For the last 9 months (since fall 2025) I said to myself: ”In one year my goal is to publish a paper in deep learning.” I wanted to simply move towards contribution, moving from being a student to a researcher, and at that time I felt mature enough to start that process.
For me, a lot of "progress" isn't about writing a certain amount of pages, but about orientation. Or at least this is how I would summarize my first months in research. It's a lot about developing a taste for what problems you consider important. It's about finding a specific direction that you believe in and that genuinely excites you. This is not trivial.
I personally was initially drawn towards work that aims to truly understand how neural network architectures work and explain black-box phenomena like scaling laws and emergence. But it took months of reading and asking questions before I felt like I had found my research direction.
Sometimes I'd doubt this path. Currently I haven't published anything. I have written down ideas, archived them, thought I had a paper ready, then realized fundamental problems about it, then write 80 pages just to boil everything down into 10 pages. It's pretty chaotic. It can feel like I "haven't done anything" because as I said, currently I haven't actually published anything. But I believe that the navigation, the orientation, and the identification of your specific direction is necessary, and even understated.
And most of the times, when working on new research, there is no map. You have to build the map while walking it. So in that sense, the navigation is a big part of the research.