The “one-shot app” story is getting out of hand.
AI has changed how I build software. It helps with scaffolding, refactoring, tests, debugging, research, documentation, and exploring options.
The idea that I can describe a serious product once, send twenty agents into the night, and wake up to a business is fantasy.
It may work for a demo, internal tool, or small prototype. Software that real users depend on is different.
A real product contains years of edge cases, customer behavior, integrations, permissions, migrations, error handling, support, security, performance, and UX decisions. Those details do not fit neatly into one prompt.
Fully describing even a “simple” mature app would fill a small book.
AI can multiply a capable builder. A good engineer with good tools can explore more options, fix more small things, write more tests, and implement faster.
It does not remove product judgment, domain knowledge, architecture, QA, customer feedback, or maintenance.
The bottleneck moves from typing code to deciding what should exist, how it should behave, and whether it is good enough for users.
That work remains hard. Teams can now produce bad software faster too.