As blockchain technology matures, the expansion from purely public state systems, pioneered by Ethereum, to privacy-centric platforms like Aztec and Aleo marks significant progress. However, the development of private shared state—where multiple parties can collaborate on data without revealing it to others—remains an unmet challenge. This talk outlines collaborative SNARKs, a cryptographic primitive that combines Zero-Knowledge (ZK) proofs with Multiparty Computation (MPC) to address this gap.
We will begin by exploring the motivations for private shared state, with applications in areas such as privacy-preserving machine learning, auctions, and gaming. The foundational research from Ozdemir and Boneh's 2022 paper on coSNARKs will be discussed, showcasing how these techniques enable secure, distributed proof generation among multiple participants.

Moving from theory to practice, we will look at the engineering challenges of implementing coSNARKs, such as optimizing the witness extension for practical use, and achieving real-world scalability. Benchmarks from recent alphanet deployments will highlight the performance and usability of coSNARKs in real-world environments.

We will mention the current limitations of this technology and the open problems in the field, inviting researchers and developers to explore its future possibilities. We will close by emphasizing how coSNARKs can broaden the Ethereum landscape by leveraging programmable cryptography to collaborate across private shared state.