The biology of suffering in intelligent agents is a deep underlying source of existential risk - and one that can potentially be overcome.