I was drowning in Claude Code content. Tutorials, threads, YouTube videos, newsletters — all of it promising to soften the learning curve, most of it inaccurate, fake, or with no proper explanation of the why.
So I stopped bombarding myself and started building. What came out the other side surprised me.
I started touching the AI surface with ChatGPT and I was impressed with what I could achieve by asking with more structured prompts. Then I tried Gemini — because you know, Google — and I'd call it a very creative model. And then I discovered Claude Code. It totally caught me off guard. It wasn't just what it could do — it was how it reasoned. It didn't just answer, it thought.
Long story short, I wanted to learn how to use this assistant to its full capacity. Trial and error, scary auto-accept and uncertain planning modes — this was impressive.
After a few months of interacting with Claude Code, taking courses from Anthropic and masterclasses, I decided to build something. At first the idea was to send myself an email with the latest features and news about Claude Code, to stay updated without losing my mind. But that didn't feel right, so instead I built an app that scrapes information from different websites — trusted ones, the ones that actually help me — and publishes it to the web.
I even started with Ollama to keep costs at zero, but after testing Haiku I saw immediately what a stronger model could do. That was the moment I stopped treating this as an experiment.
Then the moment of truth came — I deployed it. What a feeling. I've done projects as a programmer and every time there's that satisfying rush — but this deploy was different. This was purely mine. I had the problem, the tools, the knowledge — so why not? And then it worked. It's alive, and it's exactly what I needed. So satisfying. I rode this bulldozer and ended up building a fantastic house — for me, for my needs.
I'm proud of what I built with Claude Code. It even has a "tip of the day" section so you learn one thing each time you visit.
If you have a problem, the tools, and the knowledge — the only thing missing is the decision to start. And that sometimes is the harder one, but the one that will make a difference.
That's how I direct Claude Code to work with me. I knew what I needed. I had my goal clear, my instructions ready, and a vision for what I wanted.
That's my role — being the Director, not the Scriptwriter. I got that analogy from the AI Fluency course by Anthropic, and it's the best way I've found to describe it. Highly recommend it — it's free.
Worked with Claude (Anthropic) as thinking partner · Apr 2026