--- title: "Learning Report: 2020, week 37" layout: post date: 2020-09-13 parent: Learning Reports category: Learning Reports published: true --- (NOTE: not yet complete) ## Tutorials * 37 * **todo** * 38 (paths forward) * **todo** * 39 * **todo** ## Other things * Bit more haskell practice / study. Haven't made any significant progress in HFFP * Started reading FoR, and am taking notes. (_notes/fabric-notes.md on github) * I meant to go back to *the choice* but haven't this week * **todo** ### A Slower Speed of Light - WR speedrun * I got [WR (0:2:30) in *A Slower Speed of Light* (SSOL)](https://www.youtube.com/watch?v=rGs7hFcq2jI) [speedrun.com](https://www.speedrun.com/ssol) * First attempt at like a "fast casual" run was 4:04 (better than first recorded time on speedrun.com) * best I found on twitter was like 2:5x * There are ~4 or so unique ways SSOL starts the suggested tweet. I googled them in quotes with various prefixes like "It took me 00:02" * this example finds a quote on Twitter (from @CharlesEnos) from 28 May 2014 > It took me 00:02:47 to break the #threeminuteworldtime http://gamelab.mit.edu/slower #slowlight #openrelativity * I couldn't find shorter times on twitter, there were some other 2 minute times, but none faster than 2:47 * (there were some social-integration features b/c SSOL it was released in 2012) * Watched previous runs to figure out routes * Tried some more experimental routes that didn't seem to work well * Before doing much practice I wanted to get an image of the map for routing purposes * I used uTinyRipper to extract assets from the game file (Unity) * Loaded these assets in Unity 2017 and found the map and figured out (guessing/intuition) * how to load the right map * enough to show the orbs + fences in one image (search for one, ctrl+a; search for the other -> bg items are grey, searched items are visible, selected items are highlighted) [image](https://imgur.com/a/VWpZN5h) * I loaded these into photoshop (initially) to draw the previous routes over them (with timing info for WR) * Photoshop can't tell you the length of a path easily, so I moved to illustrator. The lengths for routes taken were roughly: * 3rd - 2:47 - 6711 * 2nd - 2:40 - 6543 * ~1st - 2:35 - 6385 (note: this isn't my WR run, I measured an earlier run that was 2:35) * Although the game penalises sharp changes in velocity (the character has lots of inertia), seems like shorter routes still matter * I practiced the route in 2:47 first to get a feel for it. Then I practiced the old WR route in 2:40. * practicing these and having the map let me make a few novel improvements: * I don't get the orb that's immediately in front of you when you start the game (-1s) * After the side-run of 13 orbs (bottom middle of my map) I get an extra orb -- the one between the mushrooms and the wall. You have a lot of velocity which you need to shed anyway, so it's quicker (-1s?) * In my WR 2020-09-13 run I think I had better-than-previous technical movement, but there are still obviously sloppy things (like going into the final run of 14 orbs -- I miss one and have to go back) * Doing the final 14 *well* saves >3s on it's own, if not more. Missing any one of them is quite costly * taking it further * I used [*Engauge Digitizer*](https://github.com/markummitchell/engauge-digitizer/) to get a [CSV of points of the orbs](https://imgur.com/a/MoIDmWK) - I wanted to use some route-finding software later to try and improve the route * Idea: use travelling salesman + genetic to find a faster route? * Have done a bit on this in python and haskell (separately) * Fun but not a great use of time, no results from 12hrs+ work * **BOTTLENECK** I think lookup times of pixels I'm using for a-star (which gives distances for the travelling salesman alg) is the bottleneck. looping thru all pixels (~1.3m) to test if black/white adds like +60s to my tests * can optimise this * there are other problems I identified w/ my run and issues I have that I haven't had time to write down here It was maybe ~6hrs till I beat/equalled the WR and ~8 to get a time of 2:30. Most of that time was spent doing the Unity/Route stuff. ## Issues * I had lots of things I wanted to do over the weekend and didn't do ~any. * I intend to read *easy way* again (partially due to some discord conversations about it) * need to fix some habits ## Carry over * Reading habit - I still haven't built this up, or spent time working towards it. * some improvement, not enough * one issue is I end up pausing a thinking a bunch reading FoR which means I take a long time to get through stuff * another is concentration, like doing multiple things at once (e.g. discord) * didn't write as much as I planned to ## post ideas * what would it mean for meta discussion to be bad? is it compatible with fallibilism? * philosophical oracles aren't useful (related to FoR) * postmortem on my short argument against moral relativism * comparison to logical positivism argument in fabric ch1p6 * approval of outcomes is no way to run a democracy * does all progress involve finding new universalities? * maybe just increasing reach of humanity's body of knowledge? * fabric ch1,22:30 * fabric ch1,27:00-27:59,p12 * seems wrong: bit with planets and the claim to only see facts that aren't explained by GR * postmortem: audible reading hard to follow grammatically - easier in print, multiple commas make it difficult. misinterpreted audiobook, clarified after checking print copy. * is 'replication strategy' a good label for the rational/static descriptors on rational/static memes? * finish santa claus is coming to town analysis * how do you represent not-X in CF and CR? * CR perspective: not-X the idea that X is false, which is to say that X is criticised * CF / yesno pespective: not-X ... how is it different, how does it relate to goals? * finish dialog on fi-light * heart stuff * healthypi * how is meta used socially? (like: introducing meta topics) * equal or higher status? * rejection of meta => (optimistically) rejection of social stuff? (maybe like a bit good?) * objections to meta => defensive b/c of social stuff? (not good) * what about some equivalence of action to goodness? does anything like that come in to it? * Justin/Max metaphors of violence chat log * Chat log excerpts * Why you should contribute via subscribestar * Why (and under what conditions) should you take someone seriously? * retrospective on 3 months * learning report for this week + this month * What does it mean for a philosopher to be consistent? (both WRT their behaviour and their values) * "picking fights" - can they be good? * haskell over scheme - importance of data types From FoR notes: * offtopic - FoR ch2,25:00 maybe -- beginning to talk about "tangible" photons and what it means to be in the same universe -> thinking about the "what to do with way out there ideas" post a. maybe this section answers the q "what do you think a universe is?" b. **todo**: retrospective on how my thinking has changed; - gone from - questions like "how do i talk about this" - and behaviours like 'writing a post, not knowing how to deal with it, writing a wrapper post so I can do something at least' - feeling overwhelmed by some stuff and not knowing how to proceed - to (starting to, not mastered) - being capable of knowing what i want to say and whether I can say it, - what I need to do re: organisation of ideas and research, - how to break down ideas to evaluate them earlier in my process, - how to ask questions to get better answers, - better methods of discussion, - better methods of thinking, - appreciation for own static memes and a small arsenal of methods to combat them^1 - knowing what to focus on to communicate ideas better/properly - how to understand what people are trying to say/communicate - how to evaluate texts/samples quicker for qualities and errors - how to manage my own time better and more productively - how to think and feel about myself and my actions/inactions - how to think about my own problems and life problems generally - how to make more reliable progress towards goals - how to choose better goals (and why to choose them) - how to reason about other people and social dynamics / motivations - how to read text in multiple critical frames, including social dynamics - how to choose topics, subjects, and appropriate attention for them 1: i use metaphors of violence, but particularly wrt my own static memes. how does this compare to metaphors of violence wrt criticism?