Running Ollama on iGPU on Linux
How I got 3x inference speed on a five-year-old Intel NUC iGPU

A working log of testing, tinkering, occasional detours and everything I broke trying.
How I got 3x inference speed on a five-year-old Intel NUC iGPU

Two mesh VPNs, one Viettel CGNAT line, and the four measurements that moved my homelab off Headscale.

A @reboot cron never disabled my laptop touchscreen. Finding the layer that does.

How Meta Quest headsets are setup for classroom use in the SIMPLE Project.

When IPv6 is the future, but you can't connect to it

How a filename on Windows broke my full packaged application
