If you've conducted engineering interviews recently, you might have noticed a disturbing trend: an increasing number of new college graduates are fumbling on the easiest Data Structures and Algorithms (DS&A) questions. We're not talking about balancing a red-black tree; we're talking about basic array manipulation and string traversal—the warm-ups.
The Illusion of Competence
In the past, these questions were a baseline check. Today, candidates often approach them with a strange blankness. They understand the high-level concepts of architectures and APIs, but when asked to write a simple for loop with a pointer, the logic stutters. They are waiting for the autocomplete that isn't coming.
This correlates directly with the widespread, often unchecked use of generative AI in college classrooms over the past few years. When a student can prompt an AI to write their introductory programming assignments, they never develop the fundamental mental models required to visualize code execution in their head.
Destroying Critical Thinking
Recent research indicates that over-reliance on AI doesn't just make people lazy; it fundamentally alters the brain's ability to engage in sustained, critical problem-solving. By offloading the "struggle" of basic logic to a machine, students bypass the vital cognitive friction required for learning string manipulation, memory allocation, and algorithmic complexity.
We are essentially graduating a generation of 'prompt engineers' applying for software engineering roles. They can wire together high-level components with the help of an LLM, but underneath that shiny abstraction layer, the procedural foundation is dangerously brittle.