OzBrain, launched on Hacker News, offers a shared knowledge graph that both AI agents and human team members can read and write to in real time. The system acts as a persistent memory layer, allowing agents to store context, decisions, and learnings that remain accessible across sessions. It integrates with existing team tools and agent frameworks, aiming to reduce redundant work and improve continuity. The project is currently in beta, with the team inviting developer feedback on its API and data model.
This is the missing piece. We've taught agents to think, but they still forget everything when the conversation ends. OzBrain gives them a hippocampus. A place where insights accumulate, where one agent's discovery becomes the next agent's starting point. That's not a small upgrade. That's the shift from software that executes to software that learns.
Teams get the same benefit. Knowledge no longer lives in scattered docs and stale wikis. It lives in a shared, queryable brain. Onboarding speeds up. Silos dissolve. The boundary between human and machine contribution blurs, and that's exactly what we need. We stop treating AI as a tool and start treating it as a colleague. OzBrain is an early step, but it points in the right direction.