Moving Through Transition

You have probably already used AI to some degree. Some of us use it every day. What is harder is picturing what any of it means a few years out or how we coordinate as working people in a world with superintelligence around us.

I have spent four years learning how these systems work. For the last stretch of it I have been traveling through Asia, having as many raw conversations as I could find about what people plan to do next. This field guide to the transition we see ourselves sitting inside is what came of those conversations and research.

The first half is practice. What "AI-native" means in 2026, and the three levels we see emerging. How to read the model landscape from closed to open source. What an agent harness is, and the anatomy of one: strip an agent down and a model sits at the center, and the harness is everything built around it, each part standing in for something the model cannot do alone. Context engineering, and why models run in loops. Agents, skills, and evals.

The second half is on meaning. What becomes of attention, agency, and judgment when the tools are this capable. The jagged frontier, and where taste still decides. Life after labor, who ends up owning the machine intelligence, and how work might change specifically through cultivating a life of adventure. It builds toward an argument: that the unit of production is moving away from the lone worker and the firm and toward the small group that travels together, that there are thirteen key roles that survive an AI transformation, and that small group formed runs it's own loops in exploration of a more beautiful life.