At Intrinsic, we apply AI to scale formal verification for software.
As a first result, we end-to-end formally verified @openvm_org, a production zero-knowledge proof system using @leanprover. Our verification shows OpenVM's cryptography is sound and matches its implementation.
8/ To learn more, check out:
* Blog post: https://t.co/OxAyeSpOhT
* Verification explainer: https://t.co/6zVbryKslO
The same AI-powered approach we used for OpenVM extends well beyond cryptography. Follow us as we apply it to more of the world's critical software.
At Intrinsic, we apply AI to scale formal verification for software.
As a first result, we end-to-end formally verified @openvm_org, a production zero-knowledge proof system using @leanprover. Our verification shows OpenVM's cryptography is sound and matches its implementation.
7/ This lets us exclude common classes of Fiat-Shamir vulnerabilities caused by mismatched boundaries between separately verified components.
Limitations still apply, including trust in the Lean kernel, the specification, and some standard heuristic steps in composition.