Insights and hands-on guides on frontend, backend, databases, IT infrastructure, and AI solutions.
Running an LLM on CPU Only - Real Numbers from an Old Xeon with 256GB RAM
Our Servers Catch Burglars Every Day - TREEWALL Field Test Report
NVMe Storage Benchmark: Optane to HDD - 20 Devices Compared with fio
Local AI Video Generation with ComfyUI + Wan 2.2 - An RTX PRO 6000 96GB Build Log
Structuring Unstructured Documents with an LLM - Measured Accuracy on a 19-Item Menu
How we reduced a Python automation tool's startup from 78s to 20s by adding timestamps to logs, cutting API calls from 44 to 5, and removing dead code waits.
After 1 year of smooth Selenium Naver logins, CAPTCHAs appeared. Dozens of tests and 7 failed fixes later — the real culprit was the $cdc_ variable injected by chromedriver.
Attempting to run undetected-chromedriver in a subprocess and hand the Chrome session to the main process via debug port. It seemed clean in theory but failed for 4 fundamental reasons.
Python benchmark tool tested 8 CPUs across 3 embedding models. A 16-core desktop leads at 847.3 sent/s, a 4-core mini PC at 123.7. Includes remote SSH execution and per-watt efficiency ranking.
Guarding the Front Door of Your Network - TREEWALL, an AI Security Gateway
Build a CCTV Recording Server Without an NVR - RTSP + ffmpeg + systemd, 180-Day Retention
Our MoE Model Started Spewing Garbage - What Auditing All 18,432 Expert Weights Revealed
What We Learned Building and Auditing 990 AI-Generated English Learning Scenes
BGE-M3 vs mE5-Large - Comparing Korean RAG Search Quality Head to Head
