Building
Contexta: context infrastructure for agents that need memory, provenance, and accountable action.
Now / May 2026
Current attention: context-aware AI, virtual employees, and the future of purposeful work.
This page is an operating note: what I am actively building, thinking about, learning, and protecting time for.
Testing
AI-native workflows where virtual employees can own bounded work without hiding uncertainty from humans.
Writing
A thesis on meaningful work after automation: how people keep authorship when execution becomes abundant.
Learning
Complexity science, organizational design, proof systems, and architectural spaces that make thought feel embodied.
Priority
I am optimizing for depth, proof, and useful public signal.
Contexta is the center of gravity: a way to make agent memory inspectable, accountable, and operationally useful. I am studying how context changes when it moves from static documents into workflows that can act.
Around that sits Amino.ai, where I am testing what a virtual employee should be allowed to own. The hard question is not whether an agent can complete a task. It is whether the system knows the boundaries of responsibility well enough to deserve delegation.
Work for Future is the human thesis underneath the technical work. If AI makes execution cheaper, people need better tools for direction, meaning, and career agency. Otherwise automation wins the calendar and loses the soul.
Operating notes
Build fewer things, with more memory inside each one.
Prefer systems that create compounding leverage over demos that create temporary attention.
Keep the human reason visible when designing automated work.
Write in public when the idea becomes clearer by being made accountable.