Card-game AI must first learn not to see: information sets, partnerships, and Wasm
How we kept card-game AI inside each player's observation, handled hidden deals and partnerships, and moved complete search into local browser Wasm.
JOURNAL · BUILDING IN PUBLIC
Design decisions, engineering lessons, and field notes from building The World at Play. No polished launch copy—just the problems, judgments, and work behind the site.
JOURNAL
Practical notes on browser games, local AI, content design, and static publishing.
How we kept card-game AI inside each player's observation, handled hidden deals and partnerships, and moved complete search into local browser Wasm.
How we chose different search methods for different games, moved complete search from TypeScript into Wasm, and kept NNUE, KataGo, and GNUbg on the player's device.
The project began with one board. Here is why it kept collecting games from different places and different traditions of thought.