I recently found myself in a situation I hadn’t experienced in about three years: having the opportunity and the time to learn a new language from scratch.
The company I work for has several developers with extensive experience in Rust, and the internal community is growing in a very positive way, creating an ideal environment for learning and collaboration.
However, on a personal level, I felt a bit "rusty" when it came to sitting down to study. I had been working for the same client for over three years using technologies like React and Ruby on Rails within a massive codebase. When you work on a project of that magnitude, the primary focus is gaining ownership as quickly as possible; only then do you get the chance to innovate with new technologies, if the architecture allows.
As I was saying, this led me to a crossroads: I didn’t know where to begin. Everything felt like an uphill battle.
I started by scouring all the documentation I could find. I reached the official Rust Book, but I still felt a sense of cognitive overload. I’ve always been a "learning by doing" type of person, so a lightbulb went off: "Why don't I ask the AI to bootstrap a project for me and learn as it generates the code?" WRONG.
That was not only a mistake but also the most passive approach I could have taken. Having an AI generate code for you while learning might help you understand project scaffolding or basic syntax, but it stops there. It lacks the "muscle memory" required for deep learning.
Luckily, I realized this quickly. My "Plan B" was to return to what worked when I started in web development: video tutorials. However, I discarded that idea because high-quality Udemy courses were around $60, and most YouTube content felt deprecated or outdated.
Then I thought of a different approach... What if, instead of watching the AI generate a project for me, I built the project increment by increment, using the AI as a real-time mentor?
The main challenge was avoiding the "autopilot" trap. To truly engage, I had to disable GitHub Copilot’s inline suggestions. I needed to write every line myself.
Since the AI already had the context of the initial project it had bootstrapped, I asked it to guide me step-by-step through a from-scratch implementation. Immediately, I noticed a shift: I was coding more intentionally. Most importantly, I had a "tutor" available 24/7 to answer foundational queries with full context of my specific implementation.
This shift in perspective transformed everything. By moving AI from the role of "creator" to that of "mentor," I regained control of my learning curve. We often focus on how to train models or how to get them to give us the "right" answer, but we forget that their greatest value lies in their ability to explain the reasons behind things.
If you're starting a new technical career, don't let AI do all the heavy lifting. Don't train AI to do your job; let AI train you to be a better developer.
Top comments (0)