I’m one of the skeptical people when it comes to using AI for real work. Or at least I was. I used it once to help me plan a trip, but it wasn’t doing a good job (telling me to visit a museum at 18:30, when Google says it closes at 17:00), so I thought if it makes such silly mistakes, maybe it would be a risk to use it for something more important, like, you know, actual work. As a QA with 6+ years of experience, I didn’t really see any need for it. But my opinion changed one month ago, when I found a way to implement it in my everyday life.
I’ve tried several times to add test automation to my QA repertoire, but each time so far, after I took the courses, I still didn’t understand Java at the level it was required at work. Last month I saw an opportunity to try again, and this time I felt that it was going to be different. I started using Claude as my mentor. I created a project and started feeding it with rules, practices, processes, and everything that I could think of to create this learning framework. It’s the start of week 6 right now, and it’s insane how much this whole framework has evolved in such a short time. The main idea is to write code every single day, learn theory, and create my own backend automation project. Right now I stand at 67 contributions in 5 weeks.
Of course, I’ve faced so many challenges along the way, like information overload, following random orders, Claude generating information across 15 files that I never opened (I recently solved this specific one, and it deserves a post of its own). I’ve found that this can’t be a blind-follow process. If I wanted to have a solid learning framework, I had to control every step of the way and do so deliberately. And really slow down until I had a clear reason why I do… basically everything.
The most amazing thing in this framework is that Claude will see my learning material, the exercises I do, what I struggle with, and it creates a quiz with open-ended questions, so that I’m sure that what I’ve read really sticks. I also rebuild some of the code from scratch. These two alternate every day, and I do them as a warm-up before I start learning any new material. In numbers: when I first started with the quiz (end of week 2), out of 10 questions, I got 1 correct, 5 were marked partly, and 4 were completely missed. Last Saturday (end of week 5), out of 12 questions, I got 7 correct, 3 were marked partly, and only 2 were missed. We’re talking about questions that combine code and theory – what will happen if this code runs, and why. It’s not a clean comparison, because the questions evolved with time, but the first one showed me that I didn’t learn as much as I thought.
I decided to start writing about my journey to share those principles because they’ve proven very efficient. I also wanted to see if there are other enthusiasts who use AI for learning, or if someone has another effective way of doing so. What is your way to make something stick?
Top comments (0)