Terms  /  System Transparency  /  Open State
06 · System Transparency

Open State

The system's internal state is visible and inspectable at any time. The opposite of Invisible State.

Open State means the user can look under the hood whenever they want. What does the system currently "know"? What assumptions is it operating under? What's in its memory? What context is it using? Open State transforms the AI from a black box into a glass box — the complexity is still there, but it's visible. This doesn't mean the user needs to look at the state constantly; it means they can look whenever something seems off, and they'll find answers instead of mysteries.

Example
A developer notices their AI assistant is making odd architectural suggestions. With Open State, they pull up the agent's current context and find: the agent's memory includes a conversation from three weeks ago where the developer was experimenting with microservices architecture for a side project. The agent has been applying microservices patterns to the current monolithic application because it didn't distinguish between the two projects. The developer clears the irrelevant memory, and the suggestions immediately improve. Without Open State, they'd just experience inexplicably poor suggestions.

More in System Transparency