Hi! It's been a few months since my last post. I'll explain that here! I'll also talk about some books I've been reading, including "Magick, Mayhem, and Mavericks", "Anna Karenina", "Range: Why Generalists Triumph in a Specialized World", "Grit: The Power of Passion and Perseverance", and "The 7 Habits of Highly Effective People".
Last entry, I had big long post about something very important to me, and I set a sort of mental precedent to carry that level of effort out forever, but that makes for a very difficult writing environment. Instead, I'm gonna stick to the once-a-week format of these posts, and not set any strict word- or effort-requirement I previously imagined. If a post has only a few sentences in it, then that's fine by me!
Since the last entry, some very big personal developments happened! The biggest was my desire to LEARN EVERYTHING. I made a personal goal to read much more frequently, and I've about gotten to the point where I can read one book a week! This was spurred by reading "Magick, Mayhem, and Mavericks", which I'll talk more about later in this post, but it really encouraged me to try and be my best possible self.
Books like "The 7 Habits of Highly Effective People" motivated me to employ a set of activities to organize my goals, and one of these goals was to stop using social media. Previously, I was spending a significant amount of my time on social media (at least several hours a day), and I didn't really have anything to show for it. It was fun, but I noticed that I wasn't remembering anything I was seeing. I would scroll through hundreds of images of characters, or read thousands of words detailing updates to games or politics, and I wouldn't be able to recall any of it even immediately after reading it. It was as if the time I spent just vanished.
I thought about the motivators for me using social media, and it mainly came to keeping in touch with friends. I liked to save funny posts to my notes app and then message them to my friends, and if I stopped using social media I would also drastically reduce the amount of conversation I was having with my friends. Considering my current (in)ability to socialize with other people, this seemed like an issue. I didn't want my friends to forget about me! Although, I came to the conclusion that if friendships required exchanging Twitter posts to be maintained, then it was probably a very fragile friendship to begin with. As such, I should try to have better interactions with people and actively engage with friends directly rather than doing so indirectly through posts! I'm still working on that part, but I've freed up so much more time to do other things I really feel happy about, like reading (but more on that later).
Another small update was that a few months ago, I wore girl clothes outside for the first time! It was a simple black shirt with a black skirt, since I remembered a maxim that "everybody looks good in black", and it was to people I knew were LGBTQ+ friendly, so it was very low-stress in general. The experience was great, and I'm very proud of myself for putting myself out there like that. I'm still trying to be more outgoing, and I have a few friendships I'm trying super hard to motivate myself to continue, but small steps are being made!
Also, I changed my gender (again)! Previously my pronouns were "they/it", and I sort of mentally bounced around why I was so attached to this. I found out that, in my mind, "they/them" was for conveying my gender was ambiguous, and "it/its" was to possibly communicate my gender was more abstract to me than a simple label. After thinking more about it, I took these concepts to the logical extreme by adopting new pronouns: "EVERY pronoun", in fact! Genderfluid people tend to have "any" pronouns, but I wanted a word that better conveyed how I felt about myself. The example I thought of was when someone asks you what you want for dinner: if you reply "anything", it gives off this feeling of indecision or passiveness; but if you reply "EVERYTHING", that gives big vibes of confidence, extroversion, and personal-understanding! It communicates the same outcome as "any pronouns", but it feels more personalized and "Sophie-like". So there you have it: I'm Sophie, with EVERY pronoun.
I was in a fairly busy place during the last scheduled blog entry. I had two big projects I needed to complete for school, which each involved turning in big 10-page conference-formatted reports. Each project was to be completed over the entire semester and could be done in a group of up to 4 people, but I decided to complete it alone and mistakenly started working on it during the last two weeks of the semester. I was understandably a bit frazzled, and spend several days working turbo-mode on projects I wasn't even confident would be finished in time. In the end, everything turned out great, and I'm glad I got to use them as a learning experience in both avoiding procrastination and in implementing a cohesive project.
The first project was over an implementation of a deep learning model. The class was simply an introduction to the field of deep learning, and it required a project that sufficiently demonstrated understanding of the domain. Due to my previous interest in programming languages, I thought it would be interesting to develop a programming language—with an interpreter and all—entirely using deep learning models. Each component part of the model was proven to be possible via existing research in the field, so there was no worry of whether or not I could successfully complete it. It had an LSTM model parse user-inputted strings containing simple commands, a fully-connected model to apply soft attention given the same inputted commands, and an RNN per instruction (e.g., one RNN for addition, one for subtraction, etc.) to execute the command. As a result, the user could input a string (like "SUB 145 62"), the LSTM would parse it to get each operand, the fully-connected model would determine which RNNs to activate, and the RNNs would use the operands to calculate the instruction.
The gimmick of the model was that the RNNs needed to each be trained to essentially transform each operand into their respective results, and they do this by learning an algorithm. The algorithm may be very simple, like how the "AND" instruction simply scans a bit-stream and enables bits in the output which have associated bits enabled in the input, but I had an automatic training procedure set up on my computer to train each model overnight. Fortunately, the project turned out very well, and my presentation to the class received the highest grade despite being created last-minute by only one person. Lots of people came up to me after class to ask me questions about it, and I was very happy with the outcome.
The second project was over parallel processing. This required implementing a program that contained many threads that collaborate on a task with lots of synchronization and inter-communication. Something involving the GPU seemed to be ideal for this considering my background in GPU programming, but I instead opted to implement my own polyhedral compiler that performs automatic parallelization of code. This turned out to be a very difficult task which I was not entirely confident I could finish in time. Polyhedral compilation requires lots of math to create, and I was a bit naive in thinking I could accomplish it without fully understanding the required theory.
The project itself was made in C, mainly because I'm a big fan of C and hate all the bells and whistles of C++ (though it's more realistic to say that I don't know enough about C++ to properly utilize it). Polyhedral compilation involves a bundle of dependencies to convert the code between different formats and to handle the basic operations, so the project required installation of CLooG, Candl, Clan, and OpenSCoP.
There's around 800 lines of code for extracting strings from input source files and preprocessing them with the libraries, but the main ordeal is creating a special integer matrix to transform the code to a new format. The transformation matrix is typically created through some fancy linear algebra, and there's a dozen papers that give different hints on how you can do that, but it was far too time consuming for me to learn the foundation of the domain well enough to implement it in a non-abstract way. Therefore, I instead did a process of randomly generating transformation matrices (a process that is extremely ill-advised in a genuine compiler for any number of reasons), and doing an execution of the final optimized files to get the quickest versions that remain semantically correct. The result is a compiler that is not guaranteed to produce efficient code in any reasonable amount of time and with too much boilerplate code that makes polyhedral-applicable code sections difficult to read. Despite this, my professor was very impressed with the project overall, and it also surprisingly received the best grade in the class. It was very fun to program, and I really want to clean it up and see if I can improve it for fun.
A common theme connecting the two projects was expectations: in both cases I was too intimidated by the project to start early, I rushed through it too quickly, I made big compromises that I felt sacrificed quality, and I somehow received the best grade in the class in spite of everything. I could have avoided the stress by distributing the effort across the entire semester and better tailoring my expectations to the requirements of the projects.
Boy howdy, have I been reading. At around late May I started reading a bit after noticing a peculiar book in the shelf of a local library, and I came to the realization at the same time that I wanted to be using social media a lot less. More and more books started accruing on my to-do list, and I set a daily reading goal in order to keep up. Now, I've tripled that goal and motivated myself to read one book every week. It's a lot of work, but I'm really enjoying myself a lot more as a result!
This was the first book I noticed in the library. The title was the intriguing part; I was (and still am) obsessed with the "magic" theme, and anyone that treats science like the mystic, wonderful frontier it really is gets my attention! This book was incredible. It's a history of physical chemistry (the intersection between physics and chemistry), and it's filled with stories of real scientists who made big changes in the world through their findings. It made the first step in making me think of the kind of person I want to be. Many of these people came from humble beginnings or started science especially late in their careers, which got me thinking of the ways I could make a big splash in the science community even without graduating from the best-of-the-best. The big theme they conveyed was "don't worry about how you messed up in the past; think about how you can start moving towards your goal now!" After reading this, I decided to read a lot more books to work towards gaining the elusive title of a "polymath".
I found the term "polymath" after watching the movie "Elona Holmes" last year, which made me interested in becoming the type of person who studies every topic in order to learn everything about the world. After Magick, Mayhem, and Mavericks, I remembered the term and searched online for tips on how to become a polymath, discovering this book in the process. It talks about how "range" is a much better quality to strive for rather than strict specialization. A scientist who has a very broad understanding of many different subjects with a moderately-deep understanding of some will fare better than a scientist who knows everything about one extremely specific topic. In other words, a scientist who has perhaps a graduate-level understanding of chemistry, physics, biology, botany, economics, computer science, and five other subjects will be more likely to make a groundbreaking discovery in chemistry than a separate scientist who has 50 years of lab experience in only E. Coli. While the latter scientist will be very helpful in answering problems that apply to their own domain, the former scientist could apply solutions to completely unrelated problems with their wide skillset.
My favorite example from the book was this: imagine you are a surgeon who needs to destroy a tumor with a machine that generates directed beams of energy. The beams have a certain amount of energy that can be configured, and they behave like a typical wave (e.g., with constructive and destructive interference). If any beam reaches a certain threshold of energy, it will destroy anything it touches—including healthy cells—but anything lower than that threshold of energy would be harmless to cells. How can you use the machine to destroy only a tumor from inside the body without damaging the skin or surrounding organs?
While you consider this situation, here's a completely unrelated story: the commander of an army must attack a stronghold that lies in the center of a well-fortified city. The commander cannot bring their entire army to the attack all at once or they will be detected by the enemy; surprise is essential for this army's success. Although, they have an idea: if they split their army into tiny groups and have each group take a different road, they could reconvene their army at the stronghold at the exact same moment to commence their attack. This plan turned out to be a success, as the army avoided detection and defeated the stronghold with their combined power.
Now, returning to the original problem with the surgeon, did you think of a solution? The book Range details an experiment that was carried out with groups of real-life surgeons, and those surgeons who heard the above "unrelated story" were more likely to think of a creative solution to their problem: split the beam into several smaller parts (parts that would be too weak to destroy any cells) and have them combine at the same point inside the body to destroy the tumor. This was a genuine solution to the problem, and a solution surgeons are likely to know from their training and education, but surgeons are not likely to immediately think of this solution on their own without hearing the story.
The thesis of the experiment was that "irrelevant" knowledge could be utilized to solve domain-specific problems. Even though the scientists with their specialized knowledge could have thought of the solution, scientists who have breadth and know lots of different "irrelevant" things could convene to new, more optimal solutions. As an aspiring researcher myself who is making it their job to learn new solutions to new problems, this sounds very appealing and motivating to my goal of becoming a polymath, someone with the level of breadth required to solve these problems.
This book was made by a psychologist named Angela Duckworth who was actually mentioned in both Range and The 7 Habits of Highly Effective People. She conducted a very famous experiment on a highly-prestigious and difficult military academy that required top test scores and physical ability. Despite this, the academy—which was only a few months in duration—had a dropout rate of about 30%. Duckworth tried to find a way to predict the dropouts using some metric, and she came to the conclusion that "grit" was the best quality to search for.
Rather than "intelligence" or "ability to learn new information quickly", grit was a measure of motivation. If you're someone with a lot of grit, you stick to a task and complete it no matter what the world throws your way. Duckworth gave a quick survey to each new recruit to the academy with simple questions like "how much do you stick to goals once you set them", she found that recruits who ranked high in grit were much less likely to drop out, and that other "success" qualities like intelligence had no correlation to recruits who dropped out. In other words, it didn't matter if you were the smartest in the boot camp, or if you had a recommendation letter from the vice president: if you had the motivation to succeed, it turned out to be impossible to get you to give up, and you were very likely to be successful in the end.
This was a very big inspiration for me. In the past few months, I was feeling a bit sad and unmotivated. Looking back, it was likely a very bad case of burnout. I was desperate for a solution that would make things easier for myself, and I kept imagining how much better other people must have it than me. I thought that successful researchers wouldn't have experienced burnout like I had, and I needed to stay productive if I wanted to gain success. This book showed that consistency was a much better indicator of success than raw effort. No olympian got to their position with effort alone: they just did the same things all day every day for their entire lives. This was motivating to me because it showed me I didn't need to be at my 100% all the time. As long as I took a step every day toward my goal, I would remain a candidate for success. If I kept an eye looking on the horizon and never gave up on my goal no matter how hard it got, I wouldn't need to worry about falling behind.
I found out about this book from my mom. She said this was her favorite book of all time, and the only one she read more than once. It was a bit daunting at first as it was the longest book I've ever read (800 pages!), but I can see now why so many people consider it to be one of the greatest books of all time. It was made in 1877 by Leo Tolstoy, the same person who made War and Peace, and it's mainly about a rich lady who cheated on her husband.
It's an interesting book because a huge part of it could have been omitted, as lots of little scenes happen that don't contribute anything to the plot. One scene in particular was 30 pages long and featured a hunting trip between a few characters: no characters learned anything, nothing especially interesting happened, and the event was essentially ignored after it occurred; the book is littered with scenes like these. When I first read the book, I didn't have a positive opinion of it for this reason, but after thinking it over a little longer, I completely changed my mind. Anna Karenina is the author's attempt at displaying "humanity" as a confusing, slow collection of irrelevant details about different humans. None of the characters are especially likable, and each contain fatal flaws that make them frustrating to read in their own ways, but this book was one of the best examples of "human" characters I've ever seen. It's easy to get sucked into their lives and to think about them as if they were real, and the "meaningless" scenes are absolutely essential to this effect. When I see a TV show I really like, I've found myself saying sometimes "I wish we had more scenes of these characters relaxing or being happy without conflict" because the characters are always in some conflict in order to keep the show exciting and interesting, and Anna Karenina is a prime example of this in action. Looking back, there's lots and lots of little symbolism occurring throughout the book as well, which I'm sure makes for an excellent re-read, but having it as a standalone book is still wonderful regardless. I'm very glad to have read it, and I have War and Peace on my list of books to read as a result.
The thing I learned about popular books is that "things that are popular are probably good." I think social media conditioned me to think mainstream things are bad, which made me not want to read "classics" or books that are excessively popular, but every time I read a popular book I'm blown away by the things it teaches me! This book was no exception. It sold over 40 million copies and was considered the best self-help book of the 20th century, and it still holds up tremendously well. It really made me think a lot about myself, and I thought it was super necessary for me to read.
Each of the 7 habits conveyed in the book were important, but a few ideas were especially impactful to me. The first was the idea to "sharpen your saw," and to consider such an idea as a "moral imperative." The example the author uses is a woodcutter who needs to cut down a tree; they've been working for 6 hours, and a bystander says "your saw has gotten a bit dull, why don't you take a few minutes to sharpen it?" to which the woodcutter responds "I don't have time to sharpen my saw, I'm too busy trying to cut down this tree!" The idea is that there's several things we must consider as necessary when they don't immediately seem so. A few weeks ago during the aforementioned burnout, I had a brilliant idea called "spend all day working, and it's better to do nothing than it is to do something unproductive." At the time, I wanted to read two research papers every day, which took about 8 hours of sustained effort and high concentration to do. Understandably, this led to horrible unmotivation, and I would sometimes stare at a wall for 30 minutes because I thought it was better than being on my phone. When I inevitably failed to complete the papers for the day, it would contribute to even more unmotivation and disappointment in myself. With the "sharpen the saw" mentality, I tell myself that regular sleep, exercise, and variation in my tasks with downtime incorporated in my schedule are not just healthy, but completely necessary to maintain productivity. The key idea is to reinterpret a semi-productive task like reading to be "downtime," so I would "relax" while still accomplishing a goal.
The book contained several other very important ideas to me, including "be proactive" (complete tasks that are not immediately urgent in order to make the future less busy and more productive), "think win-win" (don't treat everything in life as a competition since humans are more productive together), and "seek first to understand, then to be understood" (be more considerate of what other people have to say, don't listen to other people just to wait for the chance to speak yourself, and understand that two people can say different or opposing things while both being correct). I would recommend this book to anybody and everybody; I really think it's changed my perspective on what it means to be productive, successful, and "effective".
Woof, that's a long post. Other posts won't be this long. I'm gonna try to keep up that once-a-week schedule, so expect much shorter low-effort posts. Message me on discord at "urlslug" (though temper your expectations because I am bad at communicating). Thanks you for tuning in to this edition of: The Sophie Chronicles…
-Sophie