Weekly Best
241.
Memorilla: Latent Semantic Memory for LLMs
(arxiv.org)
242.
Evergreen: A Provably Safe Systems Language
(evergreenlang.net)
243.
GitHub Down Again?
(github.com)
244.
245.
247.
Curiosity Traps
(autodidacts.io)
248.
We can do better than line-buffered output
(nullprogram.com)
249.
SLAC researchers find proof of altermagnetism, a new magnetic phase
(www6.slac.stanford.edu)
250.
251.
252.
253.
Agent memory needs its OpenTelemetry moment
(lorekit.io)
256.
The Flash C Compiler
(blog.adafruit.com)
257.
Agentic backdoors via pretraining poisoning
(arxiv.org)
258.
260.
Microsoft fixes a Windows 11 annoyance as PC revenue declines
(arstechnica.com)
261.
TextEdit vs. Pages
(leancrew.com)
262.
Petered
(aworkinglibrary.com)
263.
Claude Opus 5.5 Can Draw Permit-Level Architectural Drawings
(stephenjgallagher.substack.com)
264.
265.
Zero Hour: Telling time was a matter of survival and strategy during WWI (2014)
(laphamsquarterly.org)
266.
Partial Operator Precedence
(ziglang.org)
267.
269.
270.
AI for scientific discovery: How researchers use Claude
(anthropic.com)