This architecture uses retroactive observation to identify and close system blind spots, leading to better alignment over time.
This architecture uses retroactive observation to identify and close system blind spots, leading to better alignment over time.
This is an open source project that allows us to create composable architectures that interact to create highly efficient, predictable, and controllable intelligence systems. It uses the idea of collective observation and retrospective improvement to create shared data sets and operations that allow composable intelligence systems to be designed and deployed.
This architecture turns every system into an observable system because it's able to reconstruct everything from components and understand more details over time. It's designed to look for things it can't explain and push sensors to that area until it understands the impact from a given segment of a system.
Team Member
The architecture has been designed and validated and is now being deployed as infrastructure within the Open Research Institute. Any funding will go to building applications on top of the architecture and educating builders on how to use these new structures.
The more money that comes in, the more sensors we'll be able to deploy and collectively learn from.
No comments yet. Be the first to share your thoughts.
