Cognitive debt

Overview

Cognitive debt is a product and engineering concept that describes the hidden cost of AI-accelerated development when teams lose shared understanding of how a system works. Instead of debt showing up mainly as messy code or architectural shortcuts, the burden shifts into team cognition: fewer people understand the implementation, design decisions become harder to trace, and the system starts to feel like a black box.

For AI Product Managers, this matters because velocity can become misleading. Agent-assisted shipping may increase output in the short term while quietly reducing the team's ability to debug, extend, govern, and safely evolve the product over time. Cognitive debt is therefore a useful framing for evaluating not just whether AI helps teams build faster, but whether it preserves enough system comprehension to sustain quality, accountability, and iteration.

Key Developments

• 2026-02-15 — Simon Willison highlighted Margaret-Anne Storey's explanation of cognitive debt as the shift of accumulated cost from traditional technical debt into developers' shared understanding when AI accelerates software creation.
• 2026-03-01 — In discussion of interactive explanations, Simon Willison connected cognitive debt to agent-written code becoming a black box, arguing that interactive walkthroughs can help teams retain implementation understanding and reduce long-term friction.

Relevance to AI PMs

• Assess delivery beyond speed. AI PMs should evaluate whether faster AI-assisted development is also weakening team understanding. Metrics like time-to-debug, onboarding friction, and dependency on a small number of experts can reveal cognitive debt before it becomes a delivery bottleneck.
• Bake explainability into workflows. When teams use agents to generate code, specs, or workflows, PMs can require artifacts such as decision logs, architecture notes, and interactive explanations so future contributors can understand why something was built and how it behaves.
• Reduce product and governance risk. Cognitive debt increases the chance that teams ship features they cannot confidently maintain, audit, or modify. For AI products especially, that can create reliability, compliance, and trust issues when model behavior or system logic needs investigation.

Related

• agent-written-code — A major driver of cognitive debt when generated implementations are accepted faster than teams can understand or review them.
• interactive-explanations — Proposed as a mitigation strategy: walkthroughs and explorable explanations can help preserve system comprehension.
• simon-willison — Amplified and contextualized the concept in newsletter coverage, especially around practical implications for AI-assisted software development.
• margaret-anne-storey — Credited with the clear explanation of how AI shifts concern from technical debt toward cognitive debt.