The Impact of AI on Developer Productivity: Evidence from GitHub Copilot
Controlled experiment showing measurable productivity gains from AI pair programming.
Scholar
Open-access research for vibe coding
These papers are highly relevant to vibe coding (AI-assisted software development, prompt design, and code quality). Each item links to Google Scholar and a free full-text source.
The Impact of Prompt Programming on Function-Level Code Generation
Evaluates prompt techniques for code correctness, similarity, and quality.
Experience with GitHub Copilot for Developer Productivity at Zoominfo
Industry report on practical adoption and outcomes using Copilot at scale.
Large Language Models for Code Generation: The Practitioners Perspective
Practitioner-focused evaluation of code generation models in real workflows.
Combining Large Language Models with Static Analyzers for Code Review Generation
Hybrid approach that combines LLM reasoning with static analysis signals.
Evaluating Large Language Models Trained on Code
Foundational benchmark paper for modern code-focused language models.