Daniel Ospina

Welcome

Current AI memory helps agents remember what happened. Tortoise — our memory substrate — helps them understand why something is true, update their world model as new evidence arrives, and make complex decisions without forgetting half the variables. Memory should be an active reasoning system, not passive storage: agents that spot contradictions before a sidequest gets out of hand, leave an auditable trail, and collaborate on a shared world model.

I've spent a decade obsessing over how we learn, coordinate, and make decisions. I founded RnD Ventures, a venture studio researching and incubating collaboration technologies, where we incubated 12 startups in partnership with L2 blockchains. Before that I was Head of Governance at Aragon and a member of the Supervisory Council at SingularityNet, developing on-chain governance systems and exploring AI governance.

The previous decade I spent in organisation design and collective intelligence: facilitating executive programmes through Saïd Business School (University of Oxford), and working with organisations including Google, BCG, the United Nations, and startups across Europe and Latin America.

My career has been unusually nonlinear. I began as a Michelin-trained chef before moving into food R&D, where I became fascinated by tacit knowledge — why experts make better decisions than they can explain. That question led me through psychology, philosophy, cybernetics, DAOs, and now AI memory.

These experiences have taught me that great ideas can come from anywhere, and that giving them fertile soil starts with building caring relationships and empowering systems.