I previously had a pretty long post that I deleted, because for a while I was really broken, and after a long time I didn’t come back, but when I finally did it turned into a tomb ![]()
I want to look back at the ridiculously stupid year 2025
Finance
Speaking of finance, it’s actually stock trading. I’m not great at placing limit orders. I previously bought some innovative‑drug ETFs, but after a few days with no progress I cut my losses and cleared the position, then started dollar‑cost averaging into gold, the Nasdaq, and the S&P. Because I didn’t have enough capital, I just bought funds directly through Alipay. After a flurry of trades, the three‑month return was around 10%, which made me very satisfied. Moreover, most of the time I don’t try to time the market; if I have money I keep DCA‑ing, if not, I skip it. Recently gold has been rising nicely.
Knowledge
In the past month I had an opportunity to interact frequently over two days with various so‑called industry bigwigs, and the more I talked to them, the more I realized that most of them are specialists—very knowledgeable in their own field, but if you ask them to speak about another domain they often produce absurd or awkward jokes.
This year I have an elective called Mathematical Modeling; the major project is to try to use a model to predict the number of taxis, and a key step is recognizing taxi license plates.
Students can be roughly divided into three types: the obedient ones who actually take their phones and photograph license plates one by one; the clever ones who try to mock plates or just scrape random data from the internet to start modeling; and the ones who know YOLO can be used, but didn’t think about using video streams, and get stuck on the instructor’s requirement to “take 2000 photos.”
I think my completion level is the highest; I used the YOLOv11 + ByteTrack approach. The code isn’t hard; I got the whole pipeline running in an afternoon. I know that if they knew this method they could also produce results at least as good as mine, so why didn’t they think of it?
The root cause, I believe, is the information gap and time investment. Some skills have nothing to do with IQ; they’re like sowing seeds and reaping what you sow. As in the earlier example, the gap between industries and between specialties is huge.
Thus, what I find most impressive are things and ideas that I could never comprehend or achieve even if I spent my whole life on them. Some people are just geniuses, giving me a sense of awe, and I fully understand and accept that.
When I talk with many people, I just feel that they’re better than me simply because they entered the field earlier and have a few more years of experience. Conversely, some work I do might look like “very good work” to classmates who haven’t seen it, but to those who understand the underlying principles it may just be toy code.
Research
Although it’s called research, I feel more like it’s a large cosplay. Overall things have been both smooth and not so smooth, perhaps due to my relatively relaxed personality. I don’t think you have to publish a paper to have truly done something.
In mid‑December I had a meal with a senior sister who was going to do a graduate program at the Chinese Academy of Sciences. She suddenly told me she didn’t want to pursue a PhD, planning to go home and take the civil service exam. She felt biology was too abstract. When she applied to the Academy she wanted to produce some work, continue to a PhD, and study abroad.
I know quite a few biology PhDs, and they give me the impression that they’re not really in that field at all
One delayed‑graduation PhD was especially abstract; he told me how to trade stocks, start a quant firm, make money, then do whole‑genome sequencing of earthworms, and use the earnings to continue research.
I think research is actually mental training: learning how to interact with people, how to search literature, how to engage with public platforms, and how to manage one’s emotions. These aspects are rarely covered in class. At least for now, I find doing research very interesting and exhilarating. It’s similar to coding: you build a rough framework, feel something is off, tweak it, adjust another part, and eventually create a decent logical loop while trying to solve a problem that perhaps no one has ever
