Right, exactly. All those disclaimers in front of various AI applications have made people forget that in fact, AI cannot fail, it can only be failed.