Posts
Page 57 of 80
-
Cloudflare DNS: Beyond Pointing Records
Stop using your registrar's janky DNS panel. Here's how Cloudflare DNS actually works, proxying, DNSSEC, dynamic DNS, and email records that don't break.
8 min read -
LiteLLM & vLLM: One API to Rule All Your Models
LiteLLM proxies every LLM, local or cloud, behind one OpenAI-compatible endpoint. Pair it with vLLM for GPU-backed serving and ditch the SDK sprawl.
· Updated:9 min read -
System Prompts: The LLM Feature Most People Ignore
System prompts are your secret weapon. How they work, why they matter more than you think, and 5 patterns that actually change model behavior.
5 min read -
Systemd Socket Activation: Start Services Only When Someone Actually Knocks
Learn systemd socket activation to start services on-demand, save RAM, and cut boot time. Includes .socket unit files, real examples, and testing with systemd-socket-activate.
· Updated:8 min read -
make for Project Automation (It's Not Just for C Code)
Forget bash scripts scattered across your repo. make is a simple task runner that's been around for 50 years and works everywhere.
6 min read -
Vault vs Infisical: Secrets Management for Teams Who've Learned the Hard Way
HashiCorp Vault vs Infisical compared: secrets management for DevOps teams, Docker Compose setup, SDK examples, and when the added complexity is worth it.
· Updated:10 min read -
Git Hooks You Should Be Using Locally Right Now
Stop committing broken code. Git hooks catch mistakes before they hit CI, save hours of debugging, and make your team love you.
5 min read -
Traefik: Docker Routing with Labels
Master Traefik's label-based routing in Docker: entrypoints, routers, middlewares, TLS, and the mental model that makes it all click.
· Updated:8 min read -
LLM Quantization: Q4_K_M Isn't Always the Best Choice
Q4_K_M is the default, but it's not magic. When Q3, Q5, or Q6 makes sense. How to benchmark quantization tradeoffs on your hardware.
5 min read -
Docker BuildKit: Stop Waiting for Your Images to Build
Your CI rebuilds node_modules from scratch every build. BuildKit, Docker's default builder since 23.0, fixes it with cache mounts, parallel stages, real secrets.
· Updated:10 min read -
Running Multiple Ollama Models Without Running Out of RAM
Ollama can load one model at a time on limited hardware. How to switch between models, use CPU offloading, and manage VRAM intelligently.
5 min read -
Piper vs Coqui: Text-to-Speech on Your Own Hardware (Because AWS Polly Charges Per Character Like It's 1999 SMS)
Piper vs Coqui TTS compared: speed, voice quality, Docker setup, and Home Assistant integration. Run offline neural TTS on your own hardware, no cloud fees.
· Updated:11 min read
