Extra Functional Requirements of Large Code Models

Large Code Models (LCMs) have revolutionized software engineering automation, enabling advanced capabilities in code generation, bug fixing, and program understanding. However, alongside their functional prowess, these models bring forth critical extra-functional requirements that must be addressed for their widespread and responsible adoption. Issues such as scalability, computational efficiency, energy consumption, interpretability, bias, and trustworthiness are becoming increasingly important as these models continue to grow in complexity and impact. This work explores these hidden dimensions, shedding light on the challenges they pose and the opportunities they create for developing more efficient, reliable, and sustainable AI-driven solutions in software engineering. By addressing these extra-functional requirements, we aim to chart a path toward the responsible evolution of LCMs that align with both technical and societal needs.