$ sort -k featured ~/articles
Hand-picked deep dives the team recommends starting with.
Self-hosted email is losing to Gmail's hard SMTP rejections. Self-hosted password managers are split down the middle. And self-hosted music is doing the opposite of what you'd expect.
Part 1 ended on a 70-gigabyte breach in 2021. Since then, hacktivists derailed trains, dumped molten steel on camera, and reached grids in three countries.
Every entity, every block update, every redstone pulse — one thread, twenty times a second, 50 milliseconds a pass. Folia's answer is to stop sharing the kitchen. Here is the tick budget and the ownership rules, read from the PaperMC source.
Forum Archaeology #1: I built plugins for this software when it ran half the internet's communities. Sixteen years after the exodus that should have killed it, vBulletin is still getting hit with critical, unauthenticated remote code execution.
Ghost Check #1: the Puppet Master hacked minds by exploiting the gap between what a system experiences and what it can verify. Thirty-one years later, a researcher did the same thing to Microsoft 365 Copilot with a single email.
JDK 27 ships nine JEPs, but only a few actually change your code today. What's default now, what's still preview, and what to do about each one now.
Self-host OpenClaw from zero to production. Server provisioning, SSH lockdown, Docker, reverse proxy, Fail2ban, performance tuning, and full backup strategy in one guide.
Build a production-grade self-hosted AI agent from scratch using Docker Compose. Covers the full stack — Ollama, n8n, Flowise, Qdrant, SearXNG, Langfuse, and Redis.
Replace Google Drive, Google Photos, 1Password, Plex, and your document scanner with a self-hosted stack that runs on your hardware. Complete Docker Compose setup with Traefik, Nextcloud, Vaultwarden, Immich, Jellyfin, and Paperless-ngx.
Can't keep up! is not a hardware message. It's the server telling you a tick took longer than 50ms, and almost every cause is a default nobody changed. Here's what actually moves MSPT, with the flags that are now wrong and the setting that isn't where every guide says it is.
A local LLM is three layers: weights on disk, a runtime that loads them, an interface that talks to it. Once you see the stack that way, the hardware question answers itself and the tuning stops being guesswork.
For three years the AI industry watched GPUs. Nobody was watching the CPU. Now Intel is warning of six-month lead times, AMD has blown past expectations, and the chip everyone assumed would always be available has become the new bottleneck in the biggest infrastructure buildout since the internet.
