Impact of AI Coding Assistants on the SW development process: Benchmarking Efficiency generated code vs. human programmers

This book explores the role of Artificial Intelligence-Based Coding Assistants, with a focus on Gemini, in software development, evaluating their impact on code efficiency and quality compared to human programmers. Through a series of programming challenges of varying complexity and using the OmegaUp evaluation platform, the performance of Gemini and human programmers is analyzed using quantitative metrics such as memory usage, test cases solved, and processing time. The findings reveal that while Gemini achieves similar performance to humans on low-complexity tasks, programmers significantly outperform the assistant on more advanced problems. This highlights Gemini's limitations in complex contexts, emphasizing that its true value lies in facilitating less demanding tasks, such as generating documentation and explaining code. This is essential reading for understanding how AI tools can complement, but not yet replace, human ingenuity in software development.