Two leading AI figures painted candid pictures of how agentic models are reshaping technical work. Andrej Karpathy said agentic AI has shifted his own role away from hands‑on coding and described a personal sense of disorientation. Geoffrey Hinton warned that many tech firms focus on short‑term profit rather than long‑term societal consequences. Both accounts were given in recent interviews and essays and reflect tensions between rapid commercial deployment and academic research norms. For universities, the messages underscore pressure on curricula, doctoral training and research ethics programs; expect students and faculty to seek clearer guidance on responsible agent development and career pathways outside academia.