Pretty much regardless of harness, running a coding agent (by general convention) means approving or whitelisting commands one by one - or skipping all that with something like Claude Code’s --dangerously-skip-permissions and living in danger-land.

My phone’s hotspot and WireGuard refused to get along: ping went through, but seemingly nothing else. It bugged me for quite some time before I finally dug in.

Agents get much more useful when they can run real commands; install packages, manipulate data, browse the web. That became clear within days of tinkering on a personal AI assistant (think OpenClaw).

Linux Mint with i3 had served me well, but Xorg was starting to feel dated, especially performance-wise. It felt like time for a change in my setup.

With the switch to Hugo, I wanted a place to test the syntax highlighting that comes with Chroma. This is that place, and we’ll use the classic “Hello, World!” example in various computer languages I have used in one way or another, past or present.