Ishan Mukherjee comments on When is AI safety research harmful?