[ Journal ]

Thinking out loud.

No.ArticlePublished

Agentic Context Engineering: What Long-Running Agents Actually Need

Long-running agents fail on two things, not one: no durable context and no task backlog. Here is what agentic context engineering means in practice, worked through a sales enablement agent.

Sep 21, 2026

The LLM Context Window: Why Bigger Isn't the Answer

An LLM context window is the working memory a model has for a single request. Here is why a bigger window does not fix your AI, what degrades as it fills, and what a context window cannot do at any size.

Sep 13, 2026

What Is a Semantic Layer? (And How It Differs From a Context Layer)

A semantic layer is a governed business definition of your data that sits between your sources and the tools that use them. Here is what it contains, what it cannot do, and how it differs from a context layer.

Sep 6, 2026

Multiplayer AI: Why Your Team's AI Is Still Single-Player

Almost every way a company uses AI today is single-player: one person, one chat, starting from zero. Multiplayer AI runs on shared, living, governed context so every person and agent builds on the same source. Here is what the single-player default costs, and how to make the shift.

Jul 31, 2026

Your AI Tools Don't Talk to Each Other. Here's What That's Costing You

AI tool sprawl is the uncoordinated pile-up of disconnected AI tools, and its real cost is hidden well beyond the subscription bill. Here is where the money actually goes, and why a shared context layer is the durable fix.

Jul 22, 2026

A Field Guide to Agentic Workflows

An agentic workflow lets an AI agent plan and carry out multi-step work toward a goal. Here is what they are, what they look like in practice, and the failure modes to watch for.

Jul 14, 2026

Context Engineering vs. Prompt Engineering: What's the Difference?

Prompt engineering shapes what you ask. Context engineering shapes what the AI knows. Here is what each term means, how they differ, and why the center of gravity has shifted.

Jul 2, 2026

Enterprise Search, Explained (and Where It Falls Short for AI Agents)

Enterprise search helps people find what their company knows. Here is how it works, where it falls short for AI agents, and what comes next.

Jun 27, 2026

[ newsletter ]

Coconut Notes

A fortnightly note on putting agents to work: what we are learning about grounding them in your team's knowledge and measuring what they take on. No noise.

fortnightly. unsubscribe any time.