Raw LLM Responses

I agree with the basic premise that students are now relying on AI coding to do project work and it means they aren't really learning practical coding. I would say that asking domain/implementation specific examples isn't fair to determine whether someone understands the principles of software design and coding in and of themselves. Say someone did CS with a C++ focus meaning a strictly-typed non-garbage-collected polymorphic object-oriented language. That checks a lot of the boxes for rigorous coding techniques, but they couldn't answer a number of questions you asked here (immediately.) As long as you are aware of other languages and implementations (E.G. functional programming, relational DBs say) that you ACTUALLY DO YOUR OWN CODING WITHOUT RELYING ON A BOT and know how to research specific implementations that is the important thing. Ultimately software engineering is about math, and specifically optimization because there are millions of ways to reach a result but in SE you have to consider constraints, time, memory, disk space and so on. Software patterns and defensive coding are an abstracted layer upon this and you don't see this a lot in CS education but for curious students they will investigate this themselves. But ultimately you could have someone just understand Excel macros and a bit of VBA and be a fabulous fit for a data science job, not knowing the first thing about python implementation. It's all about mathematic foundations, diligence and research to succeed, ultimately. I think vibe coding will soon go out of fashion.