Six months in the lab.
AI researcher. Spent half a year deep in the models — studying how context compounds, how agents degrade, where the architecture breaks. The research that became Walnut's foundation.
What I'm building
[V]ALIVE Research
publishingThe empirical basis for alive computers. What happens when you give an agent persistent context across sessions. What breaks. What compounds.
[E]Agent Architecture
activeTesting how different context structures affect agent performance. File-based vs database. Flat vs nested. The experiments that shaped the ALIVE framework.
[L]Lab Notes
ongoingSix months of findings, dead ends, and breakthroughs. The raw material behind the thesis.
This world is powered by Walnut
open source context infrastructure