Research & Thoughts.

A first-principles framework for agentic AI. Every viable entity runs six universal loops—sensing, modelling, decision, action, reflection, memory. Most AI tooling optimises the one that was already working. This paper builds the framework for doing it properly.

Why better models and better prompts still fail to produce business outcomes—and why the missing layer is a structured operating frame for the job being performed, bound to the individual's living context through a dual-graph architecture.

How Karpathy's Autoresearch validates the Domain Operating Frame thesis—from a completely different direction. When ML research and business context arrive at the same architecture, the pattern is real.

Not the manual. The 25 ways your team actually gets work done—and how to capture all of them, including the workarounds nobody can articulate, and turn them into agents.

Anthropic's new labor market report tracks AI displacement. The real crisis is context erosion—and no one is measuring it.

Why AI context must be built bottom-up, not enterprise-down. Beginning with the smallest common denominator of intelligence.

How Human Edge's Context Engine addresses every recommendation from HBR's thesis on organizational context as competitive advantage.

When AI learns to draw, what happens to the people who used to? An interactive data visualization exploring stock volatility against AI breakthroughs.