一句话：在10周入门编程课中三次引入生成式AI，发现其解法含细微错误，说明仍需教授深入编程知识。
本文关注大学计算机科学一年级入门编程课程如何应对生成式AI（GenAI）的影响。作者在10周课程开始、中期和结束阶段设计并实施三次教学干预，试图让持批判态度的学生认识到：深入理解编程仍然重要，并非所有任务都能靠“凭感觉编程”（vibe coding）完成。通过观察学生与GenAI的互动，作者发现学生对GenAI解法中存在的细微错误感到意外。对教师而言，这坚定了继续教授深入编程知识与理解的观点。作为扩展摘要，文章主要报告早期教学干预经验与反思，而非大规模量化评估。
This paper introduces and discusses an early intervention of Generative AI (GenAI) in a first year university-level introductory programming course of a Computer Science programme. In our quest to convince our critical student audience that deep knowledge and understanding of programming is still relevant, and not everything can be vibe coded, we designed 3 interventions that we deployed at the start, around the middle, and towards the end of the 10-week course. From interactions with students, we observed that they were surprised at some sublte mistakes that GenAI solutions contained. For us teachers, it solidified the idea that we should keep teaching students in-depth programming knowledge and understanding.