So you want to parse a PDF?
eliot-jones.com
Writing a good design document
grantslatton.com
Persona vectors: Monitoring and controlling character traits in language models
anthropic.com
A parser for TypeScript types, written in TypeScript types
github.com
Typed languages are better suited for vibecoding
solmaz.io
Why doctors hate their computers (2018)
newyorker.com
How Python grew from a language to a community
thenewstack.io
Life, Work, Death and the Peasant: Family Formation
acoup.blog
New quantum state of matter found at interface of exotic materials
phys.org
A dedicated skin-to-brain circuit for cool sensation in mice
sciencedaily.com
Human speech may have a universal transmission rate (2019)
science.org
Names are not type safety (2020)
lexi-lambda.github.io
"If you can rack it, you can run UniFi OS" Ubiquiti self-hosted UniFi OS release
deluisio.com
Poorest US workers hit hardest by slowing wage growth
ft.com
Efficiently Generating a Number in a Range (2018)
pcg-random.org
Learnable Programming (2012)
worrydream.com
2,500-year-old Siberian 'ice mummy' had intricate tattoos, imaging reveals
bbc.com
How to grow almost anything
howtogrowalmostanything.notion.site
This Old SGI: notes and memoirs on the Silicon Graphics 4D series (1996)
archive.irixnet.org
Twenty Eighth International Obfuscated C Code Contest
ioccc.org
Show HN: Schematra – Sinatra-inspired minimal web framework for Chicken Scheme
github.com
Tokens are getting more expensive
ethanding.substack.com
Python (or more precisely CPython) uses something like a bitmask and rejection. Alas, there's a bug in the code, so that when you generate a range whose size is a power of two, instead of getting the best case (no rejections), you get half of your values rejected.
For clarity, this worst case for this approach should happen for ranges with size 2*x+1, ie one more than a power of two.
The bug is known but not being fixed right now to keep random number output consistent.