Who would only start writing the 2025 year-end summary at the end of Q1 of 2026... I've almost forgotten what I did in 2025; I'm only making up for it now after seeing this year's record was missing. Luckily I have logs to look back on what I actually did.
On the Move
In May I went to Oakland for a conference—it was indeed an unforgettable American journey (if we don't count the nausea, dizziness and drowsiness brought on by jet lag).
After returning to China, I rushed nonstop to Changsha to participate in a competition final—a real challenge after a challenge. Fortunately the result was acceptable in the end; I got a first prize.
Forward, Pivot, and Retreat
Looking back at January when I was still hitting the wall with frama-c, and then at the agent work pushed forward in July, who would have thought that a casual word from an LLM at the time could set research progress back a full half a year? Maybe fate is just like that, punishing every person who works for the sake of working, who covers up essential laziness with false diligence (actually it's not laziness, just plain incompetence). I also realized from this matter that to deeply understand something, you need to focus on the details. Even if no results come of it, you can still gain new knowledge in the process.
By the end of the year there was no choice but to rush a paper to OSDI, and as expected it was rejected. So when it comes to research, you really have to constantly remember what you're doing, otherwise you're always wasting time.
The third work's starting point has also been very long and drawn out, hitting walls everywhere. I feel I really need to dive into the details, actually do some work, before I can discover the problems. Who would have thought that not until mid-2026 would I settle on even a vague direction? Actually there are many ideas I want to work on, but none have anything to do with my dissertation theme, so they've been shelved for now. Perhaps after graduation I can take them up as a hobby, write some papers, and submit them.