What is it about?
As AI becomes increasingly involved in writing code, producing a correct program is no longer enough: we also need to understand whether that program will remain efficient as problems grow. This article introduces ADSC — Algorithms, Data Structures, and Complexity — a complexity-first approach to teaching algorithms. It also develops the idea of Scalability Literacy: the ability to anticipate feasibility limits, recognize hidden computational costs, and critically evaluate both human-written and AI-generated code. The approach connects theoretical complexity with practical observations of execution time, memory use, data movement, and implementation choices.
Featured Image
Photo by BoliviaInteligente on Unsplash
Why is it important?
AI tools can generate code that looks polished and produces the correct output while still hiding serious scalability or resource-consumption problems. Students therefore need more than the ability to implement algorithms or name their complexity classes: they need to judge whether a solution will remain viable at scale. By making computational cost explicit and observable, ADSC provides a practical way to develop this form of critical algorithmic judgment and to support more responsible use of AI-assisted programming in computing education.
Perspectives
This work grew out of many years of teaching Algorithms and Data Structures and repeatedly observing the gap between knowing a complexity class and being able to reason about what that complexity means in practice. The rapid emergence of AI-assisted programming made this gap even more important. My aim was to turn complexity analysis from an isolated theoretical topic into a continuous habit of prediction, experimentation, comparison, and algorithmic auditing.
Dr. Valeriu Ungureanu
Moldova State University
Read the Original
This page is a summary of: Teaching Scalability Reasoning through a Complexity-First Approach to Algorithms and Data Structures, Digital Threats Research and Practice, September 2026, ACM (Association for Computing Machinery),
DOI: 10.1145/3849390.
You can read the full text:
Contributors
The following have contributed to this page
