Show HN: Web app that lets you send email time capsules
resurf.me
GitHub Actions Has a Package Manager, and It Might Be the Worst
nesbitt.io
Jujutsu Worktrees Are Convenient
shaddy.dev
Nango (YC W23) is hiring back-end engineers and dev-rels (remote)
jobs.ashbyhq.com
Emacs is my new window manager
howardism.org
I failed to recreate the 1996 Space Jam website with Claude
j0nah.com
Show HN: Lockenv – Simple encrypted secrets storage for Git
github.com
Bag of words, have mercy on us
experimental-history.com
Client-side GPU load balancing with Redis and Lua
galileo.ai
Dollar-stores overcharge customers while promising low prices
theguardian.com
Google Titans architecture, helping AI have long-term memory
research.google
Show HN: ReadyKit – Superfast SaaS Starter with Multi-Tenant Workspaces
readykit.dev
The fuck off contact page
nicchan.me
The C++ standard for the F-35 Fighter Jet [video]
youtube.com
Mechanical power generation using Earth's ambient radiation
science.org
I wasted years of my life in crypto
twitter.com
An Interactive Guide to the Fourier Transform
betterexplained.com
Solving Rush Hour, the Puzzle (2018)
michaelfogleman.com
CATL expects oceanic electric ships in 3 years
cleantechnica.com
The Anatomy of a macOS App
eclecticlight.co
Author here. We were hitting tail latency and low GPU utilization issues serving SLMs via Triton.
I built a scrappy client-side router using Redis and Lua to track real-time GPU load. It boosted utilization by ~40% and improved latencies.
Happy to hear feedback on the implementation or thoughts on better ways to do this!