The Lost Art of Reading Research Papers

During my PhD years at MIT, I spent countless evenings surrounded by printed research papers, a pen in hand, marking every paragraph, tracing every equation, and flipping back and forth between references to understand what led to a particular idea. It was slow, difficult, and often exhausting. Yet, it was also deeply fulfilling.

There was a strange satisfaction in realizing that I was engaging with ideas that represented the very frontier of human understanding. When I read a paper that cited a dozen others, and I went back to those citations, I could see the invisible tree of knowledge that humanity had built - each branch representing years of effort, collaboration, and discovery. Reading papers was not just an academic exercise; it was an act of participating in the evolution of thought itself.

But after my PhD, when I co-founded Vizuara AI Labs, that habit quietly disappeared into the background. Running a startup, managing projects, teaching, and building products all demanded quick, result-oriented action. Reading a dense 20-page paper slowly and thoughtfully no longer seemed practical. Over time, I noticed something missing. My thinking had become more operational, but less exploratory. I was learning fast, but not learning deep.

Recently, I decided to bring that habit back. Not out of nostalgia, but out of necessity. Because I realized that there is no substitute for reading knowledge directly from its source. You can watch dozens of YouTube videos or read multiple summaries, but none of them can replicate the mental discipline and clarity that comes from reading a paper line by line.

That realization led to the creation of my new YouTube series. It is not a course. There are no slides, no structured modules, and no polished edits. It is simply me, reading and explaining papers as they are: directly from the PDF. Some videos might be short, others may stretch for hours, depending on the paper. The idea is not to simplify the content but to make the original material accessible through shared focus and patience.

I will cover papers that I believe have shaped the field of AI and ML - not necessarily the latest trending ones, but those that have contributed meaningfully to the intellectual history of the field. Some of them I have already read and understood in depth; others I will be reading for the first time, along with you. The goal is to make this a shared experience of learning, curiosity, and endurance.

There is a certain honesty in reading something without shortcuts. It forces you to confront your gaps in understanding, to pause and look up references, to struggle a little before things click. And that struggle is precisely what transforms passive information into active knowledge.

I have no illusions about this being a popular series. It is not meant to be. My intention is simple: to revive the lost art of reading research papers, one page at a time. If even a few people find this process meaningful and join in, that will be more than enough.

Because, at the end of the day, understanding is not something you can outsource. It is something you must work for. And the act of reading - slow, deliberate, and sometimes painful - is still the best way I know to reach it.

You can watch the first episode on my YouTube channel, where I explain why I am doing this and what you can expect.

Let us bring back the habit of reading papers, not just for information, but for clarity, depth, and a renewed sense of curiosity about how human knowledge grows.