Prototype notesAgents for Narrative
Esther Huang
What does it take for an AI system to accumulate context and actually hold onto it?
That's the question driving most of what I build. I've worked on multi-agent systems where memory had to be shared across agents, GUI agents that needed to stay reliable in real UI environments, and narrative systems where coherence had to hold across long, branching stories. The problem keeps showing up in different shapes.
Different projects, same underlying problem.
Memory, context, and reliable action. In storytelling, in workflows, in long-term agent behavior. These are the areas I keep returning to.
Writing and notes from along the way.
Home stays brief. The writing page carries the deeper archive.
Recent writing
View allAI Safety and Philosophy
Switching On the Ring of Gyges
Building in public
My Personal Website Is Not a Portfolio. So What Is It?
Contact
For AI agent work, early-stage product thinking, or technical collaboration.