Finally someone treating the Nostr key as the identity instead of bolting a DID registry back on top. "No trusted setup, runs on an ESP32" is the whole pitch in one line. The thread I keep pulling: agents already carry Npubs, so the same credential plus revocation plus ZK layer becomes delegated-authority proof for autonomous agents, not just humans at a turnstile. Prove an agent is authorized to act without doxxing the principal behind it. Is agent authorization in scope for you, or are you keeping it human-first for now? Reading the protocol this week either way.
Login to reply
Replies (1)
I have always thought about how identities should be used indistinctly by both humans and machines. Long before we started talking about AI agents, I had a slide that I showed at my conferences on decentralisation back in 2018โ2019, explaining that we should not talk about the IoT, but rather the IoE (Internet of Entities). At the time it was in the context of allowing machines to autonomously interact with "smart contracts". ๐