Why I Prototype in Code
A static file can't tell you if an idea actually works. Code can.
A mockup can look finished and still be wrong. The gap between "looks right" and "works right" is where most product risk hides — and a static file is very good at hiding it.
Prototyping in code surfaces that risk early: real data, real states, real edge cases, instead of discovering all of it after engineering has already committed to a spec built from a flat screen.
It also changes the handoff itself. Instead of handing engineering a static file they have to interpret and rebuild from scratch, the prototype is the proven ground — behavior, transitions, and edge cases already validated before a single production sprint starts.
This matters most on AI and service-design work, where the real "interface" is a set of behaviors and decisions over time, not a fixed screen. (See Designing AI That People Actually Trust.)
It's slower up front and faster everywhere after. Fewer translation losses between what design intended and what actually ships.