April 15, 2024Open Access

Can ChatGPT Support Developers? An Empirical Evaluation of Large Language Models for Code Generation

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Empirical analysis shows limitations of LLM-generated code in practical software development, suggesting improvement is needed.

LLM-generated code typically provides examples in documentation rather than production-ready code, indicating a need for further advancement.
Empirical analysis reveals that current practices in using LLMs for code generation remain limited in applicability.
Conversations with developers using ChatGPT highlight the gap between promising capabilities and real-world effectiveness of LLMs in code support roles and usage scenarios.

Jin et al. (2024) studied this question.

View Full Paper

Ask AI

Bookmark