- 341Tailscale didn't stop the Hugging Face intrusion (tailscale.com)
- 790Elevators (john.fun)
- 373qm (github.com)
- 80Twenty-five years ago it was cryptography, today it's model weights (weeraman.com)
- 51Progressive Web Components (arielsalminen.com)
- 85June in Servo: real world compat, media queries, SharedWorker, and more (servo.org)
- 21Demystifying DRAM Read Disturbance: RowHammer and RowPress Phenomena (arxiv.org)
- 173Big Food vs. the People (lighthousereports.com)
- 15The Absurdity of Albert Camus (historytoday.com)
- 104Golang proposal: container/: generic collection types (github.com)
- Loops (YC W22) Is Hiring a Product Educator (ycombinator.com)
job - 510DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis (artificialanalysis.ai)
- 48Let's make the worst Htmx (zserge.com)
- 43Termixer (TUI DJ Mixer) (github.com)
- 122Run Kimi K3 using 29 GB of RAM at 0.50 tok/s (github.com)
- 45Increasing the lifespan of a bulb makes it worse in every other way (maurycyz.com)
- 9The First Transatlantic Telegraph Cable Was a Bold, Beautiful Failure (spectrum.ieee.org)
- 37Authorize, don't authenticate (blog.marcua.net)
- 117The most official water costs $120k a gallon (signoregalilei.com)
- 114Getting 25 Gbps Thunderbolt Ethernet on My Mac Studio (jeffgeerling.com)
- 98Dubious research tied to Red Bull has shaped energy drink policy (theexamination.org)
- 7Britain's New World of Tobacco (2017) (historytoday.com)
- 84How JPEG works: Interactively explore JPEG's lossy compression methods (cgjennings.ca)
- 15Predictive Speculative KV Replication for Bursty LLM Inference (jwlabs.vercel.app)
- 76Everyone is building LLM routers, we deprecated ours (manifest.build)
- 90Algorithms on billion-scale graph using 10GB RAM: I love DataFusion (semyonsinchenko.github.io)
- 40Using the railway network as a flatbed scanner [video] (media.ccc.de)
- 102Is AI reasoning right for the wrong reasons? (quantamagazine.org)
- 181Severance (lcamtuf.substack.com)
- 4Hope and Defeat: John Berger (newleftreview.org)