Aug 24

I didn’t make any Anki notes for anything; I’ve kind of been unwilling to learn C++, since I don’t know if I really want to do Robotics without being able to meet with my team and whatnot. (But I do know that C++ is used in TF2 [and is useful generally], which seems like it would be interesting to mod.)

---

Cool links:

• MLP: Immanetizing The Equestrian · Gwern.net: Analysis of My Little Pony. Interesting.
  • Friendship is Optimal – Fimfiction: A fan-fiction of MLP Gwern mentions: there is an MMO, and an AI “Princess Celestia” who tries to maximize the happiness of the players. Dystopian–just makes me think of AI alignment. (I haven’t actually read it, but the premise seems cool.)
• Reactionary Philosophy In An Enormous, Planet-Sized Nutshell | Slate Star Codex AND The Anti-Reactionary FAQ | Slate Star Codex: I didn’t know any details about reactionary philosophy [only that it’s alt-right], and these were a helpful guide to them. I found that I was afraid of accepting what I read in the first article, because I didn’t want to be “racist”/”sexist”/etc., and was correspondingly more accepting of Scott Alexander’s counterarguments. I found that I was easily convinced to agree and disagree with Reaction, which kind of scared me; that seems to happen to me when I read many books. (That reminds me of this.)
• Concept-Shaped Holes Can Be Impossible To Notice | Slate Star Codex: It’s hard to notice whether you “don’t actually get the idea” or “the idea is extremely obvious to you”. Very related to the typical mind fallacy.

Mobile Games Review

I played a bunch of JRPG/gachas/anime games on my phone. Not included is “Epic Seven”, which I didn’t play as much of and forgot a lot of details (I was half-asleep when I played it). A lot of them share element which don’t seem to actually be really defining: [way too] many different kinds of currencies and materials, bases/dormitories, auto-battling stages, daily log-in bonuses. It was a lot to keep track of, despite not being entirely necessary, so it just made it harder to understand what to do (though, I played them like 4 – 8 hours straight at a time, over only 2 – 3 days, and often while I was tired; so maybe my experience isn’t the expected one).

I would recommend not playing any of them, playing a different game would’ve probably been more fun and [feel like] less of a time-waste. (Maybe play Mahjong Soul, because that’s just mahjong.) I guess, if you want anime or RPG, just play an RPG (I personally like Spiral Knights) or watch anime (I want to watch Serial Experiments: Lain).

---

Game	Aesthetics	Gameplay	Story/Characters	Total
Guardian Tales	4	4	3	10
Girls’ Frontline	3	2	2	7
Arknights	4	4	4	12
Mahjong Soul	3	2 – 4	2	7 – 9

[Aesthetics] + [Gameplay] + [Story/Characters] = [Total] | Each part is ranked between 1 – 5

• Guardian Tales
  • An RPG: heroes and items, metroidvania elements (a lot of backtracking), and small puzzles.
  • I liked the main menu music and [I think] the hero-picking screen music (I can’t find anything online besides the main theme though (which I don’t like that much)). The art and pixel graphics were good.
  • I liked the metroidvania aspects, because it was satisfying to see a puzzle which I would have to figure out later. Those (the backtracking and the puzzles) were fun, but for the most part it was just hack-and-slash gameplay (I think I got really strong weapons too early: some levels featured mini-tutorials about dodging and stuff, but I just “W+M1″‘ed my way through).
  • The fantasy RPG plots was lame and uninteresting: it’s fantasy, but all the actual people and world-building are pretty mundane (just a kind of hero prophecy–which also doesn’t match [my beliefs about] reality).

• Girls’ Frontline
  • You command T-Dolls (android anime girls) to fight for you. They all carry a specific firearm, which is also their name (I thought that was kind of neat: similar to the different Homestuck trolls corresponding with a astrological sign).
  • The plot is boring: you are a new commander, and you have to take on a “difficult” job (which is actually very easy). There was a base, and a cafe/bar where you could eventually talk to your T-Dolls, but I never got to that level, so there wasn’t any story to the T-Dolls.
  • And I got T-Dolls very quickly: I would finish a mission or claim a quest and just inexplicably get a new T-Doll, and you can get the multiple copies of the same T-Dolls [since they’re robots]; so planning with them not fun (I had to settle with less-than-optimal).
  • Entering a stage: you moved your T-Doll squads around a network (nodes and connections) to capture nodes, fight enemies, and ultimately win. This was really straightforward[/boring] strategy-wise; but an annoying aspect was that, the steps to get the gold medal didn’t automatically include the silver medal, and it was hard (sometimes impossible) to get both at once for certain maps (again, I had to settle with less-than-optimal). There were also ways to just auto-play the stage, and missions that you just set your squads to enter: that just idle/not-actually-playing-the-game seemed dumb to me–it just means you have to check the game more often.

• Arknights
  • You command operators to defend, tower-defense style. There is a sci-fi plot, which I thought was kind of interesting at first, but it moved very slowly (but I didn’t play past the first chapter).
  • Seems closer to an “actually fun” game and less of just a time-waster, and there was more a of a coherent/unique vision (mainly the more interesting strategy aspect). But it became more disorganized as it progress: I would get more operators, and each operator had too many stats that was hard to really reason with/optimize anything (yet, each level was too easy, but also confusingly hard a few times: trying to optimize operator placement turned out to be unnecessary, but I wasn’t sure how to deal with the different enemies).
  • It was annoying to have to manually accept quest rewards and level up operators and skills, the last of which requires you to click on the operator, then their skills, and then the level up button, before you can see what materials you actually need [to find]; that further made it just frustrating to try to optimize how you play. There were also a lot of currencies, half of which I didn’t understand how they worked.

• Mahjong Soul
  • Mahjong, but with anime girls [and men, if you select a different path or something].
  • I didn’t play enough of this to really see what the game play entailed (I don’t know how to play Mahjong, and the tutorials weren’t very good), and gave up trying to figure most of the things out. (I remember playing Mahjong [or something with Mahjong tiles] when I was a lot younger vacationing in China, but I guess I forgot–it seems a lot more complicated than what I remembered.) Everything else was kind of “eh”: it’s just Mahjong, and so the anime girls[/men] couldn’t really be very interesting. (Maybe there was a kind of story mode, but I didn’t see anything like that.)

Aug 16

Basically same as last week [in terms of productivity]; I’m all better now, physically/medically.

I [binge] watched “Moral Orel“: it was pretty good. In summary, it’s about a kid doing bad things because he believes it’s good on religious grounds or the advice of [other] authority figures–he takes it to the logical extreme. I don’t think it’s really a critique of religion, since most of what he concludes is really far-fetched; but it’s also not that funny; but I still liked it: I think the characters were really believable and interesting [which they went into a lot more in the third season (people having existential crises)], and just the juxtaposition of an innocent child doing extremely horrific things is interesting [on a morbid curiosity level]. (One thing it made me think of was that: given that “heaven” is basically infinite goodness and “hell” is infinite badness, there are a “surprisingly” few people who are actively trying to end sinning and whatnot. (Though, I don’t know if that’s how it works; I guess, in [certain denominations of] Christianity, it’s not that “god is sending sinners to hell”, but that “sinners are sending themselves to hell by opposing god”, in which case it doesn’t make sense to try to “save” sinners as much, because they actually “deserve” it.)

I also played a lot of games on my phone, which I usually don’t: I played a lot of those “anime” games (are they called “gacha”s?) I would always see [and some people I know play them], and I decided to try some of them out. Most of them ended up being not that fun [but I played them a lot anyway–just unmotivated to do other things I guess], and it seems like a waste of time. I’m thinking of later posting a review of the games I played; but given that I don’t want to waste more time and don’t really think it’ll be that much of a use to you readers, I put more time into them than I already have. (Except, one of the games seemed genuinely pretty fun, though it still feels kind of time-sucking.)

I keep a daily diary, and I write down a lot of [what I think are] interesting thoughts I have, and I really should put them on here, both as content for you readers and as an organized archive for myself.

Aug 9

I didn’t make any Anki notes or find any interesting articles this week: I had medical thing that came up which interfered with me being productive. I really just want to write something here to keep it consistent.

Aug 2

I wasn’t productive this week, except for one day where I was quite productive [and happy] (Wednesday). On Thursday I fell while biking, and it was annoying/frustrating to deal with: since then, I haven’t been productive [or as happy], and did a lot of “nothing”: browse Reddit and Discord, and play video games (mainly TF2).

I did make some Anki cards, for C++: I did lessons 6.10 to 6.12. I think I was pretty quick with it, and found a good workflow: read + test with Code::Blocks (my IDE), and then go back and Ankify. (I stopped using Toggl: I would actively work “slowly” so I could reach my time goal, but I also felt kind of frustrated with worrying/thinking about it.) I also made some Anki notes for memorizing cello songs. (All of these were made on Wednesday/the day I was productive.)

I found some cool articles and other things. I watched some of the SpaceX Dragon crew splashing down, which was interesting. (Though, I feel “embarrassed” that I don’t care that much about space-stuff, and I don’t know many details and didn’t watch the launch [from a few months ago].)

• Alignment As A Bottleneck To Usefulness Of GPT-3 – LessWrong 2.0: Making sure AI does the things we want (alignment) is seen in the new GPT-3 as examples for the AI to predict on. How easy making good samples will impact GPT-3’s usefulness.
• Nicotine · Gwern.net: Gwern’s article Modafinil (though, which I only skimmed) was also pretty interesting: being contrarian is cool, and also useful as an easy advantage (Slate Star Codex). It seems that tobacco smoke is the main bad part, but just nicotine has less negatives and some positives (but still some negatives); it’s comparable to caffeine.
• The Cowpox of Doubt | Slate Star Codex: That the widespread “hatred” of conspiracy theorists makes us used to truth being a a really obvious thing, when it often isn’t. But also, I find that all the things that I know about the moon landing, global warming, etc., aren’t really “concrete”: I learn them in school and on trustworthy websites, but I haven’t dug into the academic research and studies; and I imagine that most people are similar (though beware typical-mind fallacy), and so in a sense, conspiracy theorists have put more effort than non-conspiracy theorists.
• Debunked And Well-Refuted | Slate Star Codex: On what studies we should trust, or what we should learn from studies: politically biased sites call studies they disagree with “debunked” by a study they do agree with. I feel like I personally should learn more about statistics: how do I interpret studies[/papers in general]?
• Thank you for doing something ambiguously between smoking and not smoking | Slate Star Codex: talking about e-cigarettes, and a concept of “funge”: whether a solution displaces the bad thing (the solution we want) or displaces the good thing (the solution we don’t want). (Or causes good thing and causes bad thing, respectively.) In the case of e-cigarettes, addiction is still bad, but it is also definitely better than regular cigarettes. I found Gwern’s Nicotine essay from here.

Jul 27

I didn’t do any work on any of the things I “wanted” to work on/learn; I didn’t make any Anki notes.

I still found interesting things to look at though. Here are some highlights of my past week’s browsing history:

• Slate Star Codex
  • Meditations On Moloch: about bad systems despite/arising from good/rational individuals
    • This was a very good/important read for me: realizing what the real problems are
  • Nobody Is Perfect, Everything Is Commensurable: how to better contribute to society and not feel guilty for not contributing
    • Reminds me of The Fallacy of Gray – LessWrong 2.0
    • Giving What We Can (was linked in this): part of the charity is a pledge to give 10% of our income, to avoid guilt but still contribute
  • The Toxoplasma Of Rage: angry memes spread better than not angry memes (‘memes’ as information)
    • Reminds me of “This Video Will Make You Angry” by CGP Grey
  • The Categories Were Made For Man, Not Man For The Categories: we create definitions to help ourselves
  • Searching For One-Sided Tradeoffs: you can expect to have an advantage when you either have insider information, can exploit some bias, or have different priorities
  • The Lottery of Fascinations: you can’t choose what you will be interested in
• LessWrong
  • Resist the Happy Death Spiral – LessWrong 2.0: rarely there is just a “Great Thingy” (in his words) that fixes everything; avoid that by splitting that idea into its smaller subcomponents
    • I find that the “Great Thingy” is often an aesthetic thing. I sort of feel like there could be a “religion” around science and rationalism, “Great Thingy”s that do work (though it’s still not straight up magic)
  • The Logical Fallacy of Generalization from Fictional Evidence – LessWrong 2.0: fictional evidence, is well, not evidence. Certain hypothetical will be more story-friendly, even if they aren’t probable. It’s dangerous when these contaminate the way we see–they “do your thinking for you”
  • The ground of optimization – LessWrong 2.0: defines optimizing systems as those that tend toward a certain goal (“configuration set”) within a certain range (a “basin of attraction”). Useful for AI.
  • Causal Universes – LessWrong 2.0 (kind of confusing for me): time travel that doesn’t change past is non-causal; causal is easily computable, non-causal isn’t. Reality seems somewhat tied/related to causality.
  • The Least Convenient Possible World – LessWrong 2.0: when answering hypotheticals/philosophical thought experiments, think of the version of the hypothetical where it is the hardest to answer
    • Commenter davidamann frames it instead as seeing what would need to be changed for their belief to change. It seems better because it’s done by the answerer. Reminds me of Anthony Magnabosco’s videos, where he asks that a lot.
  • Magical Categories – LessWrong 2.0: that values that seem obvious to humans wouldn’t be necessarily so to a superintelligent AI
  • … there are a lot more. I spend a lot of/too much time binge reading LessWrong.

I think I need some sort of process of turning these hours and hours spent reading into something useful: namely, I think I should Ankify them.

I do find the content in reading these really fascinating, and useful (the rationality stuff), so I should memorize them. In addition, maybe the process of having to Ankify them will make me read less out of work-boredom, allowing me to do something different (and hopefully productive).

I think a good process for Ankifying these kind of essays, which are sort of subjective by being, is to be cognizant of the source (the title and author). I would Ankify it after I read it (maybe more than once if I had trouble understanding it) to find out what their central thesis/claim is, and memorize that claim and the reasons for that claim.

I have kind of told myself to “just Ankify everything I see”, but I rarely do that; often I’m kind of tired, but it also seems largely just that I never started. (I used to run in the morning everyday, and it was really kind of simple to tell myself that I needed to do that every morning, despite not really feeling filter and still not really enjoying the running; but right now, I don’t have that habit, so there is a kind of extra mental obstacle that prevents me from “feeling” like running.)

A problem: a lot of essays are really long (e.g., Meditations On Moloch); I suppose I should just be more comfortable with closing out of it to finish reading it later, and more comfortable with having to reread large chunks of it when I’m Ankifying it. I feel like reading these gives me an “emotion” of rationality, of the ability to do better (not just me, but society as a whole); but if I actually want to use these ideas in my thoughts and actions I need to memorize them.

Added Anki Decks

Check it out.

Some are mainly useful for me/specific to me, and some are useful for others as well. It would be nice to receive feedback/Anki note-making advice.

Jul 22

I wasn’t very motivated for most of the week, and spent a lot of time playing video games (mainly Team Fortress 2) or otherwise doing “nothing”. I spent a little bit of time learning C++ (learncpp.com).

These are the Anki notes I have made in the past week:

• CSS:
  • Last summer I had made a bunch of Anki notes for HTML and CSS (from freeCodeCamp), but I suspended them; and now I want to relearn them. I haven’t been relearning/unsuspending them recently, because it’s pretty boring and I’m not that interested (at the moment) in making a site [from scratch].
  • They are about the property float, which pushes the element to a side (either left or right) of the element that contains it.
• C++:
  • They are from 6.8 – 9 of learnncpp.com. (I only have chapters, not the individual lessons, indicated with tags.)
  • 6.8 is about the three ways to implement global symbolic constants: as internal variables, as external variables, and inline variables (this is the best, after C++17).
  • 6.9 is about why [non-constant] global variables are bad.

I think I will make a page sharing all my Anki decks, or maybe just the ones that I have already complete (my old ones). Sharing Anki weekly doesn’t seem that useful: it’ll be a hassle for readers to update their decks every week; instead, maybe I should only share when I complete a deck [or a chapter/other milestone], and each week is just a progress update [on those decks and my life/learning in general] to hold me accountable.

Revising Blog

I had intended to update daily, but after the first day, I found doing that would take a long time. Instead, I decided to update it weekly, every Sunday; but I forgot and delayed.

I also think that the daily blog thing would be very uninteresting: who would read it?, especially with the kind of content I had planned. Instead, I’m thinking of instead sharing the Anki notes I have made during the week, and maybe a sort of summary what those notes encompass. (Or, just a summary of anything I learn, since somethings are harder to Ankify.) At anytime, unrelated to the weekly blog, I’ll share my own thoughts I find interesting (with the intent that you will also find them interesting).

I was mainly thinking of sharing my C++ notes, but I been lately unmotivated to work on learning that, so I’ll share anything in general.