/ knowledge graph /
The research map behind my AI work.
This is not a map of everything I read. It is a working research layer for turning AI noise into production judgment.
500+live wiki pages and research notes
4k+relationships across concepts, entities, and sources
1purpose: better judgment on what matters in AI
/ start with a route /
Follow the questions I keep coming back to.
The full graph is useful, but the best way in is through a concrete question: what is production-ready, what changes search, what improves decision-making, and what is just noise.
Production AI
Move from pilots to systems people can trust, operate, and improve.
Agentic workflows
Track how agents change work, handoffs, tools, and accountability.
AI search / GEO
Understand how answers, citations, and machine-readable authority are changing discovery.
Experimentation
Connect AI ideas to measurement, learning loops, and better decisions.
Decisioning
Follow the link between recommendations, personalisation, ranking, and product judgment.
Operating models
See what has to change around governance, ownership, and practical rollout.
/ live graph /
Explore the working map.
Start with the curated graph for signal. Switch to the list view if you want a readable alternative, or open the full graph when you want to inspect the whole research layer.
The graph is the research engine. De:Coded is the distilled readout.
I use this map to connect sources, ideas, and recurring patterns. The newsletter turns that work into a concise briefing on what changed, why it matters, and what is worth watching next.