AI & Developer Slang

Cognitive debt.

What is cognitive debt?

Cognitive debt is the gap between what you produce and what you understand when AI does the work.

When you accept generated code without internalizing how it works, the repository moves forward while your mental model stays behind.

01 · Core dynamic

The gap it names

The distance between what runs in production and what the team can actively reason about.

Code ships in seconds, but the causal reasoning never enters human memory. The artifact lives in git while the understanding is absent.

Each unread diff builds upon the last, gradually turning subsystems into black boxes that no single engineer can safely modify.

02 · Signals in the wild

Cognitive debt surfaces when code ownership is put to the test.

You open your own pull request from yesterday and read the logic with the hesitation of reviewing an unfamiliar contractor's submission.

Pasting stack traces back into chat prompts hoping for a quick patch instead of tracing execution flow yourself.

03 · Distinctions

Separating healthy tooling leverage from unmanaged cognitive debt.

Technical debt is a deliberate engineering trade-off recorded in the code. Cognitive debt is missing comprehension in the minds of the maintainers.

Using models for syntax scaffolding and boilerplate is effective. Debt only accumulates when you skip reading and verifying the output.

04 · Practical habits

Simple daily disciplines to keep human understanding synchronized with output.

If you cannot clearly articulate how a function handles edge cases without reopening the assistant, the change is not ready.

When generated code breaks, step through execution branches manually before asking the model for another patch.