This paper explores the impact of informal Large Language Models (LLMs) usage on undergraduate students' learning outcomes in software development education, focusing on React applications. The study involves 32 participants over ten weeks, examining the correlation between LLM usage and final grades. Key findings include: 1. **Overall Impact**: A significant negative correlation was found between increased LLM reliance and final grades, suggesting that higher engagement with LLMs may detract from academic performance. 2. **Specific Uses**: - **Code Generation**: A significant negative correlation was observed between LLM usage for code generation and final grades, indicating that overreliance on LLMs for critical tasks like code generation can undermine independent coding skills. - **Debugging**: A significant negative correlation was found between LLM usage for debugging and final grades, highlighting the importance of fostering independent debugging skills. - **Seeking Additional Explanations**: A non-significant correlation was observed, suggesting that using LLMs for seeking additional explanations does not significantly impact final grades, potentially serving as a supplementary learning tool. 3. **Educational Implications**: The study emphasizes the need for a balanced approach to integrating LLMs into programming education. While LLMs can enhance learning through supplementary explanations, their role in critical tasks like code generation and debugging appears to negatively influence student outcomes. 4. **Future Directions**: Further research is needed to explore the specific dynamics through which LLMs influence learning processes and outcomes, and to refine strategies for integrating these technologies into educational frameworks. The study underscores the importance of a thoughtful approach to incorporating LLMs in programming education, ensuring they support rather than undermine the development of core programming skills.This paper explores the impact of informal Large Language Models (LLMs) usage on undergraduate students' learning outcomes in software development education, focusing on React applications. The study involves 32 participants over ten weeks, examining the correlation between LLM usage and final grades. Key findings include: 1. **Overall Impact**: A significant negative correlation was found between increased LLM reliance and final grades, suggesting that higher engagement with LLMs may detract from academic performance. 2. **Specific Uses**: - **Code Generation**: A significant negative correlation was observed between LLM usage for code generation and final grades, indicating that overreliance on LLMs for critical tasks like code generation can undermine independent coding skills. - **Debugging**: A significant negative correlation was found between LLM usage for debugging and final grades, highlighting the importance of fostering independent debugging skills. - **Seeking Additional Explanations**: A non-significant correlation was observed, suggesting that using LLMs for seeking additional explanations does not significantly impact final grades, potentially serving as a supplementary learning tool. 3. **Educational Implications**: The study emphasizes the need for a balanced approach to integrating LLMs into programming education. While LLMs can enhance learning through supplementary explanations, their role in critical tasks like code generation and debugging appears to negatively influence student outcomes. 4. **Future Directions**: Further research is needed to explore the specific dynamics through which LLMs influence learning processes and outcomes, and to refine strategies for integrating these technologies into educational frameworks. The study underscores the importance of a thoughtful approach to incorporating LLMs in programming education, ensuring they support rather than undermine the development of core programming skills.