What are remote friends for? Unanticipated campus closures during COVID19 created uneven student-level exposures to virtual learning challenges, and physically separated campus friends. Using university administrative data, and exogenous class-level differences in pre-pandemic on-campus housing assignments for parallel trend validation, this paper unpacks student-by-course variations in grade expectations using within-semester switches in grade option choice as cues. We find causal evidence that pandemic learning challenges encouraged Satisfactory/Unsatisfactory grade-option uptake among freshmen, but having at least one (remote) friend in class nullified the effect. We explore evidence consistent with performance-improving mutual support despite physical distance between friends as underlying mechanism.