My Learning Journey (2017-2023)
I started my journey in tech back in 2017 as a first-year college student. Before college, I had zero coding experience. I always struggled and felt that there were not a lot of good resources available for new tech that was coming out. At that point, I was learning Android app development. I remember the days of reading pages after pages of Google developer docs. I felt I was the slowest learner amongst all the other people around me, but I loved the process. I loved watching videos, taking notes, attending workshops, and I made it my life’s mission to go through every possible page in this documentation to learn everything about Android. Guess what? I didn’t. I was nowhere close to being a good Android developer. I spent countless hours in documentation, which ultimately became my superpower. It created an eye for reading documentation and threads on Stack Overflow. Next, I started a company with my co-founder without knowing anything about computer vision. But CNN (Convolutional Neural Network) was the hottest topic back then, and I picked this up this time. I loved math and statistics. Seeing statistics come to life with computer vision and machine learning was probably the coolest thing I was learning at that time. The habit of reading documentation was a huge advantage. I moved from sklearn to OpenCV to TensorFlow to PyTorch so fast that I could not believe I was building production systems for a manufacturing company that was being used on a daily basis. This became my skill. But it was such a slow process that I always judged a new trend for too long. I feel like I missed all the possible trains that I could have gotten on if I had known how to learn faster.The ChatGPT Breakthrough
This journey continued until ChatGPT came into the picture. I remember one of the very early trends was “Explain a topic to a 5-year-old.” To this day, this is still my first step in learning new things. ChatGPT is probably the best tool to learn any skill you want, any topic you want. For a boy who has been shy to ask questions his entire life, something magical opened up. I could literally ask it any dumb question, and it would give me a very well-thought-out, easy-to-understand answer. This is the reason I envy anyone who has started their learning journey after ChatGPT. But recently, I get so many messages about how to learn AI Agents, how I learn so many things so fast, and more. I honestly never thought about it. Because to me, I was in the zone where it felt like I could build anything, learn anything, and do anything. All I needed to do was have a good long one-on-one with ChatGPT, and that was all it took. But I stepped back a little to see from the eyes of those who are actually starting out their career. It is difficult. It is super difficult to stick to things, learn faster, and decide when to move on to the next best thing. You are constantly under pressure and experiencing FOMO.The Learning Method That Changed Everything
Let’s look at this from a different angle.Step 1: Choose Your Topic and Get the Basics
First of all, make ChatGPT or Claude your study partner. Whatever you want to learn, just start by prompting like: “Explain X topic to me like I’m 5 years old” Ask follow-ups and shamelessly ask random questions that pop up in your mind.Step 2: Create a Learning Plan
Now, once you have a basic understanding of it, it is time to deep dive. First, create a learning plan with ChatGPT. Scope out the plan carefully. Make sure you are including the fundamentals and 2 to 3 interesting use cases that you can quickly test out and implement.Step 3: Design Your Build
Converse with it long enough to find the most interesting small piece of software you can think of building with it. Now ask it to create a prompt for building that software.Step 4: Build It
Open Cursor, Copilot, or your preferred AI coding tool. Add the system prompt, give it to the agent, and wait and watch. I would suggest using Claude Sonnet 4.5 for the model; it creates a lot of docs that help you grasp concepts.Step 5: Test and Stop
Now, once the entire app is generated, run it and test it. Now stop. No more building.Step 6: Create Understanding Documents
Now it is all about creating comprehensive understanding of the concept. Ask the agent to create three documents:- Product Understanding Document - What the system does and why
- High Level Document - Architecture and component interactions
- Low Level Document - Implementation details and code walkthrough
Real Example: Learning Anthropic’s Skills
I learned Anthropic’s Skills in this exact way recently. I started with the “explain like I’m 5” approach, created a learning plan, built a small implementation, and then generated comprehensive documentation. Within just a couple of days, I had a solid understanding of how Skills work and how to implement them effectively. This has been one of my go-to ways to keep myself on top of things for more than a year now. Be it MCP, Agent Skills, Subagents, and more.This is a great time to build things. Learning by building should be the only way you learn anything and everything. So if you are someone getting overwhelmed by all the things happening, just start having an honest conversation with ChatGPT, and I can guarantee you that within a day or two, you will end up having a decent understanding of that topic. Your goal should be to build something end to end using the most recent trending stuff going on in tech. Everybody is reading about it, but hardly anyone is actually getting their hands dirty. What will you build this week? Start with one concept, follow this method, and share your progress. Happy building!