Most AI tools still rely on a chatbox. A single stream of text between user and model. It works, but it limits how ideas form. Interaction becomes a sequence of commands instead of a space for thinking.
The question is what happens if AI is given a new medium. Not text. Not voice. Something that can hold structure, relationships, and context the way human work naturally does.
Text-based AI interfaces flatten thought. Every question or response sits on the same plane, stripped of its connections to other moments.
When people think, they map ideas visually or contextually. Notes, folders, boards, timelines. These help memory and meaning build over time. In comparison, AI tools restart context with every chat or thread, forgetting the spatial or emotional traces of previous work.
A new interaction medium would treat AI as a participant inside a workspace, not just a messenger in a box. Instead of asking for responses, users could shape an environment where the model understands objects, layouts, and progress.
The AI would recognize patterns in how people organize, not just what they say. The interface becomes a shared surface between human and model, one that remembers the work and reflects the rhythm of thought.
The open question is how space should behave when AI interacts within it. Should ideas cluster on their own?
Should memory live in the workspace or in the model?
How much structure helps before it limits exploration?
What does understanding look like when it’s not a sentence but a pattern in space?
- Sam