Post by Andy Zaidman

In response to my earlier post on how we incorporate Gen AI into our Introduction to Programming first year Computer Science course at TU Delft | Electrical Engineering, Mathematics and Computer Science, Thomas Overklift and I wrote a short paper that recently got accepted at the ACM International Conference on the Foundations of Software Engineering - Software Engineering in Education Track (FSE-SEET). The paper: https://lnkd.in/eibUZ6aM Our main take aways from the interventions and discussion in class? 1. You need in-depth knowledge to be able to judge the quality of the GenAI produced code. You need to understand the programming language at least as much as if there would not be any Gen AI. 2. Gen AI can save time, but reviewing the code to understand what the code is doing and scan for potential defects, also takes time. So (a) code reading skills are essential (see the great book by Diomidis Spinellis!) and (b) time gains need to be put into perspective. My previous post on the topic: https://lnkd.in/eKNuM6jK