⏳Future

Future of the Project

• On-Chain Randomness for Obfuscation We plan to leverage Sui’s on-chain randomness oracle (e.g., via a VRF) to drive the selection and arrangement of image tiles during obfuscation. By incorporating verifiable random values directly on-chain, each mint can use a unique, unpredictable pattern—making it impossible for attackers to predict or reconstruct the obfuscation scheme ahead of time. This also removes reliance on off-chain RNG services and enhances trustless security.

• AI-Powered Obfuscation Assessment To ensure optimal privacy without over-obscuring, we’ll train or integrate an AI model that evaluates a candidate obfuscated image and estimates how recognizable its features remain. Before finalizing the tiles, the system can run a quick inference to confirm that human and ML-based models can’t satisfactorily reconstruct the artwork from the preview. This feedback loop helps automate the balance between “too little” (risking easy reconstruction) and “too much” (making it unappealing to buyers).

• Dynamic Blocksize Selection Rather than a fixed four-tile split for every artwork, future versions will analyze each image’s complexity and distribution of detail to choose an optimal blocksize (e.g., 4×4, 8×8, or even irregular grid partitions). Simple, low-detail images might need fewer, larger tiles to preserve buyer visibility, while highly detailed pieces could split into many small tiles for stronger obfuscation. Dynamically adjusting blocksize per-image maximizes both security and aesthetic preview quality.

• Enhanced Merkle Tree Implementation We’ll replace the current static, full binary Merkle-tree scheme with an optimized Merkle forest or sparse Merkle tree that supports incremental updates, batch verifications, and smaller proof sizes. This allows faster tree construction for large tile counts, supports efficient partial reveals (e.g., streaming a subset of tiles to a buyer), and reduces on-chain overhead. By improving how we build and traverse the Merkle structure, both minting and proof-verification costs can be lowered.