Daniel Murfet comments on Prospects for Alignment Automation: Interpretability Case Study