Rei: "You're Haru-chan-san?"
Haru: "Don't put -chan and -san on my name"
The last chapter of Miki Yoshikawa's "Yamada-kun and the seven witches" #manga (chapter 243!, she published it weekly over five years!) was directly responsible for the flurry of job applications I sent to investment banks that eventually resulted in me accepting a job at one.
I wanted to make three times as much money as Yamada, just like Shiraishi, his girlfriend and the first eponymous witch 😝🤣.
(Not a recommendation, it's terrifically long, but example of manga changing lives)
Modeling the relationship between the various kinds of quizzes allowed in #Curtiz as nodes and edges in a densely-connected directed graph: reviewing which strengthens the memory of which others? Paper and pen and 9x9 adjacency matrix.
The relevant article I was reading: "Sum types are coming" (from 2015):
There's ONE hit on the first DuckDuckGo page for Python sum types on Reddit, whose top reply misses the point and implements the example using… classes.
Also, from the same blogger, this handy chart of semi?-advanced features in Java, Go, Python, Haskell, et al. and the "sum types" line highlighted, also from 2015:
Sum types are essential.
You know how some spouses are handy? With like, plumbing, or gardening, or tech, or taxes?
Yeah I'm none of those.
Just print to PDF to see the puzzle all nice (CSS media selectors omit puzzle config inputs). No Svelte, no React, no Webpack, just raw DOM, TypeScript, and Browserify.
Letter frequency to match English. Let me know if you want others.
I dealt out the following board:
🂫🃞🃎 🃕 🃓
After just the flop (first three cards), there's roughly 40% chance your hand ends up as a pair, 40% a two-pair, 10% a trip, 10% a full house, and epsilon the other results (±3%-ish).
Usually the probabilities aren't that symmetric, so I was moved to deal out ten pockets (pairs of personal cards). Not four but SEVEN two-pairs!, and, surprise!, a flush too!
Was surprised to see flush at <1% odds after the flop, since club-suited pockets are ~5% right?
The Global Historical Climate Network publishes annual gzip files of all the data collected by all its member stations each year. I made an interactive plot of the size of each zip file:
- the post-WW1/Spanish Influenza dip
- steady growth through WW2
- the crash of 1971 (two years before the Oil Shock), and the plateau till the late '90s
- the 2000–2010 spike
- downturn since
Yeah unzipped size but.
As a proxy for global scientific and economic output…
Still reeling from how amazing the first two episodes were. #Sarazanmai
Perceptions of probability, cf.,
- Sherman Kent CIA study https://www.cia.gov/library/center-for-the-study-of-intelligence/csi-publications/books-and-monographs/psychology-of-intelligence-analysis/art15.html and
- Philip Tetlock's *Superforecasting*.
Ridge plot in Bokeh (#Python plotting library).
We were hearing that students living in dormitories are hacking their crock pots to produce high enough temperatures for hot pot, which gave my brilliant spouse this idea:
#なべ in fondue pot! Works perfectly!!!
- Boy and the beast バケモノの子
- Summer wars サマーウォーズ
- Your name… 君の名は。
They were next to the Japanese learning materials for English speakers (textbooks, readers, &c.), which is absolutely brilliant product placement. And they were all quite inexpensive (¥500-700 listed on the back, $9-11 in the store).
<rolls up sleeves>
Oh how charming, the American stock market might be crashing: S&P500 5% off its recent high. I'm hearing where it might fall through 50%, in fits and starts of course, before it thinks it's low enough.
Remember in the mid-2000s, we were all worked up about Peak Oil and energy stocks and national oil companies like Saudi Aramco et al.? What happened to that? Oh right, the global economy hit the brakes and we stopped burning and pumping so much oil.
This could get fun.