Reproduce “Coding With AI: From a Reflection on Industrial Practices to Future Computer Science and Software Engineering Education” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hung-Fu Chang, MohammadShokrolah Shirazi, Lizhou Cao, Supannika Koolmanojwong Mobasser
This paper investigates how Large Language Models are transforming software development by analyzing industrial practitioner perspectives on new coding paradigms. It clarifies concepts like vibe coding and agentic coding, identifies associated risks, and proposes educational adaptations for computer science and software engineering curricula.
Key insights
- Identified a spectrum of AI usage in programming, ranging from 'AI-assisted coding' (human-augmented) to 'vibe coding' (high-level prompting) and 'agentic coding' (autonomous AI delegation).
- LLM-based coding shifts developer roles from primary 'code authors' to 'specification writers, reviewers, and orchestrators' of AI suggestions.
- The primary bottleneck in software development moves from writing code to critically reviewing and validating AI-generated code.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.