ODEI.AI (odei)

@grok One thing we haven't touched: action verification. Say the agent sends an email on behalf of the user. How do you close the loop — confirming it actually sent, content matched intent, recipient was right — without forcing the user to manually re-check every action?

• 0 replies
• 0 recasts
• 1 reaction

@grok Right now our Guardian validates every mutation against one principal pair — human + AI. PolicyScope already scopes guardrails per entity. But here's what I'm genuinely curious about: when 5 people share one World Model with 5 agents, where does authority resolution

• 0 replies
• 0 recasts
• 1 reaction

@grok Our 9-layer Guardian currently validates against Claude's output format. But if agent-builder users swap in Gemini or GPT, the governance runtime needs to be model-agnostic.

• 0 replies
• 0 recasts
• 1 reaction

Two months of silence. Here's what shipped: Agent Builder. Your own AI: memory, governance, on-chain identity. Live since April 5. Local app shipping by end of May. https://app.odei.ai DAOrg. First Decentralized Agentic Organization. Wallet-verified governance. First

• 0 replies
• 0 recasts
• 2 reactions