Even if you bought with Linux in mind, there’s no guarantee what works today works tomorrow. I have a Linux box that had working HDMI audio. A recent set of updates to the audio subsystems now means that when the box boots, the HDMI devices are muted by default. It seems like something changed in the order that things are initialized and the HDMI devices aren’t present when the audio system is initialized so none of your saved settings are re-applied. Every boot now requires manually fiddling with the audio settings to reset where audio should be going and unmute the devices. If my choice when going into an interview is the box that you need to buy specific hardware for and hope that no one re-configured the audio subsystems in the last update or the box that runs an os by a company known to put explicit code paths to recreate a prior edition bug that other software relies on… I might choose the latter too, even if it is windows. Of course I’d also probably choose the latter because there’s a non zero chance the interviewer is going to want to use some conferencing software or website that doesn’t work in Linux regardless of how well matched my hardware is.

I simply don't use a headset or a webcam with my home Linux setup, because it's often been a hassle to deal with, and my work setup is separate. For screen-sharing type coding interviews, I have a perfectly functional remote SSH setup from my windows laptop, which took less time to setup than the last time I fiddled with a bad headset connection on Linux a long time ago. I find the "not using all of screen real estate" accusation to be strange as well, because usually screen-sharing is ergonomically limited to 1 screen, but people may usually work with two screens and not have developed habits to make 1 screen work.