Problem
Prototyping used to be expensive, so we'd only try one or two approaches. We'd pick the "most likely" solution and hope it worked. Better alternatives stayed unexplored.
Pattern
Prototyping is now so cheap that you can try all paths. Build 10 variations, not 1. Test them all. Pick the best. Learn as you're prototyping. With AI, a working prototype costs almost nothing. You can build and test multiple prototypes per day. Actually feel how each one works instead of imagining it.
Example
Claude Code's todo list: 20 prototypes in 2 days
- Inline progress vs drawer vs pill vs spinner
- Each variation fully functional, tested with users
- Found the best through exploration, not guesswork
UX decisions: Instead of mockups, build 10 working versions. Play with all of them (can also borrow behaviors and combine)
The power: You're no longer limited to your first idea. Explore the whole solution space.
