Engineer Builds AI App in Six Hours, Spends a Year Reworking It

Generative AI is accelerating software prototyping by turning natural-language instructions into interfaces and working features. But British software engineer Alex Hyett’s habit-tracking app illustrates the gap between a functional demo and production-ready software. Architecture, maintainability, data integrity and synchronization still demand experienced engineering judgment, particularly when an app must meet the reliability standards expected for distribution through Apple’s App Store.

Hyett said AI-assisted tools produced an initial version of the app in just six hours, taking the project roughly 80% of the way toward completion. The remaining 20% proved far more demanding: tangled code and potential system-sync failures required a year of manual refactoring and detailed fixes before the app could reach the App Store. His experience shows that AI can compress early development dramatically while leaving the most consequential quality work to human engineers.