
Article URL: https://github.com/microsoft/BitNet
Comments URL: https://news.ycombinator.com/item?id=41877609
Points: 105
# Comments: 25
Article URL: https://technicalwriting.dev/strategy/decisions.html
Comments URL: https://news.ycombinator.com/item?id=41881872
Points: 16
# Comments: 9


Article URL: https://subvert.fm/
Comments URL: https://news.ycombinator.com/item?id=41880829
Points: 60
# Comments: 23


Article URL: https://www.ycombinator.com/companies/distro/jobs/XGMjSPe-tech-lead
Comments URL: https://news.ycombinator.com/item?id=41881304
Points: 0
# Comments: 0
https://www.ycombinator.com/companies/distro/jobs/XGMjSPe-tech-lead
As a fan of dense New York Times-style crosswords, I challenged myself to create topic-specific puzzles. It turns out that generating crosswords and efficiently placing words is a non-trivial computational problem.
I started the project, "Joystick Jargon" combining traditional crossword elements with gaming-related vocabulary. Here's the technical process behind it:
1. Data Source: Used a 3.8 Million Rows Reddit dataset from Hugging Face (https://huggingface.co/datasets/webis/tldr-17).
A quick dive into the Plan9 assembly I picked up while developing my SIMD package for Go, and how it led to a 450% performance boost in calculations.
Comments URL: https://news.ycombinator.com/item?id=41879854
Points: 93
# Comments: 33

Article URL: https://spliit.app/blog/spliit-by-the-stats-usage-costs-donations
Comments URL: https://news.ycombinator.com/item?id=41879845
Points: 69
# Comments: 39
https://spliit.app/blog/spliit-by-the-stats-usage-costs-donations