Every decision earns its place or it doesn't ship.
No feature enters the build because it was technically convenient. Each one is held against a single question: does a real hand need this during real work?
Every interaction is tested against one outcome: does it interrupt the work or vanish into it? If a UI element demands attention it hasn't earned, it's cut — regardless of how long it took to build.
The tool disappears.
Precision over scale is not a slogan — it's the constraint that shapes the build list. A narrow tool that holds up under deadline pressure is worth more than a wide one that buckles when the work gets dense.
Two filters. Everything else is noise.
Need vs. possible
Surface area vs. depth
A feature that's technically possible but solves no friction in the workflow doesn't enter the build. The question is always: what breaks without this at the moment it's needed?
Fewer capabilities, held to a higher standard of reliability. Scope is deliberately kept small so each function can be trusted under the conditions where the work actually gets done.