Generative and Agentic AI Shift Concern from Technical Debt to Cognitive Debt (simonwillison.net)

A recent article by Margaret-Anne Storey highlights a growing concern within the AI and machine learning community: the rise of "cognitive debt." This term refers to the accumulating mental burden on developers who rush through projects, leading to a breakdown in understanding their systems. While technical debt, like messy code or poor architecture, is often blamed for project challenges, Storey argues that the more pressing issue is cognitive debt—where teams lose clarity on design decisions and system functionalities. This disconnect can severely hinder a team's ability to implement changes or innovate effectively. Storey illustrates this with a poignant anecdote from her experience coaching a student team. As they approached their project deadline, they found themselves unable to make even minor adjustments without causing unforeseen issues. Initially attributing the problems to technical debt, further exploration revealed that the team's lack of shared understanding about the system's design was the real culprit. This shift in focus from technical to cognitive debt emphasizes the importance of fostering a clear mental model among team members, as losing sight of collective knowledge can stymie progress and decision-making in AI-driven projects.