Phone screens, at least in SRE, were chosen from the same pool of questions as the on-sites except for the whiteboard-heavy ones. I imagine practices differed

Phone screens, at least in SRE, were chosen from the same pool of questions as the on-sites except for the whiteboard-heavy ones. I imagine practices differed org to org, but I didn't really see a drop in question quality by the time I left in 2018. Calibration still looked like a fairly meaningful signal but I suppose everything regresses to the mean eventually. Honestly, I had a lot more time and energy back then to devote to algorithms and data structures and knowing Linux and networks inside and out and not sure how I would do now.