In this video I’m basically reacting to the usual “Linux audio stack” pain after watching Luke Smith talk about PulseAudio and the headaches around trying to run ALSA by itself. On Linux you start at ALSA, but if you actually want normal modern stuff—multiple audio sources at once, easy desktop capture, routing outputs as inputs—you quickly end up stacking PulseAudio or PipeWire on top. That layering works, but it adds complexity, and complexity means more places for things to break. Then I contrast that with OpenBSD (and BSDs that use it) where sndio is just… refreshingly simple. With sndio/sndiod + sndioctl, I can mix devices, monitor outputs, and treat an output like an input for recording—without building a monster stack. The documentation is actually good, the man pages tell you what you need, and I can script basically everything (volume keys, per-program behaviors, automation) with a couple simple commands. I’m not saying “switch to BSD for pro audio production,” because OpenBSD isn’t exactly aimed at audio engineering, but for everyday audio reliability sndio has been way more consistent for me than the typical Linux PulseAudio/PipeWire situation.