Viviane Clay present to Jeff Hawkins and the Thousand Brains Project team about the Cortical Messaging Protocol (CMP, referred to here as Common/Cortical Communication Protocol), the critical idea that enables all modules in the Monty system to communicate. Developed from early ideas about a shared “AI Bus,” the CMP standardizes interactions between Monty’s sensors, learning modules, and motor systems, making Monty flexible and modular. 
This is a great video which makes the project more understandable. I am very keen on the release of the open source project:)
Great clear video on CCP. It seems similar to distributed messaging protocols I have worked on, The original one being the TIB (Tibco Information Bus). It had a few interesting characteristics that CCP might want to consider. First is was complete distributed with no centralized server, each process linked to an end point, so no single point to failure. Second was subject based addressing over publish and subscribe, a process would come up and start listening and publishing independently. Third was self describing messages allowing processes to evolve independently which becomes important as systems scale - versioning becomes important. This architecture (and similar ones) enable distributed models - I could imagine remote sensors connected over a distributed CCP acting with multiple “Montys” communicating through distributed CCP to create a large dynamic smart system (not sure if that is the correct terms). I could even imagine using CCP based on this to leverage LLMs as static data sources, similar to how humans use books. I think LLMs are a good way to take knowledge and normalize it into static machine communicable sensor data. Perhaps LLMs are the libraries of the Thousand Brain Project :-).
Thanks @Doonhammer. These are great points to consider as we scale the system to thousands of learning modules communicating together. Currently, as we experiment, we’re keeping the number of columns very small so we can test ideas in isolation.