AGIFORGE
Decentralizing Intelligence Through Contributions
About
The problem it solves Current AI development is largely centralized and often unfair—individual contributors rarely receive proper credit, profits are distributed unevenly, and intellectual ownership is typically unclear. Our platform addresses these issues by allowing anyone to contribute AI models and prove their authorship on-chain. It ensures automatic revenue sharing whenever a contributor’s work is used, making the entire process transparent, auditable, and governed by the community. By enabling open collaboration on modular AGI components, the platform fosters a fairer, decentralized ecosystem that makes AI development more inclusive, rewarding, and trustworthy for everyone.
How it's made
Challenges we ran into While building the project, we faced several key challenges. Initially, we tried storing AI models directly on-chain, but due to gas and size limitations, this wasn’t feasible. We resolved it by using IPFS for decentralized storage and linking the file hashes in smart contracts. Another major hurdle was ensuring fair evaluation of AI contributions without using zero-knowledge proofs. To address this, we made all evaluation scripts and scores publicly accessible and incorporated DAO-based community voting for validation. We also encountered bugs in our revenue-sharing smart contract, particularly with rounding errors when distributing funds. This was fixed by using SafeMath and extensive testing with edge cases. Lastly, wallet integration posed a challenge for users unfamiliar with Web3. We improved this by enhancing the UI experience and providing clear onboarding instructions.