couple of nights.
That's when it clicked. This wasn't about replacing my knowledge — it was about amplifying it.
Learning Something Completely New
I decided to tackle something I'd never done before: building a native macOS app with SwiftUI. I'd always tried React Native for mobile apps but felt unsatisfied, like something was missing.
Instead of learning Swift through tutorials or videos, I learned by doing. I'd ask AI to implement something, try to replicate it, see where it failed, research the issues, and learn Swift concepts only when I needed them. Getting that app into the App Store was satisfying.
The learning structure was completely different and far more effective and efficient.
How I Work Now
These days, I use AI like a junior backend engineer. I ask Claude Code to achieve something, review the code it produces, give feedback, and let it iterate until it's complete. While it's working, I focus on the front-end and UX parts of the application.
I also use AI for rubber ducking: debugging, discussing ideas to see if I've missed anything, and coming up with implementation plans. It's like having a coding partner who's always available and never gets tired of my questions.
What's Next
I'm working on integrating AI into my fitness app, Maxout, building features like voice-note calorie tracking and workout planning that adapts to each user's lifestyle.
I'm also diving deeper into prompt engineering. Eventually, I want to circle back to the technical side — learning LLM internals and building small models like I did in the past. But now I have the practical experience to know what's actually useful.
The Real Lesson
The key mindset shift wasn't about the technology but about taking ownership. Instead of waiting for someone else to tell me how to use AI, I went out and figured it out myself.
If you're skeptical about AI, you're probably right to be cautious. AI isn't perfect and it won't replace your knowledge, but using it mindfully and intentionally can help amplify your skills. Find a project you know well, set up your own environment, and start small.
The hype might be annoying, but the technology is real. You just have to find your own way to it.
Photo by Nahrizul Kadri on Unsplash