AI Coding Transforms Software Development, Insights from 57 Industrial Practitioner Videos

The rapid evolution of software development driven by large language models presents both opportunities and challenges for computing professionals and educators, and a new study investigates how these tools are reshaping industrial practice. Hung-Fu Chang from the University of Indianapolis, Mohammad Shokrolah Shirazi, Lizhou Cao from the University of Maryland Eastern Shore, and Supannika Koolmanojwong Mobasser from The Boehm Center for Systems and Software Engineering analysed reflections from software practitioners shared in online videos to understand current usage, associated risks, and changes to established workflows. The research reveals significant productivity gains and reduced barriers to entry with these new coding methods, but also highlights emerging concerns around code quality, security, and the potential erosion of fundamental programming skills. This industry-focused analysis provides crucial insights for adapting computer science and software engineering curricula to prepare future engineers for a rapidly changing professional landscape, advocating for greater emphasis on problem-solving, architectural design, and rigorous code review practices.

Vibe Coding, Themes and Critical Perspectives

A comprehensive analysis of YouTube videos reveals diverse perspectives on Vibe Coding, a rapidly evolving approach to software development. These videos demonstrate a wide range of opinions, falling broadly into pro-Vibe Coding, critical, or nuanced categories, with additional content focused on explanation, the future of the practice, and security considerations. The videos highlight core themes including speed, maintainability. Many practitioners emphasize the potential benefits of Vibe Coding, while others express concerns about its long-term viability and the potential for technical debt. Balanced perspectives acknowledge both the advantages and risks, focusing on best practices for secure and effective implementation. A significant number of videos explore the future of Vibe Coding in relation to artificial intelligence, agents, and broader shifts in software development. These discussions often center on the potential for AI to automate coding tasks, lower barriers to entry, and accelerate innovation, while also raising questions about the evolving role of developers and the need for new skills.

YouTube Videos Reveal AI Coding Practices

Researchers conducted a novel study of AI-assisted coding by analyzing content shared by professional developers on YouTube. They curated a collection of 57 videos published between late 2024 and 2025, focusing on practitioner experiences with large language models (LLMs) in real-world software development. This approach addresses a gap in existing research, which often concentrates on educational applications of LLM coding tools. The team employed a rigorous qualitative analysis process, carefully selecting videos that accurately represented current industrial practices. This detailed examination revealed definitions of both ‘vibe coding’ and ‘agentic coding’ practices, distinguishing them from conventional code generation techniques.

Industry reports indicate that a substantial portion of code in some organizations is now generated using these technologies, demonstrating significant productivity gains and lowered barriers to entry. The analysis highlights a shift in development bottlenecks towards code review and concerns surrounding code quality, security, and the potential erosion of foundational programming skills. The study provides valuable insights into AI-assisted coding and offers guidance for adapting computer science and software engineering education to meet the demands of this rapidly changing professional environment.

LLMs Reshape Professional Software Development Workflows

This research details how large language models (LLMs) are impacting professional software development, based on a qualitative study of 57 videos featuring experienced practitioners. The work demonstrates a shift towards AI-supported coding approaches, including vibe coding, AI-assisted coding, and agentic coding, all accelerating code generation and changing development challenges. The study reveals that development bottlenecks are moving away from initial code writing and towards areas like code review, testing, and ensuring system-level security. This highlights the growing importance of skills focused on evaluating and validating AI-generated code.

Practitioners describe “vibe coding” as a focus on high-level instructions and functional output, emphasizing orchestration over direct coding. “Agentic coding” involves delegating coding tasks to autonomous AI agents, simulating collaboration with multiple engineers. While AI tools accelerate prototyping and experimentation, concerns are rising regarding code quality, maintainability, security vulnerabilities, and the potential erosion of foundational problem-solving skills.

AI Reshapes Software Development Practices

This research presents a systematic analysis of how large language models are reshaping software development, based on detailed examination of industry practitioners’ experiences shared in online videos. The study identifies a shift towards various AI-supported coding approaches, demonstrating accelerated code generation and a corresponding change in development challenges. Bottlenecks are moving away from initial code writing and towards areas like code review, testing, and ensuring system-level security, highlighting the growing importance of skills focused on evaluating and validating AI-generated code. The findings also reveal a transformation in developer roles and skill requirements, with AI tools lowering barriers to entry but simultaneously raising concerns about potential skill erosion. Consequently, the research suggests a need for adaptation within computer science education, advocating for a greater emphasis on problem-solving, architectural thinking, and specification-driven development, alongside increased use of project-based learning that incorporates AI tools and real-world workflows. This work represents an early snapshot of a rapidly evolving field, and future research should focus on longitudinal studies to better understand the long-term co-evolution of AI-based development practices.