This site is about our ideas on embodied AI’s requirements for learning how to behave in certain places. That means we think they will have their own ways of learning and doing things, but we should not leave it at that, we have to have ways to oversee what’s going on.

Every site has it’s hazards. Ideally everyone is aware of them. How you learn about hazards can happen in a variety of ways. Let’s just say that embodied AI likes to learn from data. If you have a changing environment you are going to need streams of data.

No one is going to be able watch all the data streams and make decisions. Well, maybe something running in the data stream can.

Agents, or special code running on the network paths of the data streams can watch everything. However you still need to tell them what to do when they see something. Agents need well thought out policy. This is our mission, determine which agents work best and what policy solves the problem.