September 19 2025.
Recently I looked into many things outside tech. some time was spent on a variety of ideas and people in the humanities and social sciences. I also thought a bit about how I got to whatever I am at now. Lots were interesting. Not too many years ago I was into some of these things, but missed a lot. I recall sometime in middleshool I stumbled upon Hume's treatise of human nature in middle school and tried to read through it. One thing led to another and I brushed shoulders with the writings of a few other of the "greats". What I took away was quite limited, slamming my head at the texts directly was difficult. I remember I decided not to read through some wikipedia articles and what not back then, probably because over my entire school years to that point my school teachers had hammered into my mind the myth that sources like wikipedia were totally untrustable.
This time round I have looked through some secondary sources more. Many of the themes and ideas are laid out clearly, which was what I was looking for.
Sometime in highschool I found myself in pure math. I'm not sure I really enjoyed it, but I found myself always intrigued by the genius of people who had left a mark. Two classmates of mine developed an interest in math in my 8th grade year. I orbited around them, but didn't do much myself. If they had not developed an interest before me I would never have later. They took the ap calculus exam that year and thought they had reached the pinnacle of mathematics, so they never went further. That summer I went through some materials for a course on calculus and found it wasn't magical in anyway. I had the impression it would be extremely hard because in my math classes some of my clasmates occasionally talked about how calculus was extremeley difficult. I figured it was true and the myth sat unactivated in my mind for many years until I realized it was completely false.
Later during the covid lockdown I went a bit further. At this point I did not exactly know what was next and went to the internet, and realized I had not even set foot into pure mathematics. After reading through some reddit posts I decided to work through baby rudin. The exercies were hard and eventually I just started skimming through the pages. It didn't take long, probably around a week to skim through the whole book. I then went back and tried to do some exercises but still couldn't, and realized skimming just made me vaguely familiar with things. I tried to work through some chapters to get a better grasp of the material but didn't have the motivation to grind through them. I eventually decided to keep going forward and skimmed through numerous texts in pure maths. it was certainly eye opening, although I had not really learned to solve anything from them, I got some idea of what was out there.
At one point I decided I wanted to do some research, since it seemed that was where all this knowledge came from and I looked up to lots of the people who produced this knowledge. One summer in highschool I visted france and I remember searching for the graves of some mathematicians and scientists in a few of the cemeteries in paris. In math I nearly went down the route of competitions and what not.
I competed in fencing for a bit before. All was well until I started doing well and eventually even won some competitions. Suddenly, the coaches were fanning the fires and wanted me to train more, and my parents seemed to buy it. The coaches were formerly part of the national fencing team in china, and they were running the only fencing club in my area, which was a relatively affluent suburb in Beijing where all the international schools were. Whatever vision they sold was quite effective, I suppose it involved me getting into some elite university or something, They had all kinds of incentives, they would make more money if I trained more, and it would also lend them a good amount of prestige for their members to do well in these competitions. For a while I found myself training 6 days a week, and being sent to compete every other weekend, despite complaining about wanting to have more time for myself. I always objected, but it was unfruitful. My parents tried to punish me for refusing training or competitions, but I didn't have anything they could take away. They were much more emotionally invested in how well I did in these competitions than myself, I didn't really care much for how I did at all.
This gave me some idea that there were all these competitions and forces to join them and I tried to avoid their influences later. Getting into a presitigious university or job struck me as something quite similar. My relatives and friends and parents friends and what not always seemed to be more interested in how I did in these things than myself. In highschool many of my peers spent all their time doing things like extracurricular activities for college admissions that were useless. Few enjoyed what they were doing at all and most of their lives were based around getting themself into a good college. Years later after some encouters with peter thiel's writing and talks, in some places I found a sense of familiarty. From my personal experience I would add that the forces of conformity are often much smaller than they appear. For me it was just these few people around me, in this case namely... my fencing coaches, who uprooted my entire life by selling some vision to my parents. I don't think it's much different for when you get sold some similar vision to grind it out in big law or tech or banking or what not.
People around me were trying to tell me that over time if I did well I would learn to enjoy these things. I couldn't learn to like these things, and even if I could I wouldn't want to, since in the present I still didn't enjoy them. Years later when I asked some people in startups why they were working on boring problems they used a similar justification. I don't know where they got this from, at the least I hope they thought of it themself. It would be quite a tragedy if they were just parroting it from somebody else.
After complaining failed I finally got myself out of the grueling 6 day a week training schedule by faking being unable to train. I have been unfortunate to be plagued by asthma for my entire life, which I inherited from my mother, but here I used it as an excuse to quit and even found myself a bit grateful for this condition that had caused so many problems for me in this life.
To be good at something you don't enjoy is a strange curse, and by all chances I got some taste of it. When I think about this, the life of Paul morphy comes to mind. His was an interesting case, in his early years he enjoyed chess and was probably the greatest natural talent to have been. Not long after he became the world champion and retired at like 22 to try and start a law practice. Everybody labeled him as a chess prodigy despite trying to distance himself from the game, his law career was mostly a failure, and he was miserable. He later developed all kinds of paranoias and what we now call "mental diseases" and isolated himself from society then died at like 47. (I fully agree that mental diseaess exist, but I think a lot of what we call mental diseases today are actually state of mind problems). I was nowhere near as talented as Paul Morphy, but I had no intentions to continue grinding it out at fencing for years more just because I could win some regional competitions.
I was careful enough to not find myself in the same situation with mathematics, by simply not starting to get into these competitions against the vigorous insistence of my mother. I shut up about what I found fascinating in math and did not try to impress anybody. In that brief period of time just after finishing some courses in calculus and linear algebra and before discovering what was out there in pure maths, if I had participated in any of these competitions and by some misfortune did well in them, things would have turned out radically different for me. I would have lost interest in maths years earlier then when I eventually had a glimpse of what was out there, and it would have led to me never having the courage to do any research in quantum error correction or any such field later (which I guess was also a formative experience), because I would have been too intimidated and the math invovled would have striked me as totally arcane. I have no idea where my life would have gone in that case. Maybe I would have become a musician, there was a time later where I was quite infatuated with music and it's possible things could have gone that way. Although somehow my love for music at the time eventually morphed into hanging out at this luthier's workshop and making instruments. I think I was inspired by Alexander Scriabin, he made pianos, composed, and played.
In my first semester of college. I tried to get to know the place and the people, but I didn't enjoy them. For a final project for a natural language processing course, I trained some authorship attribution models, and they got up to like 89% accuracy on a handful of authors I scraped from project gutenberg. They classify if a text was written by a certain author or not. This authorship attribution approach uses a few thousand words of each persons text to learn their personal writing style, which I believed would be more accurate than approaches that try to distinguish between the broader categories of human and AI generated text. I remember there were some things out there like GPT zero and originality.ai and what not. Since I already had the core technical part working, I just threw it behind a website and tried to launch a startup out of it. Maybe partly becase I thought doing a startup in a college dorm seemed like a quite mythical thing. I talked to some school admins and tried to gauge their interest in such a thing. I decided to stop because I was using AI generated text in my own school assignments, and figured that I would rather do many other things than spending a few years building it out.
I quit college after a first semester since I didn't enjoy it and one thing led to another and I started chasing the whole trend of quantum computing. I hung around some labs but they all wanted to use me to do dirty work so I left them and was suddenly beginning to work on quantum error correction. I had some vague understanding of how chain complexes and their properties and operations worked, which was a language used quite heavily to study quantum error correction. My brief encounters with topics in homological algebra and algebraic topology and what not from skimming through all those books years earlier in highschool had paid off here. At the very least it gave me the confidence and patience to go back and figure it out. It's quite probable the limited knowledge I had acquired from before may not have been of any other use than the courage I found from it. At one point I chatted with who would later become my coauthor and he helped polish it up and do story telling and tell me where some extra results could be helpful after I wrote up a draft and produced all the main results.
In pure math it was never clear to me what I could contribute, since I was well aware that I wouldn't really be able to solve any open problems that were interesting and relevant, much less introduce any problems that would be found interesting and relevant. So I later ended up doing research in quantum information largely because it was somewhere that I could actually do something meaningful, and my tastes had changed to be more applied at that point after brushing shoulders with some people who talked about changing the world and what not.
I eventually churned out a paper and became the SOTA which gave me some confidence in being able to do things that could be meaningful. It surpassed methods from a team at Xanadu as well as Florian Marquadt's group, and his post doc who was now a senior scientist at nvidia. All of a sudden I realized I could produce results better in some key metrics than researchers at huge startups with hundreds of millions in funding and top academic groups, even though at the same time it was obvious to me that what I did was utterly insignificant in the grand scheme of things. I didn't really know what to think of it, but I figured I chose something too small to work on and next time around I should choose something bigger. The type of problems I work on will never be the same after that experience. For the fall of the following year I had went back to university, if not I probably could have done it in like half the time. This time I still only stayed for a semester, I took courses in a variety of engineering fields which I did not know much of before, and not a single course in computer science. Same as last time, I quit because I didn't enjoy the environment or the people.
Around mid April this year, I was very curious to see what would happen after I reported our harvard law school educated ex partner landlord to the berkeley housing inspection office. Of course, his place had some problems that warranted me reporting it, and it was not just out of curiosity, but if he wasn't such a well decorated lawyer I would have just moved on as my lease ended in late May. I thought he might do something interesting in return, what ended up happening was he made up some bill for refunding like 150 dollars and asked for my address so he could send it to me. Of course, I didn't give my address, if I did there's defnitely a lot more things he could do to me that I was not going to risk. Long after I moved out, on the day of a scheduled reinspection, I got an email notifying me that the follow up inspection was cancelled due to "short-staffing", and they haven't rescheduled it. He must have pulled some strings.
After reporting the landlord I was looking to find a friend's place to live at, I moved in to join a startup as cofounder with a friend I made about 2 years ago who raised 36m in his last startup and trained a bunch of diffusion models, so I thought we were setup to at the very least fail gloriously, but I quit some months in when they decided to do a consumer social app. I figured cursor was raising huge amounts of cash to train their own foundation models, and later I asked a friend of theirs at cursor (who joined in feb 2025 but whose equity is already worth like 300m somehow), he deflected the question at the time, but about a month later he admitted to it and showed us some stuff he had "vibe coded" up with one of the checkpoints he was testing. Not long after this cofounder of mine decided we should also just do something with high growth for like half a year or so while cursor is training their model, and then if cursor succeeds we could also raise to train a model since investors would believe in the whole narrative of agents companies turning around to train their own foundation models, and if they failed to train a model we would also probably fail. The reasoning was solid but I left anyways. If I didn't decide to report my landlord I may have not worked on this startup and shopped around a bit more, as I needed to find somewhere to live within a week. btw, the 200$ subscription that cursor offered was in part because we kept telling this friend they should add it, and after he got it added they were loosing too much money on it so they cut usage limits not long after and got tons of backlash for it.
The strangest thing is when looking back a bit I feel I have never really chosen what I would do very deliberately, even though it felt like I had made those choices. It seems I have often found my hand forced in new places and situtations. Of course now that I have somehow miraculously developed some awareness of these things, I should be more deliberate about these small things as well. Much of my life must have been the sum of effects of some small pertubations propagating into huge effects.