How do we Plan? First we Sense. How do we Sense? First we Plan the Sensing... argh....

OK, I want to formulate a Plan to achieve my goal of aligning the code and funcationality to new intentions.

Let’s look at what we need to do.. yup this looks like a good start…

Sense (run characterisation experiments)                                                                      
    ↓ produces                                                                                                
s₀ (initial state: model capabilities + instance properties)                                                  
    ↓ input to                                                                                                
Plan (HTN decomposition: select methods whose preconditions match s₀)                                         
    ↓ produces                                                                                                
π (plan: ordered generator preferences per step)                                                              
    ↓ executed by                                                                                             
DualStateAgent (via DI: injects generators from π)                                                            
    ↓ monitored by                                                                                            
Guards (sense execution state, feed back to agent)                                                            
    ↓ when stuck                                                                                              
Replan or Escalate                                                                                            

Sweet. Let’s Sense the environment first then we can make a Plan.

Right how do I Sense the environment… I need a Plan for Sensing… :face_palm: :exploding_head: :man-facepalming:

Time to revisit the AI Planing Bible… :laughing: