Essays on human-centered AI, organizational intelligence, multi-agent systems, and what it actually takes to make enterprise AI work inside real organizations. Most of it starts as a question I cannot find a good answer to.
I built The AI Power Map by using AI agents to collect and verify public data at scale, then applying network-science analysis to turn it into an interactive map and a ~70,000-word book. It maps 420 people and 1,724 sourced relationships to show how power, capital, talent, and trust move through the AI industry, and where outsiders may find entry points. I designed it as a research artifact that makes influence visible through evidence, structure, and network patterns.
How every major enterprise AI category converges on the same blind spot, and why closing that gap represents an entirely new product category.
Why agents fail at routing despite solid retrieval systems, and how behavioral context is the overlooked solution to organizational intelligence gaps.
How I built and scaled a team of AI agents for real operations using OpenClaw. Ten design principles, and why building an agent system feels more like organizational design than programming.
Less about work, more about how I think.