Difference between revisions of "User:Cjq"
|Line 196:||Line 196:|
did i go to any lectures? what were they like?
did i go to any lectures? what were they like?
what did i *actually* learn after the class?
what did i *actually* learn after the class?
Revision as of 01:23, 4 August 2022
|Courses||Mathematics, Electrical Engineering and Computer Science|
|Classpect||Seer of Mind|
Now every single crack, every penny that I pass
Says I should either leave or pick it up
Oh, with every single buck I've made
I'm saddled with bad luck that came
- 1 Classes
- 2 Clubs
- 3 Puzzles
- 4 Some favorite media
- 5 External links
- 6.3900 [6.036]: Introduction to Machine Learning. I did not go to any lectures. I think I grok some basic machine learning concepts after taking this class. It's given me enough confidence to pursue my own projects. The real mathematical prereqs are knowing the chain rule and how matrix multiplication works. The class is more code-intensive, with homework every week having a programming problem. Some psets are fun, others tedious. Don't worry too much about the code, as tests are entirely conceptual. I can now read OpenAI papers without looking up every other word. One hour a week on psets. (I've been an LA or TA for 6.3900 [6.036] every semester since, so talk to me if you want insight about how it's changed over time.)
- 8.022: Physics II (Harlow). I only took this class because my adviser forced me to take a GIR. I'm in a love-hate relationship with this class. On one hand, the physics is cool, and exciting, and seeing how things are derived fills me with ecstatic joy. The format was probably nicer for my schedule than TEAL, the lecture notes are great, and it actually feels like I solved problems. On the other hand, this was my most time-consuming class this semester. The correct move was to probably test out of the 8.02 requirement, take a different GIR, and shuffle other subjects around to get things to work. Two hours a week on psets.
- 15.000: Explorations in Management. Decent food every Monday night. I found the lectures to be pretty hit or miss in terms of quality; it's probably a good idea to bring a pset to work on if it turns out to be on the bad end. But the good lectures are really good. Try not to miss the nights when you talk to current Course 15 students, alumni night, or the night you play the simulation. Has swayed me to consider taking more 15 classes, since some of them have some good math. No out-of-class work.
- 18.701: Algebra I (Artin). I knew some of the material coming in, so I'm not sure if taking this class was the right decision. I pretty much just took the class because everyone I knew was taking it. I should have taken 18.700 instead. I'd probably advise freshmen considering 18.701 to consider whether they really want to. I did meet two of my best friends from the class that I wouldn't have met otherwise (non-MOPpers take the class, who knew?) and it fulfills my 18.06 requirement. Three hours a week on psets.
- 21W.022: Writing and Experience: Reading and Writing Autobiography (Harrison Lepera). My favorite class this semester, because the essays were literally like writing blog posts, except I got graded for them. A CI-HW where you get to write about yourself, which was a good way to get rid of my CI-HW requirement without having to write philosophy essays. Be aware that different sections of the class do vastly different things. Spikes of work with each paper to submit.
24.93: The Search for Meaning. Dropped. I think I went to the first two lectures, realized it was a class about semantics, and then lost interest in going. I often had better things to do with my Thursday nights. Semantics just isn't my cup of linguistic tea.
- 6.S087: Mathematical Methods for Multidimensional Statistics. "All the math I wish I knew before 6.3900 [6.036] but didn't." Difficulty ramps up fast. Lectures were okay; psets were really good. The class claimed to teach linalg from scratch, but I doubt you can make much progress in the class without knowing some linalg concepts beforehand, unless you spend a lot of time during the class learning linalg. One hour on psets.
due to the coronavirus pandemic, spring 2020 did not proceed as usual, so it's hard to make judgments about how these classes would have been in a normal semester.
- 6.1010 [6.009]: Fundamentals of Programming. I only went to office hours to get checkoffs. I think the only thing I learned was "how to not accidentally make Dijkstra's O(VE)", as I already had lots of experience writing in Python. Like many programming classes, you write projects from scratch, and then throw away your code after a week, which doesn't really feel like "real-world" programming. I did have fun doing the psets, especially thinking about the nicest possible ways to write things, and constant optimizing silly things. Labs took me two hours on average.
- 6.1220 [6.046]: Design and Analysis of Algorithms. I was familiar with the first half of the course from preparing for the IOI. Lecture and recitation notes are well-written, so I didn't go to anything. I enjoyed thinking about reduction from NP, and probabilistic and approximation algorithms; probably not worth taking just for these topics, though. For this year's tests, it was usually clear what algorithm to apply, making it slightly less "tricky" than a CF round or a 6.1210 [6.006] test. So 6.1220 [6.046] + another algorithms class might actually be easier than 6.1210 [6.006] + 6.1220 [6.046], which could be one reason to take the class. If you do all the psets you can save lots of time preparing for tests; I think psets took me an hour on average.
- 18.218: Topics in Combinatorics (Postnikov). This year was about polytopes. Tripled my understanding of matroids, which I think will be useful. Postnikov's lectures were locally very organized, but globally disorganized; individual classes were clear and had lots of examples, but as a whole each week felt self-contained, less reading a textbook and more reading papers. There were only two problem sets, you only have to solve a small subset of problems, and there are a mix of easy and hard problems to choose from, so it's a pretty light load. Eight hours each pset. The kind of class where you get as much as you put in.
- 18.702: Algebra II (Shankar). Felt very fast pre-COVID, and then very slow afterward, so it's hard to give an enduring judgment. As a lecturer, Shankar is dynamic and patient: he's energetic, but will slow down to answer questions. His coverage has a very slight categorical tilt, choosing to think of representations in the abstract, having smaller "toy" examples in the Galois theory part. As a professor, Shankar is very kind and approachable. I was kinda sad we didn't get to talk much about algebraic number theory; I wanted to learn more about quadratic fields.
- 21G.012: Exploring Globalization through Chinese Food (Teng). For the price of writing page-long essays every two weeks, you get to eat food in class, which is a good deal. The writing isn't that bad, and the coursework was actually interesting; I learned a lot about Chinese food, and thought a lot about globalization. Teng puts in the effort to remember not just everyone's name, but where everyone's from. Fun. Less than an hour out-of-class work.
- 24.900: Introduction to Linguistics (Pesetsky). The least work you can get away with in a CI-H. Pesetsky posts his slides, which can substitute paying attention to lecture. Attendance wasn't checked past the first week. Not worth buying the textbook. Non-syntax psets feel like linguistic olympiad problems. Drawing syntax trees is fun once you get the hang of it. The research critique is the only "hard" writing assignment; the rest are just describing what you've learned in fieldwork. Post-COVID, the Piazza was extremely helpful. Fieldwork is very fun if you pick a language with sounds you can understand. Two hours a week on psets.
- 9.URN: Undergraduate Research. Did a UROP for the Computational Psycholinguistics lab. Wasn't great.
- 18.098: Internship in Mathematics. Was a counselor for PROMYS.
ditto spring 2020 note, but this is probably closer to a regular semester
- 3.091: Introduction to Solid-State Chemistry. Due to the semester, lectures were all pre-recorded and recitation wasn't mandatory, because quizzes could be taken asynchronously. I didn't watch the lectures or go to recitation, instead just reading the problem set solutions before taking quizzes and exams. Which was fine because quizzes and exams this semester were also open-notes, and I was also taking it on P/NR. I averaged spending an hour a week on it.
6.5310 [6.849]: Geometric Folding Algorithms: Linkages, Origami, Polyhedra (Demaine). Switched to listener after three weeks, then stopped attending after two more weeks. Took this class because I know 20ish of the 80 or so people originally enrolled. Genuinely interesting material at first, but it stopped being interesting to me after a month in. Lectures were all pre-recorded so classes were all problem-solving sessions, which I really enjoyed. Demaine is apparently someone to pay attention to when they teach a class.
- 6.5210 [6.854]: Advanced Algorithms (Karger). The heaviest class I've taken so far. I think I averaged 15 hours a week, mostly through the challenging and time-consuming psets. Psets were hard for like two weeks, then easier for a while, and then hard for three weeks, and then easier again, so maybe wait a while before considering dropping. All the algorithmic parts of 6.1220 [6.046] that I liked, and way, way more. Material's interesting, and Karger is a fun, excellent lecturer, with the polish that twenty years of teaching algorithms brings, although you could probably get away with skipping class and watching recordings sped up. If you're unambitious and have decent groupmates, the final project shouldn't be too heavy, even if you're shooting for an A.
6.8180 [6.883]: Advanced Topics in Artificial Intelligence (Drori). Dropped. Topic was meta learning, and the first two lectures were really interesting, but it turns out I wasn't actually enrolled in the class?
- 18.404: Theory of Computation (Sipser). Didn't show up to lectures or recitations. Sipser's book is amazingly well-written, and as you can imagine, the class follows the book pretty much exactly. This is the class where you get to do all the fun NP-hardness-style proofs in 6.1220 [6.046]. Psets were once every two weeks, and I averaged around eight hours per pset, inclusive of the time it took to read the book. Exams are pretty tightly timed, probably expect to finish all questions only if you never stop writing solutions, but you don't need all the questions to get a good score anyway.
- 21M.600: Introduction to Acting (Torn). The class was actually pretty fun, and I'm not sure if it's because of Torn or because there were only seven of us in the section. The first two-thirds were entirely improv-style exercises in pairs and triplets, which I enjoyed because I actually kinda like improv, and I got to learn things like playing high-status characters. The latter third of the class was doing scripted scenes and monologues. There was only one two-hour synchronous class per week, and homework was some reading assignments with two-paragraph long essays, so I averaged around three or four hours a week.
- 24.902: Language and Its Structure II: Syntax (Pesetsky). The first three weeks were mostly review of 24.900, perhaps because Pesetsky also taught my 24.900. But the rest of the class was some really exciting syntax, and I enjoyed drawing trees and moving things and adding structure to sentences. In an alternate universe this class could have convinced me to become a linguistics major. I went to the three hours of lecture per week and skipped recitation. Homework was similar to 24.900 syntax homework and averaged me around two hours per week. I crammed the final paper over an afternoon and was fine.
didn't take any classes, just worked on mystery hunt
ditto fall 2020 note.
- 6.5120 [6.822]: Formal Reasoning About Programs (Chlipala). Hands-down favorite this semester, contender for favorite of all time, made me consider further work in the area. Coq is incredibly fun, and playing with it makes the six-hour-ish average time I spent on psets really worth it. Class is 100% weekly psets, and I found the psets to be relatively high-variance in terms of how long they took, from three to twelve hours. Chlipala is an excellent lecturer too.
6.5230 [6.851]: Advanced Data Structures (Demaine). Switched to listener. I want to contribute to the class forum and coding projects but I don't want to do the psets or watch the lectures or do a project. Probably would've loved the class in-person, though. Also couldn't make Monday afternoons, and Thursday nights are just hard.
- 18.212: Algebraic Combinatorics (Postnikov). On exploratory. Second class with Postnikov and quite similar, style-wise, to the first class I took with him (18.218). Genuinely fascinating material, lots of brilliant bijections, excellent lecture notes. Grade is 100% psets, there were three psets for the whole semester, and each took six to eight hours.
18.408: Topics in Theoretical Computer Science (Moitra). Dropped. I was scared with the amount of analysis in the first lecture. Also three-hour lectures are just hard.
- 18.453: Combinatorial Optimization (Franks). First class was really exciting, but I stopped going to anything after two weeks. Sadly, I was more familiar with the material than I initially expected; notable overlap with 6.5210 [6.854] and 18.218. Franks pretty faithfully followed Goeman's notes, which are really well-written. Biweekly psets, each of which took me around four hours. Midterm and final are weighted heavier but are easier than psets.
- 21W.755: Writing and Reading Short Stories (Lewitt). Fiction is hard, but Lewitt has a great personality and her classes were really fun! Class is probably really professor-dependent. We wrote only two short stories, between 1000 and 5000 words, for the whole semester. They get workshopped and revised, but that's most of the grade. Reading-wise, it's two short stories before each class. First third was spent discussing random short stories, rest of class was spent workshopping everyone's stories. Three or four hours a week reading, around eight hours of writing the week before a story was due.
- 24.903: Language and Its Structure III: Semantics (Elliott). Went to three hours of lecture per week, skipped recitation. I'd probably have loved this class if I was less familiar with the material, or if it was taught by someone else. Elliott is a kind, great, organized lecturer! But his approach to the class was less mathematical than I expected. For example, the class never officially used lambdas, although I did use them in psets. Psets were weekly-ish, took me anywhere from two to four hours each. There's a final, but no final paper, unlike 24.902.
- 18.098: Internship in Mathematics. Returned to PROMYS as a head counselor this time.
ditto fall 2020 note.
changed from 18c to 18. declared double in 6-3. declared linguistics concentration.
- 6.5110 [6.820]: Foundations of Program Analysis (Solar Lezama). Did not go to recitation. Similar content to 6.5120 [6.822] to begin with, but diverges at times; this class has a more practical bent. Solar Lezama has a great sense of humor. Six biweekly psets, in Haskell; Haskell and Coq; Coq; OCaml; OCaml; OCaml; averaged six hours each. Psets somewhat related to lectures, but are also relatively self-contained. I don't think I learned much formal verification, but I did get exposed to more functional programming.
- 18.03: Differential Equations. Went to nothing but the exams. Weekly psets, averaged two hours each, inclusive of reading. You can frontload a decent amount of work. I learned basically nothing but did well on the exams because I memorized the practice exam solutions. Three midterms and a cumulative final, no psets on weeks with exams. Boring, but it's a REST and a math major requirement.
- 18.510: Introduction to Mathematical Logic and Set Theory (Cohn). Cohn is one of the best math lecturers I've seen, and he has an excellent sense of humor too. But this is also an 18.404-like situation, where Cohn's book is really well-written and can definitely substitute going to lectures. First third felt slow to me, but the rest of the class was really good, and satisfied what I wanted to know from logic. Weekly-ish psets, average five hours. Don't be scared off by the first pset; it's harder than the rest. Take-home, open-book midterm and non-cumulative final, so exams are not too big of a deal as long as you remember to do them.
- 18.821: Project Laboratory in Mathematics. A CI-M and a lab, so it's focused tightly on communication and class period tends to be freeform. Instructors care a lot about how the class works, which is really awesome. Work on three different projects that are either well-known or open. You'll be with the same team of three through the semester, so take the class with two friends if possible. But if you don't, not the end of the world; I didn't and I ended up becoming friends with my teammates. Really tight turnaround on write-ups; you're basically writing up results as soon as you get them, so write things down as early as possible. Spikes of work near deadlines, averaged eight hours a week in- and out-of-class. Learned a decent amount about how to write math up well. (On the other hand, 18.821 might change formats after this semester, which might be good? Who knows.)
- 24.917: ConLangs: Construct a Language (Richards). Didn't go to anything and didn't learn anything, but I hear Richards is an excellent lecturer. Large overlap with 24.900 material-wise, except instead of working on a different language, you work on your own. Incredibly fun weekly psets, averaged two hours a week, plus a final project that took me six hours. Worth it to take even if you already know 24.900.
didn't take any classes nor had any commitments
ditto fall 2020 note.
- 6.1910 [6.004]: Computation Structures. Did not go to anything. I learned a decent bit about writing assembly and specifying circuits, but I doubt I'll ever use that info. Technically 6.1900 [6.0004] is a prereq, but assembly and C aren't strictly necessary. Labs were two to four hours each, I got checkoffs two at a time. Tests are high-stakes, unrelated to labs and will need study. I can see why some people would like it, and indeed, most do. (Is it because it's satisfying to "build your own computer"?) But working with RISC-V and Minispec is frustrating.
- 6.1800 [6.033]: Computer Systems Engineering. "The class that taught the most useful things in the worst way," which I agree with. Did not go to lectures, went to (mandatory) recitation and tutorial. Systems design has some of CS's most fundamental and fascinating ideas, but the class is sullied by its awful design project. 6.1910 [6.004] is a fake prereq. You will work closely in a group of three all semester. Two hours per week to prepare for recitation, biweekly hands-ons each took an hour. Writing for the design project is a timesink, averaged two hours weekly. Exams are important, but so is participation. Probably an ideal class to take listener if you can help it, so you can only do the lectures and readings. Take it as soon as you know you'll need it; try not to take it senior year.
- 6.5151 [6.905]: Large-Scale Symbolic Systems (Sussman). It's worth going to lectures just to listen to Sussman, who is old but funny. Liking this class will be largely a function of how much you like writing Lisp and the Lisp mindset. I'm comfortable with Scheme now, and the class does have nice ideas. Weekly psets each took me four hours; they were one of the few things I looked forward to doing every week. YMMV, because my friends didn't like them. No tests. Low-stakes final project. Might as well take 6.5150 [6.945]; it's literally one extra pset and you get more credit.
- 6.UAT: Oral Communication. Scheduling consideration: lectures aren't real, only recitations are. Polarizing because so much depends on your RI and TA; check previous evals, preregister, and be ready for first-come first-served sectioning. People always complain about subjective grading. Your RI will determine your entire grade. The class I learned the most from recently, even if I was a confident speaker coming in. It's semseter-long Splash practice. Zero writing is good. Biweekly presentations that took two hours average weekly. You can frontload a lot of work; I did. Worth taking even if you're not course 6. Take it earlier (think first- or second-year) rather than later.
- 7.014: Introductory Biology. Went to nothing. Biweekly psets took two hours each, inclusive of doing research to answer them. Tests were entirely online and closely related to psets. Likely the least work you can get away with for the Biology GIR, if you're not taking the ASE, although the ASE might be less work. Took this class on P/NR.
24.910: Advanced Topics in Linguistics Analysis (Abramovitz).Changed to listener after a month. I love Rafael and Mitya, and I love morphology, and I love the lectures, I really do! I just don't want to do the psets because there was so much writing involved, the class being a CI-M. Stopped going to lectures for last third of semester due to a lack of time. But I love this class so much.
- 6.9820 [6.920]: Practical Internship Experience. Will be a front-end intern for Instabase.
- 6.5620 [6.875]: Cryptography and Cryptanalysis.
- 18.704: Seminar in Algebra. I got lotteried into this (out of 18.204, 18.434, this one).
- 21A.155: Food, Culture, and Politics.
- 21L.489: Interactive Narrative.
- 21L.580: Translations.
- CMS.300: Game Studies.
Remaining requirements, in the rough order I plan to take them.
- (18) CI-M. Will keep lottery-ing until I get into one of 18.204, 18.434, 18.504.
- (GIR) HASS. I'll probably get this done.
- (6) Intro.
- (6) 6.1200 [6.042]. Need to petition, I guess?
- (6) 6.1210 [6.006]. Ditto 6.1200 [6.042].
pain, absolute pain, absolute awful horrible terrible pain.
- floorpi. I am the hall chair. I just needed to list this down so I remember to list this when I say "these are the extracurricular commitments I have".
- Tech Squares. MIT's square dancing club. I'm currently a member-at-large. We dance every Tuesday night; it's one of the only ways I get exercise. Ixa dragged me into it. It's surprisingly mentally taxing. I think the only way to get an idea of what it's like is watching a video, or seeing a dance live.
- Educational Studies Program (ESP). We run learning programs for middle school and high school students. I'm an admin and help run our programs. The flagship one is Splash, where over two thousand students from around the country come to MIT's campus to learn from classes taught by the MIT community. If you do end up being ESP admin, my advice is to keep careful track of how much commitment you give.
- Assassin's Guild. A live-action roleplaying group. I'm currently the Secretary of the Exchequer, so, the treasurer. There's a LARP every other weekend during the semester; I participate in a game around once a month or so.
- Student Information Processing Board. Glorified computer club. I'm the secretary. Contrary to the name there is no board, nor information being processed. We just like computers, I guess. Come to meetings, Mondays 7:30 PM, W20-557.
- Filipino Students Association (FSA). We run events celebrating Filipino culture. We run events once or twice a semester; last semester we had a grocery outing and a karaoke event.
- MIT Puzzle Club. We help run the yearly Mystery Hunt. We also run puzzle events every two months or so. I'm currently the web person.
- Live-Action Mafia. We play two to three live-action mafia games per semester. During a game, the commitment is roughly an hour a day while alive. Unlike usual mafia games, everyone has a power role and gets to do something. There's a lot of investigative roles; the emphasis is on deducing the mafia through actions rather than psychreading. It's fun to tap on people's shoulders and say "bang".
i only really started doing puzzlehunt-style puzzles in september 2019, very vaguely working on mark halpin's labor day extravaganza with pi. i was aware of the existence of these puzzles before that, and I guess the first of those kinds of puzzles I solved were for hall rush. i went to the puzzle club event that month, and maybe in november and in january. the first hunt I really participated was with galactic for mitmh 2020, so talk about a puzzly education from a firehose. hunts i've joined:
- January 2020: MIT Mystery Hunt 2020: ✈✈✈Galactic Trendsetters✈✈✈, 1st.
- February 2020: My Little Pony: Puzzles Are Magic: [pi], 16th.
- June 2020: Puzzle Potluck 3: pi, 7th.
- July 2020: Smogon Puzzle Hunt 3: Pi, 13th.
- September 2020: Balancing Act
- September 2020: DP Puzzle Hunt: pros and cons[pi]racy photographs, 22nd.
- October 2020: teammate hunt 2020: ⌊π⌋, 7th.
- November 2020: CMU Puzzlehunt F'20: pi.
- January 2021: MIT Mystery Hunt 2021: ✈✈✈Galactic Trendsetters✈✈✈, wrote.
- January 2021: Inexact Puzzlehunt: ⌊π⌋, 48th.
- March 2021: HMMT Puzzlehunt Spring 2021: edited.
- March 2021: PI Hunt 7: pi.
- March 2021: Princeton Puzzle Hunt: pi, 2nd.
- April 2021: CMU PUzzlehunt S'21: pi.
- April 2021: Inaudible Hunt: [pi], 3rd.
- May 2021: Huntinality: [pi], 6th.
- July 2021: Galactic Puzzle Hunt 2020: wrote.
- August 2021: Puzzle Potluck 4: [pi], 16th.
- October 2021: teammate hunt 2021: fires[pi]nning, 63rd.
- December 2021: Silph Puzzle Hunt: [pi], 12th.
- December 2021: Puzzle Rojak: [pi], 13th.
- January 2022: MIT Mystery Hunt 2022: ✈✈✈Galactic Trendsetters✈✈✈, 5th.
Some favorite media
might expand with descriptions later on
- Homestuck. Comparable to something that's 1,300,000 words.
- Worm. Around 1,680,000 words.
- Unsong. Around 200,000 words.
- There Is No Antimemetics Division. Around 75,000 words.
- Ra. Around 150,000 words.
- The SCP Foundation (in general)
- The Unbearable Lightness of Being. Around 95,000 words.
- Cordyceps. Around 40,000 words.
- Saga, 17776, A Practical Guide to Evil, Kill Six Billion Demons, Edith Can Shoot Things And Hit Them, etc. (I should probably make a subpage for all the things I read, just so I can tell myself that I read things, because apparently not everything is on Goodreads.)
- The Man Who Loved Only Numbers
- Universal Principles of Design
- The Art of Calligraphy
- The Design of Everyday Things
- Made to Stick
- Primary Surgery
- How to Teach Mathematics
- The Mathematical Experience
- Human Accomplishment
- Reality is Plastic
- The Empathy Exams
- Short Term 12 (2013)
- Boyhood (2014)
- Mommy (2014)
- X+Y (2014)
- 2 Cool 2 Be 4gotten (2016)
- I'm Drunk, I Love You (2017)
- Love, Simon (2018)
- The End of the F***ing World
- Person of Interest
- The Last Airbender / The Legend of Korra
- Rupaul's Drag Race
- Bojack Horseman
- Undertale / Deltarune
- Ace Attorney
- Zero Escape
- Butterfly Soup
- Slay the Spire
- Imagine Dragons
- Ed Sheeran
- Sleeping at Last
- Christopher Tin
- Ang Bandang Shirley
- The Ransom Collective
- Reese Lansangan
- Rusty Machines
- The Experimental Generation of Interpersonal Closeness
- Dear Evan Hansen
- MIT Confessions
- anime, which is in my MyAnimeList
- and most likely a bunch more im forgetting about right now
- i have a private blog, and you can ask me for access