learning by reverse engineering
As a heavy user of LLMs, I’ve gradually realized that I’ve developed a habit, almost unconsciously, of reverse engineering.
What I mean by that is whenever I encounter an article that I want to learn from, I no longer feel the need to understand every unfamiliar term or reference along the way. Doing that often feels inefficient, and it’s very easy to get lost. Instead, when necessary, I’ll simply copy the article into an LLM and ask it to help me understand what is actually going on underneath the piece.
Most of the time, this turns out to be a good decision. Not because the article itself is bad, but because many articles, even those that claim to be beginner friendly, are not written in a fully logical or reader centered way. They are written with assumptions, narrative shortcuts, and compromises made for a general audience. After the LLM helps me break down the core fundamentals, I begin placing the pieces back together myself, forming an understanding that feels coherent to me. The real test is whether I can retell it in my own way, not repeating the article, but explaining the idea as something I genuinely understand.
That is usually the moment I feel when learning actually happens.
Over time, when I do this often enough, something interesting begins to emerge.
I no longer treat an article, or anything meaningful, at face value. To me, it is simply a form sitting in front of me. What I actually care about is the function underneath. A fundamental idea can be explained in many different forms, but the function stays the same. My task is to peel away everything that is covering that underlying function.
Sometimes, during this process, I find myself asking questions that I do not even expect to ask. Because I am not waiting passively for the knowledge to enter my brain, I actually try very hard to tap into whatever I have experienced and learned to understand a specific piece of knowledge. My mind is often moving faster than my mouth. It is not until I hear myself say something out loud, something I did not think I would say, that clarity arrives. That is often the moment when insight happens.
In those moments, it feels as if something clicks into place. Not because I have memorized more information, but because I have finally articulated something that had been forming quietly in my mind. Hearing my own words forces structure onto my thoughts. It turns vague understanding into something concrete.
Looking back, I realize that this habit of reverse engineering is not really about learning faster. It is about changing how I relate to knowledge itself. Articles are no longer authorities to accept or reject. They are starting points.