Really interesting breakdown of RPC data quality, especially the replay + majority approach.
One thing we’ve been digging into is how to measure state correctness beyond node agreement. Majority voting works well, but can still break when multiple nodes return the same incorrect result.
We’re working on a complementary layer around verifiable execution (WASM + zk) and consistency scoring, which could potentially reduce replay overhead and improve node selection.
Curious if you’ve thought about confidence in state as a separate primitive from node-level trust.
That's really neat and would love to learn more. The question is how do you do this in a way that make the unit economics make sense imo. Also not sure if the challenge is always around malicious calculations but rather the hardware not setup correctly.
Really interesting breakdown of RPC data quality, especially the replay + majority approach.
One thing we’ve been digging into is how to measure state correctness beyond node agreement. Majority voting works well, but can still break when multiple nodes return the same incorrect result.
We’re working on a complementary layer around verifiable execution (WASM + zk) and consistency scoring, which could potentially reduce replay overhead and improve node selection.
Curious if you’ve thought about confidence in state as a separate primitive from node-level trust.
That's really neat and would love to learn more. The question is how do you do this in a way that make the unit economics make sense imo. Also not sure if the challenge is always around malicious calculations but rather the hardware not setup correctly.
see DM 🫡