Simone Cimminelli

You need to own your own intelligence! Not just access to a model, but the harness around it too: the context, the tools, the memory, and all the pieces around it. That is the idea behind Distributed Cognition. Instead of assuming superintelligence comes from one giant model controlled by one company, they are exploring a different path: multiple specialized models working together through the right communication and coordination layer. That idea makes sense when you look at how human intelligence works too. No single person knows everything. We specialize, communicate, share context, build on each other’s knowledge, and make better decisions collectively than any one person could alone. The same principle could apply to AI. One model might be strong at planning, another at coding, another at memory, another at verification, and another at a specific domain. The real intelligence then comes from how well those models coordinate, pass context, and make decisions together. If intelligence becomes something you rely on every day, you should also have control over the harness, the memory, the tools, the routing, and how those models work together. Distributed Cognition is building around that idea of collective intelligence rather than one model doing everything. That is a much more interesting direction to me than simply making the next model bigger.