What is Technical Debt?
Technical Debt: The future cost of shortcuts in code and architecture, paid as slower changes, more bugs, and riskier releases until the shortcuts are fixed.
Technical debt is a metaphor coined by Ward Cunningham: taking a shortcut to ship sooner is like borrowing money. You get speed now, and you pay interest later in the form of code that is harder to change, bugs that are harder to find, and systems that are harder to scale. Some debt is deliberate and sensible; some is accidental, from inexperience, rushed work, or AI-generated code nobody reviewed.
Debt shows up as missing tests, duplicated logic, hard-coded values, outdated dependencies, no error handling, tangled data models, and fragile deployments. It compounds: each new feature built on a shaky base adds more. Paying it down means refactoring, adding tests, and fixing structure, ideally a little every sprint rather than in a rare big rewrite.
Why it matters when you build or buy AI
Taking on debt in an MVP is fine, because most MVPs change direction and much of that code gets thrown away. The mistake is scaling a product on MVP-grade code after it finds traction. When growth starts, budget time for paying down the debt that slows you most. If every small change now takes days and breaks something else, you are already paying heavy interest.
Example
An MVP stored all settings in one JSON column to ship fast. Once customers needed per-team settings, every change risked breaking others, so the team spent a sprint migrating to a proper table.
Related terms
More in Engineering
Book 20 minutes, or send a brief
A former CTO replies. You leave with a written plan and a fixed estimate.
Prefer LinkedIn? Connect with Harish.