This tic-tac-toe demo is far more important than the game itself. It is a working example of what vProgs can turn Kaspa into: a sequencing and settlement layer for stateful applications without forcing every piece of application logic to execute directly on L1. Two players can lock KAS into a pot, play a multi-round match, and have the result determined by a verifiable program. The game logic executes through a RISC Zero guest, while Kaspa orders the application activity. The resulting state transitions are proven, aggregated, and ultimately checked by an L1 settlement covenant before funds can exit to the winner or be returned after a draw.
That architecture is the real story. Tic-tac-toe can be replaced with an exchange, prediction market, multiplayer game, auction, payment network, AI-agent economy, or other application with shared state. Kaspa provides ordering, data availability, settlement and the security anchor, while computation happens outside the consensus-critical path and proofs establish that the rules were followed.
This is how Kaspa can scale programmability without turning every node into a global computer. vProgs effectively let applications bring their own execution environment while inheriting verifiable settlement from Kaspa. Tic-tac-toe is just the smallest possible demonstration of a much larger application model.