#sk-snarks

Notes from the talk by Elad Verbin at Web3Summit.

https://oceanprotocol.com https://www.openmined.org

Cryptographically provable privacy Private computation Crypto Governance

Privacy on the blockchain, why should I care?

Reasons:

1. Blockchain privacy is broken! Public keys are visible to everyone if you use it for any transactions all activity is visible on the chain. 2. Private computation is necessary for Web3. Uber example, complex system, post private information on the blocks, reputation change, a business logic, that means computation changes. With public computation, you can’t do that without publicly visible, with private computation you can without revealing the business logic. 3. More privacy - > more information shared - > more value all around. It fosters collaboration. In a world where everything is computed privately, we would have a lot of collaboration.

Private computations: Zero-Knowledge Proofs, Secure Multiparty Computation, sk-Snarks, homographic encryption, secure function evaluation.

Secure Multiparty Computation (SMPC)

Zero-knowledge: I want to prove something without telling you the details about it, without teaching you how I do it.  

https://zkp.science http://www.multipartycomputation.com/mpc-software

How is private computation used?

Example: Ad Auctions

• publishers are trusted middlemen with massive access to our private data, we must decentralize them • Run SMPC-based auction • User data is kept private • they can now advertise based on private data (browser history, bank records, medical records, DNA)

I can advertise based on your DNA. This is not the world I want to live in, it’s a world I’ll be living in.

Reward user for sharing data. More data -> more reward

Algorithmic Redistribution of Wealth combining:

• Algorithmic (reinforcement learning, control, and optimization theory) • Mechanism design (game theory) • Economics (public policy, economics, behavioral economics) • Cryptography (SMPC)

Practical issues:

• How do I use SMPC in my dApp? • Is SMPC too slow? Yes, but we can believe in Moore’s Law for private computation. We can use Trusted Execution Environments (enigma) https://enigma.co • Will the user trust you/your dApp?

Conclusion: