I think it's becoming more and more clear that the "AI job skills" for the future will not be focused on using AI as an assistant, but in understanding and mitigating the impacts of AI as an adversary.