# Len P. van der Hof

> Official site of Len P. van der Hof, MSc — Dutch entrepreneur, AI innovator and author of
> Systems for the Strategic Self, an operator-grade nonfiction series on founder psychology,
> agentic AI systems, evidence-based decision-making, and building under pressure. Every title
> is published as two native editions, English and Dutch.

This non-standard file is an experimental discovery aid. The canonical HTML pages remain
authoritative. Everything below is generated from the same data that renders the site, so it
cannot drift from the pages themselves.

## Primary pages

- [Homepage](https://lenvanderhof.com/)
- [Books (English)](https://lenvanderhof.com/books/)
- [Boeken (Dutch)](https://lenvanderhof.com/boeken/)
- [Series](https://lenvanderhof.com/series/systems-for-the-strategic-self/)
- [Frameworks](https://lenvanderhof.com/frameworks/)
- [Start here](https://lenvanderhof.com/start-here/)
- [Worksheets](https://lenvanderhof.com/worksheets/)
- [About](https://lenvanderhof.com/about/)
- [Speaking](https://lenvanderhof.com/speaking/)
- [Podcasts](https://lenvanderhof.com/podcasts/)
- [Work with me](https://lenvanderhof.com/work-with-me/)
- [Undominated.ai](https://lenvanderhof.com/undominated/)
- [MCP](https://lenvanderhof.com/mcp/)
- [Contact](https://lenvanderhof.com/contact/)
- [English blog](https://lenvanderhof.com/en/blog/)
- [Dutch blog](https://lenvanderhof.com/nl/blog/)
- [Author profile](https://lenvanderhof.com/en/authors/len-p-van-der-hof/)
- [Long-form machine summary](https://lenvanderhof.com/llms-full.txt)

## Desk companion

POST JSON to https://lenvanderhof.com/api/companion (trailing slash accepted) with the current page path and a question. No client API key; the NVIDIA token stays on the server. Same-origin, rate-limited, best-effort read-help. It is not an authenticated write API. Jobs: ask, syllabus, explain, book-fit, lint. Answers are grounded markdown that recommend existing URLs on this site only. Never invent ASINs. In-production books are not for sale. Contract: https://lenvanderhof.com/openapi.json. Health: GET https://lenvanderhof.com/api/companion/health.

## About the series

Systems for the Strategic Self is written for founders, operators and builders who work under
real pressure. Each book delivers one named framework the reader can run, not just read, and is
published in a native English edition and a native Dutch edition, never a machine translation.
51 titles are designed, covered and documented; 11 are available now on
Amazon. Formats follow each title's live purchase ASINs: Kindle, paperback and hardcover where
those editions exist. The wider portfolio holds 133
active book projects.

Books are produced through a gated pipeline: sourced research, a claim-level evidence ledger,
structural review, fact-checking, red-team critique and a bilingual final edit. AI agents do
specialist work inside those gates; judgment, voice and accountability stay human. That production
system is itself documented in The Agentic Author.

## Available now

### The Psychology of Founder Resilience: How High-Stakes Leaders Stay Grounded, Clear, and Effective Under Pressure

Most founder advice tells you to grind harder. This book gives you something rarer: a trainable psychological system for the specific stress of building a company. Drawing on cognitive psychology, affective science, and decision research, it introduces the RESILIENT framework, nine learnable capacities that separate founders who survive from founders who stay clear under pressure. Operator-grade psychology with worksheets, protocols, and a dashboard, designed to compound.

- Page: https://lenvanderhof.com/books/psychology-of-founder-resilience/
- Dutch edition: De psychologie van oprichtersveerkracht — https://lenvanderhof.com/boeken/psychology-of-founder-resilience/
- Theme: Founder psychology
- Framework: RESILIENT (Reappraisal, Emotional granularity, Somatic regulation, Identity flexibility, Learning orientation, Interpersonal support, Executive function, Narrative integration, Titrated exposure)
- Length: 12 chapters, 276 pages in print
- ISBN (English paperback): 9798183948950
- Buy, English Kindle: https://www.amazon.com/dp/B0H6PH4V7D?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=psychology-of-founder-resilience&utm_content=kindle&tag=lenvanderho0f-20
- Buy, English paperback: https://www.amazon.com/dp/B0H6Q4CP83?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=psychology-of-founder-resilience&utm_content=paperback&tag=lenvanderho0f-20
- Buy, English hardcover: https://www.amazon.com/dp/B0H7PJ3S8Z?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=psychology-of-founder-resilience&utm_content=hardcover&tag=lenvanderho0f-20
- Buy, Dutch Kindle: https://www.amazon.nl/dp/B0H7TGWKNW?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=psychology-of-founder-resilience&utm_content=kindle&tag=lenvanderho01-21
- Buy, Dutch paperback: https://www.amazon.nl/dp/B0H85644VB?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=psychology-of-founder-resilience&utm_content=paperback&tag=lenvanderho01-21
- Buy, Dutch hardcover: https://www.amazon.nl/dp/B0H7XX8LTT?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=psychology-of-founder-resilience&utm_content=hardcover&tag=lenvanderho01-21
- Kindle Unlimited: yes, free for KU members
- Availability: Available now

### AI Agents for Startup Strategy: How Founders and Operators Deploy Agentic AI to Move Faster and Think Smarter

Markets move faster than human analysis allows, yet most founders still use AI to draft emails while rivals build autonomous intelligence layers. This book gives you the SENSE framework: a deliberate architecture for agents that continuously monitor competitors, sense markets, and feed your judgment instead of replacing it. A strategic lens, not a coding manual, for operators who would rather run a tested protocol than improvise under pressure.

- Page: https://lenvanderhof.com/ai-agents-for-startup-strategy/
- Dutch edition: AI-agenten voor startupstrategie — https://lenvanderhof.com/boeken/ai-agents-for-startup-strategy/
- Theme: AI & agents
- Framework: SENSE (Scan, Evaluate, Navigate, Simulate, Execute)
- Length: 17 chapters, 274 pages in print
- ISBN (English paperback): 9798187502974
- Buy, English Kindle: https://www.amazon.com/dp/B0H96R4573?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=ai-agents-for-startup-strategy&utm_content=kindle&tag=lenvanderho0f-20
- Buy, English paperback: https://www.amazon.com/dp/B0H96GF2PG?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=ai-agents-for-startup-strategy&utm_content=paperback&tag=lenvanderho0f-20
- Buy, English hardcover: https://www.amazon.com/dp/B0H96HQYY9?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=ai-agents-for-startup-strategy&utm_content=hardcover&tag=lenvanderho0f-20
- Buy, Dutch Kindle: https://www.amazon.nl/dp/B0H961ZRB9?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=ai-agents-for-startup-strategy&utm_content=kindle&tag=lenvanderho01-21
- Buy, Dutch paperback: https://www.amazon.nl/dp/B0H96CL1BZ?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=ai-agents-for-startup-strategy&utm_content=paperback&tag=lenvanderho01-21
- Buy, Dutch hardcover: https://www.amazon.nl/dp/B0H96F8CSK?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=ai-agents-for-startup-strategy&utm_content=hardcover&tag=lenvanderho01-21
- Kindle Unlimited: yes, free for KU members
- Availability: Available now

### The Agentic Author: The BooksOS Method for Writing Serious Nonfiction with AI Agents

Most writers use AI to go faster. This book shows serious nonfiction authors how to use it to go deeper. Built around the BooksOS production system, it treats AI agents not as ghostwriters but as a team of specialists working behind evidence ledgers, quality gates, and red-team review. You keep the judgment and the voice. The agents raise the ceiling on rigor, accuracy, and output.

- Page: https://lenvanderhof.com/books/the-agentic-author/
- Dutch edition: De agentische auteur — https://lenvanderhof.com/boeken/the-agentic-author/
- Theme: Evidence & learning
- Framework: BooksOS (Brief, Bible, Map, Draft, Edit, Package, Launch)
- Length: 12 chapters, 320 pages in print
- ISBN (English paperback): 9798187632381
- Buy, English Kindle: https://www.amazon.com/dp/B0H9GSLQKN?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agentic-author&utm_content=kindle&tag=lenvanderho0f-20
- Buy, English paperback: https://www.amazon.com/dp/B0H9HHXTF9?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agentic-author&utm_content=paperback&tag=lenvanderho0f-20
- Buy, English hardcover: https://www.amazon.com/dp/B0H9H2BN4D?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agentic-author&utm_content=hardcover&tag=lenvanderho0f-20
- Buy, Dutch Kindle: https://www.amazon.nl/dp/B0H9H4WCLD?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agentic-author&utm_content=kindle&tag=lenvanderho01-21
- Buy, Dutch paperback: https://www.amazon.nl/dp/B0H9GXYQNT?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agentic-author&utm_content=paperback&tag=lenvanderho01-21
- Buy, Dutch hardcover: https://www.amazon.nl/dp/B0H9H2R45W?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agentic-author&utm_content=hardcover&tag=lenvanderho01-21
- Kindle Unlimited: yes, free for KU members
- Availability: Available now

### The Agentic Codebase: The Five-Layer Architecture That Makes Agent Work Survive Production

Prompts live in chat history, rules get copy-pasted across tools, and MCP servers sprawl without contracts, until the repo no longer survives a model upgrade or a second engineer. The Agentic Codebase introduces STACK, a five-layer framework for repositories that humans and agents maintain together. This is architecture, not a tutorial: the repo is the operating system, and every layer is versioned, tested, and owned like production code.

- Page: https://lenvanderhof.com/books/the-agentic-codebase/
- Dutch edition: De agentische codebase — https://lenvanderhof.com/boeken/the-agentic-codebase/
- Theme: AI & agents
- Framework: STACK (Structure, Toolchain, Agent configuration, Connection, Knowledge and quality)
- Length: 18 chapters, 482 pages in print
- ISBN (English paperback): 9798187627509
- Buy, English Kindle: https://www.amazon.com/dp/B0H964XXTS?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agentic-codebase&utm_content=kindle&tag=lenvanderho0f-20
- Buy, English paperback: https://www.amazon.com/dp/B0H96NPJ1B?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agentic-codebase&utm_content=paperback&tag=lenvanderho0f-20
- Buy, English hardcover: https://www.amazon.com/dp/B0H96LZH39?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agentic-codebase&utm_content=hardcover&tag=lenvanderho0f-20
- Buy, Dutch Kindle: https://www.amazon.nl/dp/B0H9CRJV5P?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agentic-codebase&utm_content=kindle&tag=lenvanderho01-21
- Buy, Dutch paperback: https://www.amazon.nl/dp/B0H9G3T2DJ?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agentic-codebase&utm_content=paperback&tag=lenvanderho01-21
- Buy, Dutch hardcover: https://www.amazon.nl/dp/B0H9GH4KKM?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agentic-codebase&utm_content=hardcover&tag=lenvanderho01-21
- Kindle Unlimited: yes, free for KU members
- Availability: Available now

### The Multi-Agent Organization: How Humans and AI Agents Share Work Without Losing the Plot

You added agents to your team faster than you added protocols, and now context dies at every handoff while nobody knows which agent ran last, on what input, with what authority. This book introduces CHORUS, six coordination skills that turn a loose collection of humans and agents into a multi-agent organization with explicit charters, clean handoffs, orchestrated workflows, and real review gates. It is the operating layer between having AI tools and running an AI-powered team.

- Page: https://lenvanderhof.com/books/multi-agent-organization/
- Dutch edition: De multi-agent organisatie — https://lenvanderhof.com/boeken/multi-agent-organization/
- Theme: AI & agents
- Framework: CHORUS (Charter, Handoff, Orchestrate, Review, Update, Sync)
- Length: 13 chapters, 276 pages in print
- Buy, English Kindle: https://www.amazon.com/dp/B0H9Z6QCTR?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=multi-agent-organization&utm_content=kindle&tag=lenvanderho0f-20
- Buy, English paperback: https://www.amazon.com/dp/B0H9YG78TK?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=multi-agent-organization&utm_content=paperback&tag=lenvanderho0f-20
- Buy, English hardcover: https://www.amazon.com/dp/B0H9YQR1WB?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=multi-agent-organization&utm_content=hardcover&tag=lenvanderho0f-20
- Buy, Dutch Kindle: https://www.amazon.nl/dp/B0HBNH93V4?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=multi-agent-organization&utm_content=kindle&tag=lenvanderho01-21
- Buy, Dutch paperback: https://www.amazon.nl/dp/B0H9ZRLL4B?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=multi-agent-organization&utm_content=paperback&tag=lenvanderho01-21
- Buy, Dutch hardcover: https://www.amazon.nl/dp/B0HB46WQQT?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=multi-agent-organization&utm_content=hardcover&tag=lenvanderho01-21
- Kindle Unlimited: yes, free for KU members
- Availability: Available now

### The Calibrated Mind: How to Hold Judgments at Their Right Strength Under Uncertainty

You cannot tell the difference between a judgment you should bet on and one you should question, because your confidence feels like a feeling rather than a measurement. The Calibrated Mind introduces the CALIBRATE system: nine skills that let you assess your own certainty, update it as evidence arrives, and act decisively without pretending the uncertainty is gone. The system Superforecasting proves is possible and Thinking, Fast and Slow proves is necessary.

- Page: https://lenvanderhof.com/books/the-calibrated-mind/
- Dutch edition: De gekalibreerde geest — https://lenvanderhof.com/boeken/the-calibrated-mind/
- Theme: Decisions & philosophy
- Framework: CALIBRATE (Calibrate confidence, Assess assumptions, Lean on probabilities, Investigate alternatives, Build updating rules, Record and review, Act with appropriate uncertainty, Teach and extend, Evolve your model)
- Length: 14 chapters, 424 pages in print
- ISBN (English paperback): 9798190866223
- Buy, English paperback: https://www.amazon.com/dp/B0HFWX8KSL?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-calibrated-mind&utm_content=paperback&tag=lenvanderho0f-20
- Buy, English hardcover: https://www.amazon.com/dp/B0HFWMV6YB?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-calibrated-mind&utm_content=hardcover&tag=lenvanderho0f-20
- Buy, Dutch paperback: https://www.amazon.nl/dp/B0HFZGQJ79?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-calibrated-mind&utm_content=paperback&tag=lenvanderho01-21
- Buy, Dutch hardcover: https://www.amazon.nl/dp/B0HFZHBR63?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-calibrated-mind&utm_content=hardcover&tag=lenvanderho01-21
- Availability: Available now

### The RAG Engineer: Retrieval Systems That Agents Can Actually Trust in Production

Embedding a PDF is a demo; surviving production is engineering. The RAG Engineer takes you past the tutorials into the failure modes that actually break retrieval (chunking, freshness, reranking, grounding, and silent drift), with a protocol for pipelines agents can depend on.

- Page: https://lenvanderhof.com/books/the-rag-engineer/
- Dutch edition: De RAG-engineer — https://lenvanderhof.com/boeken/the-rag-engineer/
- Theme: AI & agents
- Framework: GRAIN (Gather, Rank, Assemble, Inspect, Navigate)
- Length: 15 chapters, 514 pages in print
- ISBN (English paperback): 9798192483626
- Buy, English Kindle: https://www.amazon.com/dp/B0HGTMW9YG?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-rag-engineer&utm_content=kindle&tag=lenvanderho0f-20
- Buy, English paperback: https://www.amazon.com/dp/B0HGTTJ7QV?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-rag-engineer&utm_content=paperback&tag=lenvanderho0f-20
- Buy, English hardcover: https://www.amazon.com/dp/B0HGT628X9?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-rag-engineer&utm_content=hardcover&tag=lenvanderho0f-20
- Buy, Dutch Kindle: https://www.amazon.nl/dp/B0HGTDTH5R?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-rag-engineer&utm_content=kindle&tag=lenvanderho01-21
- Buy, Dutch paperback: https://www.amazon.nl/dp/B0HGTLRD13?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-rag-engineer&utm_content=paperback&tag=lenvanderho01-21
- Buy, Dutch hardcover: https://www.amazon.nl/dp/B0HGTB3935?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-rag-engineer&utm_content=hardcover&tag=lenvanderho01-21
- Kindle Unlimited: yes, free for KU members
- Availability: Available now

### The Model Portfolio: How to Route, Mix, and Govern LLMs Like a Strategic Asset Allocator

Choosing one model off the leaderboard is a bet you keep re-losing. The Model Portfolio treats your LLMs like a governed asset mix, routed and managed by cost, latency, privacy, and capability, so model choice becomes a strategy, not a guess.

- Page: https://lenvanderhof.com/books/the-model-portfolio/
- Dutch edition: Het model-portfolio — https://lenvanderhof.com/boeken/the-model-portfolio/
- Theme: AI & agents
- Framework: ROUTE (Register, Objective typing, Utilize policy, Track, Evolve)
- Length: 13 chapters, 490 pages in print
- ISBN (English paperback): 9798171745295
- Buy, English Kindle: https://www.amazon.com/dp/B0HHVQ48JQ?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-model-portfolio&utm_content=kindle&tag=lenvanderho0f-20
- Buy, English paperback: https://www.amazon.com/dp/B0HHWDFYS1?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-model-portfolio&utm_content=paperback&tag=lenvanderho0f-20
- Buy, English hardcover: https://www.amazon.com/dp/B0HHVZBYDP?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-model-portfolio&utm_content=hardcover&tag=lenvanderho0f-20
- Buy, Dutch Kindle: https://www.amazon.nl/dp/B0HHVQ2NHJ?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-model-portfolio&utm_content=kindle&tag=lenvanderho01-21
- Buy, Dutch paperback: https://www.amazon.nl/dp/B0HHWBBH5J?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-model-portfolio&utm_content=paperback&tag=lenvanderho01-21
- Buy, Dutch hardcover: https://www.amazon.nl/dp/B0HHXD2Q7X?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-model-portfolio&utm_content=hardcover&tag=lenvanderho01-21
- Kindle Unlimited: yes, free for KU members
- Availability: Available now

### The Verified Research Loop: Deep Online Research for Entrepreneurs Using AI Without Trusting It Blindly

AI returns summaries that feel researched but hide their sources. The Verified Research Loop is a discipline for operators: a five-step protocol that turns a fluent paste into linked records a stranger can check, and a decision memo that keeps evidence, inference, and action apart, so you can size a market or vet a deal before you bet on it.

- Page: https://lenvanderhof.com/books/the-verified-research-loop/
- Dutch edition: De geverifieerde research-loop — https://lenvanderhof.com/boeken/the-verified-research-loop/
- Theme: Evidence & learning
- Framework: TRACE (Target, Retrieve, Assess, Cross-check, Export)
- Length: 13 chapters, 330 pages in print
- ISBN (English paperback): 9798172832499
- Buy, English Kindle: https://www.amazon.com/dp/B0HJ9ZXGQ6?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-verified-research-loop&utm_content=kindle&tag=lenvanderho0f-20
- Buy, English paperback: https://www.amazon.com/dp/B0HJ9XKM89?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-verified-research-loop&utm_content=paperback&tag=lenvanderho0f-20
- Buy, English hardcover: https://www.amazon.com/dp/B0HJ9YMYTB?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-verified-research-loop&utm_content=hardcover&tag=lenvanderho0f-20
- Buy, Dutch Kindle: https://www.amazon.nl/dp/B0HJB2X5HY?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-verified-research-loop&utm_content=kindle&tag=lenvanderho01-21
- Buy, Dutch paperback: https://www.amazon.nl/dp/B0HJ9Y3RZX?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-verified-research-loop&utm_content=paperback&tag=lenvanderho01-21
- Buy, Dutch hardcover: https://www.amazon.nl/dp/B0HJB9M64L?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-verified-research-loop&utm_content=hardcover&tag=lenvanderho01-21
- Kindle Unlimited: yes, free for KU members
- Availability: Available now

### The Agent Dream Team: Build Your Autonomous Specialist Roster for a One-Person Company

A single chatbot mixes contexts and never hands off. The Agent Dream Team shows the solo operator how to build a coached roster of specialist agents (each with a charter, triggers, and a review gate), so you direct the team instead of doing every job yourself.

- Page: https://lenvanderhof.com/books/the-agent-dream-team/
- Dutch edition: Het agent-droomteam — https://lenvanderhof.com/boeken/the-agent-dream-team/
- Theme: AI & agents
- Framework: ROSTER (Roles, Objectives, Skills and tools, Triggers, Evaluation, Rotation)
- Length: 14 chapters, 484 pages in print
- ISBN (English paperback): 9798173121790
- Buy, English Kindle: https://www.amazon.com/dp/B0H852YJXJ?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agent-dream-team&utm_content=kindle&tag=lenvanderho0f-20
- Buy, English paperback: https://www.amazon.com/dp/B0HJDB6RT1?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agent-dream-team&utm_content=paperback&tag=lenvanderho0f-20
- Buy, English hardcover: https://www.amazon.com/dp/B0HJDGJLS5?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agent-dream-team&utm_content=hardcover&tag=lenvanderho0f-20
- Buy, Dutch Kindle: https://www.amazon.nl/dp/B0HH1BYBY1?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agent-dream-team&utm_content=kindle&tag=lenvanderho01-21
- Buy, Dutch paperback: https://www.amazon.nl/dp/B0HJDH7V6F?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agent-dream-team&utm_content=paperback&tag=lenvanderho01-21
- Buy, Dutch hardcover: https://www.amazon.nl/dp/B0HJDKN5B3?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-agent-dream-team&utm_content=hardcover&tag=lenvanderho01-21
- Kindle Unlimited: yes, free for KU members
- Availability: Available now

### The Evidence-Aware Life: How to Think in Levels Instead of Opinions

We do not suffer from a shortage of information, but from a shortage of ways to judge it. This book hands you the instrument: the Evidence Ladder to name the kind of support a claim really has, and the LEVEL check to choose a proportionate next step before you bet on it. It is a calibration manual, not a skepticism lecture, taught through real decisions across health, business, and AI. You will not believe fewer things. You will believe them better.

- Page: https://lenvanderhof.com/books/the-evidence-aware-life/
- Dutch edition: Het bewijsbewuste leven — https://lenvanderhof.com/boeken/the-evidence-aware-life/
- Theme: Evidence & learning
- Framework: LEVEL (Locate, Estimate, Value, Expose, Lean)
- Length: 13 chapters, 238 pages in print
- ISBN (English paperback): 9798174312623
- Buy, English Kindle: https://www.amazon.com/dp/B0HJSN8YXX?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-evidence-aware-life&utm_content=kindle&tag=lenvanderho0f-20
- Buy, English paperback: https://www.amazon.com/dp/B0HJSZVGLK?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-evidence-aware-life&utm_content=paperback&tag=lenvanderho0f-20
- Buy, English hardcover: https://www.amazon.com/dp/B0HJT9PMQ7?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-evidence-aware-life&utm_content=hardcover&tag=lenvanderho0f-20
- Buy, Dutch Kindle: https://www.amazon.nl/dp/B0HJSX6Y5C?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-evidence-aware-life&utm_content=kindle&tag=lenvanderho01-21
- Buy, Dutch paperback: https://www.amazon.nl/dp/B0HJSWPYNL?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-evidence-aware-life&utm_content=paperback&tag=lenvanderho01-21
- Buy, Dutch hardcover: https://www.amazon.nl/dp/B0HJT7MC5F?utm_source=lenvanderhof&utm_medium=buy_grid&utm_campaign=the-evidence-aware-life&utm_content=hardcover&tag=lenvanderho01-21
- Kindle Unlimited: yes, free for KU members
- Availability: Available now

## The rest of the library

- [ML for Agent Builders](https://lenvanderhof.com/books/ml-for-agent-builders/) [TRAIN] — Auto-Training, Fine-Tuning, and Eval Loops for LLM-Powered Systems. Status: Manuscript complete.
- [ML for Strategic Founders](https://lenvanderhof.com/books/ml-for-strategic-founders/) [SIGNAL] — Read Models, Not Math - Decision-Grade Machine Learning for Operators. Status: Manuscript complete.
- [Stoic Systems for Decision-Making](https://lenvanderhof.com/books/stoic-systems-for-decision-making/) [STOIC] — A Practical Framework for Clear Judgment Under Pressure. Status: Manuscript complete.
- [The Predictive Founder](https://lenvanderhof.com/books/the-predictive-founder/) [PREDICT] — When to Trust Your Gut, When to Override It, and How to Keep Score. Status: Manuscript complete.
- [Strategic Solitude](https://lenvanderhof.com/books/strategic-solitude/) — How Deliberate Alone Time Sharpens Decision-Making. Status: Manuscript complete.
- [The Org Memory](https://lenvanderhof.com/books/the-org-memory/) [ARCHIVE] — Build a Company That Remembers What It Learns. Status: Manuscript complete.
- [The AI Team Dynamics](https://lenvanderhof.com/books/the-ai-team-dynamics/) [SYNERGY] — Designing Mixed Human and Agent Teams So Productivity Compounds and Accountability Stays Human. Status: Manuscript complete.
- [The Second-Order Thinker](https://lenvanderhof.com/books/the-second-order-thinker/) [CASCADE] — Trace the Full Cascade of a Decision Before You Commit to It. Status: Manuscript complete.
- [The Identity Forge](https://lenvanderhof.com/books/the-identity-forge/) [FORGE] — The Operating System for Identity That Survives Growth. Status: Manuscript complete.
- [The Downstream Effects](https://lenvanderhof.com/books/the-downstream-effects/) [DOWNSTREAM] — Trace the Delayed Consequences of Your Decisions Before They Cost You. Status: Manuscript complete.
- [The Experimental Founder](https://lenvanderhof.com/books/the-self-experimenting-founder/) [TRIAL] — N=1 Protocols for Cognitive and Physical Performance. Status: Manuscript complete.
- [Marketing Without Manipulation](https://lenvanderhof.com/books/marketing-without-manipulation/) [TRUST] — How to Sell with Integrity in a World of Persuasion Tactics. Status: Manuscript complete.
- [The Human-in-the-Loop Life](https://lenvanderhof.com/books/the-human-in-the-loop-life/) [AGENCY] — How to Stay Essential in an Age of AI Automation. Status: Manuscript complete.
- [The Delegated Mind](https://lenvanderhof.com/books/the-delegated-mind/) [PROVISO] — Stoic Judgment in the Age of AI Delegation. Status: Manuscript complete.
- [The Attention Architect](https://lenvanderhof.com/books/the-attention-architect/) [PACE] — Reclaiming Cognitive Sovereignty by Delegating Distraction to AI. Status: Active development.
- [The Red Team Founder](https://lenvanderhof.com/books/the-red-team-founder/) [ASSAULT] — Adversarial Thinking for Operators Who Cannot Afford Surprise. Status: Manuscript complete.
- [The Constraint Founder](https://lenvanderhof.com/books/the-constraint-founder/) [FENCE] — Stoic Limits for Bootstrapped Ambition. Status: Manuscript complete.
- [The Cognitive Budget](https://lenvanderhof.com/books/the-cognitive-budget/) [LEDGER] — Spend Focus, Judgment, and Emotion Like Capital. Status: Manuscript complete.
- [The Judgment Gym](https://lenvanderhof.com/books/the-judgement-gym/) [REPS] — Maintain the Thinking That Makes You Worth Hiring When AI Thinks Faster. Status: Manuscript complete.
- [The Simulation Table](https://lenvanderhof.com/books/the-simulation-table/) [TABLE] — War Games for Teams Too Small for a Strategy Department. Status: Manuscript complete.
- [The Authentic Operator](https://lenvanderhof.com/books/the-authentic-operator/) [VOICE] — Voice, Identity, and Trust When AI Writes With You. Status: Manuscript complete.
- [AI Agents for Personal Health](https://lenvanderhof.com/books/ai-agents-for-personal-health/) [AGENT] — Deploy Intelligent Systems to Monitor, Analyze, and Optimize Your Biology. Status: Manuscript complete.
- [The Founder CFO](https://lenvanderhof.com/books/the-founder-cfo/) — Cash, Metrics, and Calm Between Rounds. Status: Manuscript complete.
- [The Founder Sales System](https://lenvanderhof.com/books/the-founder-sales-system/) [SALES] — Pipeline, Discovery, and Deals Before You Have a Sales Team. Status: Manuscript complete.
- [The Findability Engine](https://lenvanderhof.com/books/the-findability-engine/) [CITATION] — How to Be Cited, Trusted, and Discovered in the Age of AI Search. Status: Drafting (English edition first).
- [The Negotiation Operator](https://lenvanderhof.com/books/the-negotiation-operator/) [NEGOTIATE] — Evidence-Based Deal-Making for Founders. Status: Manuscript complete.
- [Proof in Public](https://lenvanderhof.com/books/the-proof-engine/) [PROVE] — Trust Velocity for Evidence-First Founders. Status: Drafting (English edition first).
- [The Fundraising Operator](https://lenvanderhof.com/books/the-fundraising-operator/) [CLAUSE] — Term Sheets, Power, and Calm Under Dilution. Status: Manuscript complete.
- [Decision-Making Under Chaos](https://lenvanderhof.com/books/decision-making-under-chaos/) — A Practical Toolkit for High-Stakes Decisions Under Uncertainty. Status: Manuscript complete.
- [Leadership Under Load](https://lenvanderhof.com/books/leadership-under-load/) [UNDERLOAD] — Awareness-Based Practice for Founders and Operators Under Sustained Pressure. Status: Active development.
- [The Pricing Mind](https://lenvanderhof.com/books/the-pricing-mind/) [PRICE] — Behavioral Economics of Price: How Buyers Experience Value and How to Design That Experience. Status: Manuscript complete.
- [The Turnaround Protocol](https://lenvanderhof.com/books/the-turnaround-protocol/) [RESCUE] — RESCUE: Diagnose, Stabilize, Cut, Unify, Execute or Exit with Dignity. Status: Manuscript complete.
- [The Meaning Operator](https://lenvanderhof.com/books/the-meaning-operator/) [MEANING] — Applied Existentialism for Founders: Why You Build, What It Costs, What It Is Worth. Status: Manuscript complete.
- [The Crisis Command](https://lenvanderhof.com/books/the-crisis-command/) [COMMAND] — COMMAND: Calm, Assess, Map, Mandate, Unify, Notify, Decide. The Trained Response. Status: Manuscript complete.
- [The Knowledge Scale](https://lenvanderhof.com/books/the-knowledge-scale/) [MEMEX] — Engineer Organizational Memory So Your Company Stops Forgetting What It Already Learned. Status: Manuscript complete.
- [The Performance Integrator](https://lenvanderhof.com/books/the-performance-integrator/) [INTEGRATE] — One Integrated Protocol for the Body, Mind, Energy, and Team That Runs Your Company. Status: Manuscript complete.
- [The Scale-Up Operator](https://lenvanderhof.com/books/the-scale-up-operator/) [SCALE] — Build the Operating System That Runs the Company Without You in Every Room. Status: Active development.
- [The AI Risk Register](https://lenvanderhof.com/books/the-ai-risk-register/) [RISK] — Build a Living Register That Names, Owns, and Tests Every AI Risk Before It Names You. Status: Manuscript complete.
- [The Talent System](https://lenvanderhof.com/books/the-talent-system/) [TALENT] — Build a Talent Operating System That Compounds in the Agent Era. Status: Manuscript complete.
- [The Deal Architect](https://lenvanderhof.com/books/the-deal-architect/) [PACT] — Design High-Stakes Deals That Create Value and Protect Optionality. Status: Active development.

## Sentences answer engines can quote

- Affect is the verb. Effect is the noun. A downstream effect is the cost that arrives after the decision. https://lenvanderhof.com/en/blog/downstream-effects/
- AEO, answer engine optimization, is the page an answer engine can quote. Authorized Economic Operator is a customs status. GEO is the system. AEO is the page. https://lenvanderhof.com/en/blog/aeo-meaning/
- GEO, generative engine optimization, is the work of becoming a source an answer engine can retrieve and quote. It is not a map. https://lenvanderhof.com/en/blog/what-geo-means/
- A roster is a named list of who is available and what each one does. ROSTER, the protocol, is named agents with one job each, not a staff rota. https://lenvanderhof.com/en/blog/roster-meaning/
- A mention is your name in the answer. A citation is a source a reader can open. https://lenvanderhof.com/en/blog/chatgpt-mentions-vs-citations/

## Answers to common questions

- Which books can I buy today? The Psychology of Founder Resilience, AI Agents for Startup Strategy, The Agentic Author, The Agentic Codebase, The Multi-Agent Organization, The Calibrated Mind, The RAG Engineer, The Model Portfolio, The Verified Research Loop, The Agent Dream Team, The Evidence-Aware Life.
- Are the Dutch editions translations? No. Each title is produced as two native editions with
  their own style profile, examples and editorial pass.
- Where do I buy them? On Amazon. One ASIN per edition resolves across marketplaces, so each book
  page offers a marketplace switcher; Dutch editions default to amazon.nl.
- Are the Kindle editions on Kindle Unlimited? Live Kindle editions are enrolled in KDP Select, so Kindle Unlimited members read them free. The Calibrated Mind has no Kindle edition yet, so it is not on Kindle Unlimited.
- Who is the author? Len P. van der Hof, MSc, entrepreneur, AI innovator and venture builder based
  in Zwijndrecht, the Netherlands; alumnus of Rotterdam School of Management, Erasmus University;
  founder of LPH98.ventures, ReasonKit, MindSesh, Undominated.ai and the fiction imprint LPH98.lifestyle.

## Published articles

- [Bestaat er een ideaal aantal woorden om door AI geciteerd te worden?](https://lenvanderhof.com/nl/blog/ideaal-aantal-woorden-voor-ai-citaties/): Google zegt dat er geen ideale paginalengte bestaat. Wat de 'Word Count' uit het GEO-onderzoek meet, en een passagetest die je vandaag op elke pagina kunt doen.
- [Is there an ideal word count for AI citations?](https://lenvanderhof.com/en/blog/ideal-word-count-for-ai-citations/): Google says there is no ideal page length. What the GEO study's 'Word Count' measures, and a passage test you can run on any page today.
- [Zo houd je een pre-mortem in 20 minuten](https://lenvanderhof.com/nl/blog/pre-mortem-in-20-minuten/): Houd een pre-mortem van 20 minuten: ga ervan uit dat het plan al mislukt is en leg uit waarom. Kopieer de agenda, geef elk risico een eigenaar en lees waar het bewijs ophoudt.
- [How to run a pre-mortem in 20 minutes](https://lenvanderhof.com/en/blog/how-to-run-a-pre-mortem/): Run a 20-minute pre-mortem: assume the plan has already failed and explain why. Copy the agenda, give each risk an owner, and see where the evidence stops.
- [Nepurgentie: zo onderscheid je eerlijke schaarste van een valse deadline](https://lenvanderhof.com/nl/blog/nepurgentie/): Nepurgentie is een valse deadline of voorraadclaim. De EU verbiedt valse 'zeer beperkte tijd'-claims; volgens de ACM moeten timers en voorraadmeldingen kloppen.
- [Fake urgency: how to tell honest scarcity from a false deadline](https://lenvanderhof.com/en/blog/fake-urgency/): Fake urgency is a false deadline or stock claim. EU law blacklists false 'very limited time' claims; the Dutch ACM says timers and stock counts must be true.
- [Een AEO-bureau inhuren? Zeven vragen voordat je tekent](https://lenvanderhof.com/nl/blog/aeo-bureau-inhuren/): Zeven vragen voordat je een AEO-bureau inhuurt: hoe goede en slechte antwoorden eruitzien, de alarmsignalen en wat het contract moet vastleggen.
- [Hiring an AEO agency? Seven questions to ask before you sign](https://lenvanderhof.com/en/blog/hiring-an-aeo-agency/): Seven questions to ask before you hire an AEO agency, what good and bad answers look like, the red flags, and the deliverables to put in the contract.
- [Waarom multi-agent-LLM-systemen falen: de 14 MAST-faalwijzen](https://lenvanderhof.com/nl/blog/multi-agent-faalwijzen-uitgelegd/): Het MAST-onderzoek labelde 1642 traces en vond 14 faalwijzen. Houd de fout van een agent en de werkstroom die hem doorliet uit elkaar, en test één mislukte run.
- [Why multi-agent LLM systems fail: the 14 MAST failure modes](https://lenvanderhof.com/en/blog/multi-agent-failure-modes-explained/): The MAST study labelled 1,642 traces and found 14 failure modes. Separate the agent's mistake from the workflow that let it pass, then test one failed run.
- [Planning fallacy: een plan is geen voorspelling](https://lenvanderhof.com/nl/blog/planningsdrogreden/): Planning fallacy (planningsdrogreden): je rekent op je plan, hoewel je weet dat vergelijkbare projecten uitliepen. Het onderzoek, de grenzen ervan en een werkblad.
- [The planning fallacy: a plan is not a forecast](https://lenvanderhof.com/en/blog/planning-fallacy/): The planning fallacy is expecting your project to go to plan when you know similar ones ran late. The research, its limits, and a worksheet to test a deadline.
- [Het cobraverhaal is onbevestigd. De rattenpremie van Hanoi is gedocumenteerd.](https://lenvanderhof.com/nl/blog/cobra-effect/): Het cobra-effect: een beloning die het op te lossen probleem verergert. Zet het onbevestigde verhaal uit Delhi naast de rattenpremie van Hanoi en test je prikkels.
- [The cobra story is unverified. Hanoi's rat bounty is documented.](https://lenvanderhof.com/en/blog/cobra-effect/): The cobra effect: a reward that worsens the problem it pays to fix. Weigh Delhi's unverified story against Hanoi's rat-tail bounty, then test your incentives.
- [utm_source=chatgpt.com: zo zie je verkeer uit ChatGPT en andere AI-assistenten in GA4](https://lenvanderhof.com/nl/blog/ai-verkeer-zien-in-ga4/): ChatGPT zet utm_source=chatgpt.com op zijn links. Waar je dat in GA4 terugziet, een regex voor AI-bronnen om te kopiëren en wat GA4 niet kan zien.
- [utm_source=chatgpt.com: how to see ChatGPT and other AI traffic in GA4](https://lenvanderhof.com/en/blog/see-ai-traffic-in-ga4/): ChatGPT tags its links with utm_source=chatgpt.com. Where that shows up in GA4, a copy-paste regex for AI sources, and what GA4 cannot see.
- [Stoïsch of stoïcijns? De filosofie ging nooit over gevoelens wegdrukken](https://lenvanderhof.com/nl/blog/stoisch-of-stoicijns/): Stoïsch of stoïcijns? Allebei goed. Het echte verschil zit tussen een strak gezicht en een filosofie die je oordeel toetst. Met een oefening van tien minuten.
- [Being stoic is not Stoicism: the philosophy was never about hiding your feelings](https://lenvanderhof.com/en/blog/stoic-vs-stoicism/): Stoic vs Stoicism: everyday stoic means showing no feeling. The Stoics taught you to test the judgment behind a feeling, not to hide it. A 10-minute exercise.
- [Stroomopwaarts en stroomafwaarts: één rivierbeeld voor olie, toeleveringsketens, software en besluiten](https://lenvanderhof.com/nl/blog/stroomopwaarts-en-stroomafwaarts/): Stroomafwaarts betekent later in de stroom, stroomopwaarts dichter bij de bron. Eén rivierbeeld, zes vakgebieden in een tabel en waarom je positie telt.
- [Upstream vs downstream: one river picture that explains oil, supply chains, software, and decisions](https://lenvanderhof.com/en/blog/upstream-vs-downstream/): Upstream vs downstream: upstream is nearer the source, downstream later in the flow. One river picture, a six-field table, and why it depends where you stand.
- [GEO-voorbeelden: zes voor-en-na-aanpassingen die je pagina voor AI beter citeerbaar maken](https://lenvanderhof.com/nl/blog/geo-voorbeelden/): Zes GEO-voorbeelden met een voor- en een na-versie: kleine aanpassingen waardoor ChatGPT, Perplexity en Google je pagina makkelijker vinden en citeren.
- [GEO examples: six before-and-after edits that make a page easier for AI to quote](https://lenvanderhof.com/en/blog/geo-examples/): Six GEO examples, each with a before and an after: small edits that make a page easier for ChatGPT, Perplexity and Google's AI answers to find and quote.
- [Nederlands kost 20% meer GPT-tokens dan Engels. We maten het in 78 artikelparen](https://lenvanderhof.com/nl/blog/nederlands-kost-meer-tokens/): Nederlands kost meer tokens dan Engels: in 78 artikelparen had Nederlands 20% meer tokens nodig bij de tokenizer van OpenAI, en 49% meer voor embeddings.
- [Dutch costs 20% more GPT tokens than English. We measured it on 78 article pairs](https://lenvanderhof.com/en/blog/dutch-costs-more-tokens/): Dutch vs English token count, measured on 78 article pairs with OpenAI's tokenizer: Dutch needed 20% more tokens, and 49% more for embeddings.
- [De 90%-test: waarom 'ik weet het zeker' veel vaker misgaat dan je denkt](https://lenvanderhof.com/nl/blog/de-90-procent-test/): Tien vragen, één regel: geef een marge waarvan je 90% zeker bent. De meeste marges zijn veel te smal. Doe de quiz, tel je score en leer ze eerlijk verbreden.
- [The 90% test: why 'I'm sure' is right far less often than it feels](https://lenvanderhof.com/en/blog/overconfidence-90-percent-test/): Ten questions, one rule: give a range you are 90% sure of. Most people's ranges are far too narrow. Take the quiz, score it, and learn to widen ranges honestly.
- [ChatGPT, Perplexity, Google AI Overviews: three engines, three doors to your page](https://lenvanderhof.com/en/blog/three-answer-engines-three-doors/): ChatGPT search, Perplexity and Google AI Overviews reach your page through different crawlers and switches. A table from official docs and a 15-minute check.
- [ChatGPT, Perplexity en de AI-overzichten van Google: drie machines, drie deuren naar je pagina](https://lenvanderhof.com/nl/blog/drie-antwoordmachines-drie-deuren/): ChatGPT, Perplexity en Googles AI-overzichten bereiken je pagina via eigen crawlers en schakelaars. Eén tabel uit officiële bronnen en een check van 15 minuten.
- [Heb je een literair agent nodig voor non-fictie? Wat een agent doet, wat het kost en wanneer je zelf uitgeeft](https://lenvanderhof.com/nl/blog/literair-agent-nodig-voor-non-fictie/): Heb je een literair agent nodig voor non-fictie? Wat een agent doet, de gangbare commissie van 15% en 20%, de Nederlandse markt en wat zelf uitgeven vraagt.
- [Do you need a literary agent for nonfiction? What agents do, what they cost, and when to publish yourself](https://lenvanderhof.com/en/blog/do-you-need-a-literary-agent/): Do you need a literary agent for nonfiction? What agents do, the usual 15% and 20% commissions, the Dutch market, and what self-publishing asks of you.
- [Faalt 90% van de startups? Zo check je een viraal cijfer in tien minuten](https://lenvanderhof.com/nl/blog/faalt-90-procent-van-startups/): Faalt 90% van de startups? Cijfers van CBS en BLS zeggen van niet. Waar het getal vandaan komt, wat falen betekent en een check van tien minuten.
- [Do 90% of startups fail? How to fact-check a viral statistic in ten minutes](https://lenvanderhof.com/en/blog/do-90-percent-of-startups-fail/): Do 90% of startups fail? US and Dutch government data say no. Where the number comes from, what 'fail' means, and a ten-minute method to check any statistic.
- [Is SEO dood? Nee. Antwoordmachines hebben er een toegangskaartje van gemaakt](https://lenvanderhof.com/nl/blog/is-seo-dood/): Is SEO dood? Nee. Google, OpenAI en Perplexity moeten je pagina nog steeds crawlen en Google moet haar indexeren, wil hun AI haar citeren. Veranderd is de klik.
- [Is SEO dead? No. Answer engines turned it into the entry ticket](https://lenvanderhof.com/en/blog/is-seo-dead/): Is SEO dead? No. Google, OpenAI and Perplexity still need to crawl your page, and Google to index it, before their AI can quote it. What changed is the click.
- [MCP is geen USB-C: waar de geliefde vergelijking mank gaat en waarom dat ertoe doet](https://lenvanderhof.com/nl/blog/mcp-is-geen-usb-c/): Wat MCP betekent, waarom de eigen documentatie het met USB-C vergelijkt en op welke vier punten dat beeld mank gaat. Plus een checklist voor het koppelen.
- [MCP is not USB-C: where the favourite analogy breaks, and why that matters](https://lenvanderhof.com/en/blog/mcp-is-not-usb-c/): What MCP means, why its own docs compare it to USB-C, and the four places that analogy breaks. Plus a checklist before you plug in an MCP server.
- [Downstream effects of affects? De regel in één zin, plus twee uitzonderingen](https://lenvanderhof.com/nl/blog/downstream-effects-of-affects/): Downstream effects of affects? Effect is het zelfstandig naamwoord, affect het werkwoord. De regel, de twee uitzonderingen en een quiz van vijf zinnen.
- [Downstream effects or affects? The one-line rule, and the two exceptions](https://lenvanderhof.com/en/blog/downstream-effects-or-affects/): Downstream effects or affects? Effect is the noun, affect is the verb. The one-line rule, the two real exceptions, and a five-sentence quiz.
- [MCP-OAuth is geen toolautorisatie: een geldig token is nog geen ja](https://lenvanderhof.com/nl/blog/mcp-oauth-is-geen-toolautorisatie/): Een geldig OAuth-token laat een client je MCP-server in, maar bepaalt niet wat een tool met welk record mag. Wat de specificatie eist en wat je moet toevoegen.
- [MCP OAuth is not tool authorization: a valid token is not a yes](https://lenvanderhof.com/en/blog/mcp-oauth-is-not-tool-authorization/): A valid OAuth token gets a client into your MCP server. It does not decide which tool may act on which record. What the spec checks and what you must add.
- [Hoe je een productie-agent buiten gebruik stelt (stoppen is niet genoeg)](https://lenvanderhof.com/nl/blog/productie-agent-buiten-gebruik-stellen/): Stel een AI-agent in productie buiten gebruik: sluit triggers, sleutels, aanroepers en lopende kosten af en test elke oude route. Eén checklist, één ticket.
- [How to retire a production agent (stopping it is not enough)](https://lenvanderhof.com/en/blog/retire-a-production-agent/): Retire a production AI agent by closing its triggers, keys, callers and recurring costs, then testing every old path. A checklist you can run in one ticket.
- [Wat betekent 'waarschijnlijk'? Vraag het vijf mensen en je krijgt vijf getallen](https://lenvanderhof.com/nl/blog/wat-betekent-waarschijnlijk/): 'Waarschijnlijk' heeft geen vast getal: in een Nederlands onderzoek kozen negen op de tien iets tussen 57 en 91 procent. Waar dat misgaat, en wat wel helpt.
- [What does 'likely' mean? Ask five people and you get five numbers](https://lenvanderhof.com/en/blog/what-does-likely-mean/): 'Likely' has no fixed number: most people put it somewhere between about 55% and 90%. Where that spread comes from, what it costs a team, and the fix.
- [Wat is ReasonKit? Gestructureerde beslissingen via MCP](https://lenvanderhof.com/nl/blog/wat-is-reasonkit/): ReasonKit MCP uitgelegd: wat ReasonKit Think doet, hoe het via MCP aansluit op AI-assistenten, een uitgewerkt besluit, wat het niet is en de installatiestatus.
- [What is ReasonKit? Structured decisions through MCP](https://lenvanderhof.com/en/blog/what-is-reasonkit/): ReasonKit MCP explained: what ReasonKit Think does, how it plugs into AI assistants through MCP, a worked decision, what it is not, and its install status.
- [Agent-schaduwmodus: beoordeel voorstellen voordat je actierechten verleent](https://lenvanderhof.com/nl/blog/agent-schaduwmodus/): Agent-schaduwmodus: de agent stelt voor, een mens voert uit, jij vergelijkt. Wat je meet, hoeveel gevallen je nodig hebt en een promotieregel om over te nemen.
- [Agent shadow mode: evaluate proposals before granting action rights](https://lenvanderhof.com/en/blog/agent-shadow-mode/): Agent shadow mode: the agent proposes, a person still acts, and you compare. What to measure, how many cases you need, and a promotion rule you can copy.
- [Wat is citeturn0search0? Het ChatGPT-restje dat verraadt dat een antwoord zonder bronnen is gekopieerd](https://lenvanderhof.com/nl/blog/wat-is-citeturn0search0/): citeturn0search0 blijft achter als een geplakt ChatGPT-antwoord de link naar zijn bron verliest. Wat de markering betekent, wat ze bewijst en wat je eraan doet.
- [What is citeturn0search0? The ChatGPT leftover that shows an answer was copied without its sources](https://lenvanderhof.com/en/blog/what-is-citeturn0search0/): citeturn0search0 is a ChatGPT citation marker left behind when an answer is pasted without its link. What it means, what it proves, and how to fix it.
- [Agent-runtimebudget: houd retry's binnen hetzelfde plafond](https://lenvanderhof.com/nl/blog/agent-runtimebudget/): Een runtimebudget voor agents begrenst tokens, geld, tijd en toolaanroepen per taak. Retry's vermenigvuldigen de kosten. Zo houd je ze onder één plafond.
- [Agent runtime budget: keep retries inside the same ceiling](https://lenvanderhof.com/en/blog/agent-runtime-budget/): An agent runtime budget caps tokens, money, time and tool calls per task. Why retries multiply cost, and how to keep them inside one ceiling.
- [Agent-audittrail: reconstrueer de handeling, goedkeuring en uitkomst](https://lenvanderhof.com/nl/blog/agent-audittrail/): Wie er niet bij was, reconstrueert met een agent-audittrail wat een AI-agent deed, wie akkoord gaf en wat er gebeurde. De velden en een oefening van 20 minuten.
- [Agent audit trail: reconstruct the action, approval, and outcome](https://lenvanderhof.com/en/blog/agent-audit-trail/): An agent audit trail lets someone who was not there reconstruct what an AI agent did, who approved it, and what happened. The fields, and a 20-minute drill.
- [Downstream impact, consequentie, implicatie, uitkomst, effect: vijf woorden, vijf taken](https://lenvanderhof.com/nl/blog/downstream-impact-betekenis/): Wat downstream impact betekent, in gewone taal: het verschil tussen impact, consequentie, implicatie, uitkomst en effect, toegepast op één besluit.
- [Downstream impact, consequence, implication, outcome, effect: five words, five jobs](https://lenvanderhof.com/en/blog/downstream-impact-meaning/): Downstream impact meaning in plain words: how impact, consequence, implication, outcome and effect differ, with one decision traced through all five.
- [Kalibratie of calibratie? De spelling, de betekenis en het verschil met ijken](https://lenvanderhof.com/nl/blog/kalibratie-of-calibratie/): Je schrijft kalibratie, met een k. Wat kalibreren betekent, waarom het iets anders is dan justeren en hoe het verschilt van ijken, de wettelijke keuring.
- [What does calibrate mean? Compare, record, then decide: calibration is not adjustment](https://lenvanderhof.com/en/blog/calibrate-meaning/): To calibrate means to compare a reading with a trusted reference and record the difference. Adjusting the instrument is a separate step, and here is why.
- [Technische GEO: het leidingwerk dat bepaalt of AI je pagina kan gebruiken](https://lenvanderhof.com/nl/blog/technische-geo/): Technische GEO is het ontwikkelwerk in generative engine optimization: crawlertoegang, leesbare HTML, indexering, snippetregels, datum. Checklist in 7 punten.
- [Technical GEO: the plumbing that decides whether AI can use your page](https://lenvanderhof.com/en/blog/technical-geo/): Technical GEO is the developer side of generative engine optimization: crawler access, readable HTML, indexing, snippet rules, dates. A 7-point checklist.
- [STACK versus AGENTS.md: waar instructies ophouden en harde controles beginnen](https://lenvanderhof.com/nl/blog/stack-versus-agents-md/): STACK versus AGENTS.md: AGENTS.md vertelt codeeragents hoe je repo werkt, maar dwingt niets af. Welke regels erin horen en welke om een harde controle vragen.
- [STACK vs AGENTS.md: where instructions end and controls begin](https://lenvanderhof.com/en/blog/stack-vs-agents-md/): STACK vs AGENTS.md: AGENTS.md tells coding agents how your repo works but enforces nothing. Which rules belong in it, and which need a real control.
- [CHORUS en Chorus AIDLC: welke bedoel je?](https://lenvanderhof.com/nl/blog/chorus-naamcollision/): CHORUS versus Chorus-harness: CHORUS is een coördinatieprotocol van zes stappen uit een boek. Chorus van Chorus-AIDLC is opensourcesoftware, een agentharness.
- [CHORUS and Chorus AIDLC: which one do you mean?](https://lenvanderhof.com/en/blog/chorus-name-collision/): CHORUS vs Chorus harness: CHORUS is a six-step coordination protocol from a book. Chorus from Chorus-AIDLC is open-source agent-harness software.
- [Wat is een roster? En waarom het net iets anders is dan een rooster](https://lenvanderhof.com/nl/blog/wat-is-een-roster/): Wat is een roster? Een lijst van wie er beschikbaar is en wat ieder doet. Waarom dat net iets anders is dan een rooster, en wat een roster van AI-agents is.
- [What is a roster? From the Monday rota to a team of AI agents](https://lenvanderhof.com/en/blog/roster-meaning/): What is a roster? A named list of who is available and what each one does. Staff, sports, agency and AI agent rosters, and the Dutch grill behind the word.
- [Bewijs dat je een agent kunt stoppen: onderbreek de run en trek de toegang in](https://lenvanderhof.com/nl/blog/bewijs-dat-je-een-agent-kunt-stoppen/): Wie kan de agent stoppen? Een benoemde persoon moet run, wachtrij en toegang binnen minuten stilleggen, met bewijs per laag. Met een oefening voor deze week.
- [Prove you can stop an agent: interrupt the run and revoke access](https://lenvanderhof.com/en/blog/prove-you-can-stop-an-agent/): Who can stop the agent? A named person must halt the run, the queue, and its access within minutes, and prove each one. A stop drill you can run this week.
- [Agentgeheugen met regels: wat mag blijven en wie mag erin schrijven](https://lenvanderhof.com/nl/blog/bestuurde-agent-memory/): Agentgeheugen met regels: wat een AI-agent tussen taken bewaart, waarom geheugen zonder regels veroudert of vergiftigd raakt, en een beleid op één pagina.
- [Governed agent memory: what may persist, and who may write it](https://lenvanderhof.com/en/blog/governed-agent-memory/): Governed agent memory explained: what an AI agent keeps between runs, why unmanaged memory goes stale or gets poisoned, and a one-page memory policy.
- [Len of een fractional CAIO: welk werk heb je nodig?](https://lenvanderhof.com/nl/blog/len-versus-fractional-caio/): Fractional CAIO of AI-adviseur: wat een parttime chief AI officer meestal doet, wat het werk van Len wel en niet dekt, en een zelftest in zes vragen.
- [Len or a fractional CAIO: which work do you need?](https://lenvanderhof.com/en/blog/len-vs-fractional-caio/): Fractional CAIO vs AI advisor: what a part-time chief AI officer usually owns, what Len's advisory and systems work covers, and a six-question self-test.
- [ChatGPT-citaties meten: een wekelijks logboek, geen screenshot](https://lenvanderhof.com/nl/blog/citatie-meetlog/): Meet ChatGPT-citaties met vaste vragen, de zoekfunctie aan en per run één regel: genoemd of niet, gelinkt of niet. Met sjabloon, afspraken en rekenvoorbeeld.
- [How to measure ChatGPT citations: a weekly log, not a screenshot](https://lenvanderhof.com/en/blog/citation-measurement-log/): To measure ChatGPT citations, ask fixed questions with search on and log each run: were you named, were you linked? Template, rules and a worked example.
- [Is één AI-model genoeg? Wanneer één model volstaat en wanneer je een portfolio nodig hebt](https://lenvanderhof.com/nl/blog/een-model-is-geen-portefeuille/): Eén AI-model volstaat bij weinig volume, één soort taak en geen gevoelige data. Anders heb je een portfolio nodig: vastgelegde keuzes, fallback en reviewdatum.
- [Is one AI model enough? When a single model works, and when you need a portfolio](https://lenvanderhof.com/en/blog/one-model-is-not-a-portfolio/): One AI model is fine for low-volume, single-task work with no sensitive data. Past that, you need a portfolio: written choices, a fallback, and a review date.
- [Wat is hybride zoeken? Twee indexen, één inspecteerbare treffer](https://lenvanderhof.com/nl/blog/wat-is-hybride-zoeken/): Hybride zoeken draait een lexicale index en een vectorindex over hetzelfde corpus en fuseert de twee ranglijsten voordat het model schrijft. Een vector store alleen is niet hybride. Kun je de passage die het model gebruikte niet openen, dan heb je nog steeds een vloeiende gok.
- [What is hybrid search? Two indexes, one inspectable hit](https://lenvanderhof.com/en/blog/what-is-hybrid-search/): Hybrid search runs a lexical index and a vector index over the same corpus and fuses the two rankings before the model writes. A vector store alone is not hybrid. If you cannot open the passage the model used, you still have a fluent guess.
- [Wat is een zekerheidsgraad? Waarom "hoge zekerheid" geen 90 procent is](https://lenvanderhof.com/nl/blog/wat-is-een-zekerheidsgraad/): Een zekerheidsgraad zegt hoe goed het bewijs één bewering onderbouwt. Het is geen kans en geen groen licht. Zo houden het IPCC, GRADE en analisten ze apart.
- [What is a confidence grade? Why "high confidence" is not 90%](https://lenvanderhof.com/en/blog/what-is-a-confidence-grade/): A confidence grade rates how well the evidence supports one claim. It is not a probability or a go-ahead. How the IPCC, GRADE and analysts keep them apart.
- [Wanneer geen AI-agent gebruiken? Vijf signalen en wat je in plaats daarvan inzet](https://lenvanderhof.com/nl/blog/wanneer-geen-agent-gebruiken/): Sla de AI-agent over als de stappen vastliggen, één prompt volstaat of een fout onomkeerbaar is. Vijf signalen, alternatieven en een test van 15 minuten.
- [When not to use an AI agent: five signs, and what to use instead](https://lenvanderhof.com/en/blog/when-not-to-use-an-agent/): Skip the AI agent when the steps are known, one prompt does the job, or a mistake cannot be undone. Five signs, what to use instead, and a 15-minute test.
- [TRACE versus RAG: de een haalt de tekst op, de ander toetst wat die bewijst](https://lenvanderhof.com/nl/blog/trace-versus-rag/): RAG is software die tekst ophaalt voor een AI-antwoord. TRACE is een onderzoeksmethode die vóór een besluit toetst wat bronnen bewijzen. Verschil en samenhang.
- [TRACE vs RAG: one fetches the text, the other checks what it proves](https://lenvanderhof.com/en/blog/trace-vs-rag/): RAG is software that fetches text for an AI answer. TRACE is a research method that checks what sources prove before a decision. How they differ and fit.
- [Wat is een gedomineerd AI-model? Beter en goedkoper is maar de halve toets](https://lenvanderhof.com/nl/blog/wat-is-een-gedomineerd-model/): Een AI-model is gedomineerd als een alternatief minstens zo hoog scoort voor niet meer geld. Wat dat betekent, twee echte voorbeelden en de controle vooraf.
- [What is a dominated AI model? Better and cheaper is only half the test](https://lenvanderhof.com/en/blog/what-is-a-dominated-model/): A dominated AI model has an alternative that scores at least as high for no more money. What that means, two real examples, and the check before you switch.
- [ROSTER versus CHORUS: AI-agents voor één persoon of voor een team?](https://lenvanderhof.com/nl/blog/roster-versus-chorus/): ROSTER is voor een solo-oprichter met AI-agents, CHORUS voor teams van 2 tot 15 mensen met agents. Hoe je kiest, waar ze overlappen en wat je eerst repareert.
- [ROSTER vs CHORUS: AI agents for one person, or for a team?](https://lenvanderhof.com/en/blog/roster-vs-chorus/): ROSTER runs AI agents for a founder working alone. CHORUS coordinates teams of 2 to 15 people and agents. How to choose, where they overlap, what to fix first.
- [Wat is ROSTER? Zo stuur je in je eentje een team van AI-agents aan](https://lenvanderhof.com/nl/blog/wat-is-roster/): ROSTER legt in zes besluiten vast hoe je AI-agents als benoemde specialisten inzet in een eenmansbedrijf. Met één uitgewerkte rol en een check van 15 minuten.
- [What is ROSTER? How one person manages a team of AI agents](https://lenvanderhof.com/en/blog/what-is-roster/): ROSTER is a six-decision framework for running AI agents as named specialists in a one-person business: plain definitions, a worked role, a 15-minute check.
- [LEVEL versus CALIBRATE: toets je de bewering of je eigen oordeel?](https://lenvanderhof.com/nl/blog/level-versus-calibrate/): LEVEL toetst één bewering vóór één beslissing. CALIBRATE toetst over veel beslissingen of je zekerheid klopt met wat er gebeurt. Zo verschillen ze.
- [LEVEL vs CALIBRATE: check the claim, or check your own judgment](https://lenvanderhof.com/en/blog/level-vs-calibrate/): LEVEL checks one claim before one decision. CALIBRATE checks, over many decisions, whether your confidence matches what happens. How they differ and fit.
- [Wat is TRACE? In vijf stappen onderzoek controleren voordat je handelt](https://lenvanderhof.com/nl/blog/wat-is-trace/): TRACE is een controle in vijf stappen: Target, Retrieve, Assess, Cross-check, Export. Wat elke stap inhoudt, een uitgewerkt voorbeeld en een kaart voor vandaag.
- [What is TRACE? Five steps to check research before you act on it](https://lenvanderhof.com/en/blog/what-is-trace/): TRACE is a five-step research check: Target, Retrieve, Assess, Cross-check, Export. What each step means, a worked example, and a card to use today.
- [MCP versus function calling: twee lagen van dezelfde toolaanroep](https://lenvanderhof.com/nl/blog/mcp-versus-function-calling/): Function calling is hoe een model om een tool vraagt. MCP is een standaardroute om tools bij de app te krijgen. Verschil, voorbeeld en wanneer MCP loont.
- [MCP vs function calling: two layers of the same tool call](https://lenvanderhof.com/en/blog/mcp-vs-function-calling/): Function calling is how a model asks to use a tool. MCP is a standard route for getting tools to the app. The difference, a worked example, and when MCP pays.
- [Wat is LEVEL? Een check in vijf stappen voordat je op een bewering afgaat](https://lenvanderhof.com/nl/blog/wat-is-level/): LEVEL is een check in vijf stappen voor elke bewering waar je op wilt handelen: Locate, Estimate, Value, Expose, Lean. Met uitleg, voorbeeld en een kaart.
- [What is LEVEL? A five-step check before you act on a claim](https://lenvanderhof.com/en/blog/what-is-level/): LEVEL is a five-step check for any claim you might act on: Locate, Estimate, Value, Expose, Lean. A plain definition, a worked example and a card to copy.
- [Wat is een modelcascade? Escaleren met een stopregel](https://lenvanderhof.com/nl/blog/wat-is-een-modelcascade/): Een modelcascade is een gepland kostenpatroon: begin bij het goedkoopste model dat de klus aankan, escaleer op een geschreven trigger, stop op een geschreven regel. Het is geen fallback, geen retry-lus en niet het CASCADE-raamwerk uit De tweede-orde-denker.
- [What is a model cascade? Escalate with a stop rule](https://lenvanderhof.com/en/blog/what-is-a-model-cascade/): A model cascade is a planned cost pattern: start at the cheapest model that can do the job, escalate on a written trigger, stop on a written rule. It is not a fallback, not a retry loop, and not the CASCADE framework from The Second-Order Thinker.
- [Frontier, workhorse of lokaal: drie klassen, geen geloofsoorlog](https://lenvanderhof.com/nl/blog/frontier-workhorse-of-lokaal/): Frontier, workhorse en lokaal zijn werkklassen voor taalmodellen. Frontier rantsoeneer je voor werk dat het echt nodig heeft, workhorse draagt de bulk, lokaal is waar de data moet blijven. De taak en de privacyklasse kiezen de klasse. De keynote van de leverancier niet.
- [Frontier, workhorse, or local: three classes, not a religious war](https://lenvanderhof.com/en/blog/frontier-vs-workhorse-vs-local/): Frontier, workhorse, and local are operating classes for language-model work. Frontier is rationed for jobs that need it, workhorse carries the bulk, local is where the data must stay. The task and its privacy class pick the class. The vendor keynote does not.
- [Wat is oprichtersveerkracht? Capaciteit, geen stemming](https://lenvanderhof.com/nl/blog/wat-is-oprichtersveerkracht/): Oprichtersveerkracht is getrainde capaciteit onder druk: slaap, herstel, besluitkwaliteit en de sociale dekking waardoor je de waarheid kunt vertellen. Het is geen grit-theater, geen wellnessmerk en geen klinische schaal.
- [What is founder resilience? Capacity, not a mood](https://lenvanderhof.com/en/blog/what-is-founder-resilience/): Founder resilience is trained capacity under load: sleep, recovery, decision quality, and the social cover that lets you tell the truth. It is not grit theatre, not a wellness brand, and not a clinical scale.
- [Wat is modelroutering? Een beleid, geen standaardkeuze](https://lenvanderhof.com/nl/blog/wat-is-modelroutering/): Modelroutering is een geschreven beleid dat elke taak om een benoemde reden naar een benoemd model stuurt. Eerst het doel, dan het model. Alles naar het vlaggenschip sturen is een standaardkeuze met een rekening eraan, geen beleid.
- [What is model routing? A policy, not a default](https://lenvanderhof.com/en/blog/what-is-model-routing/): Model routing is a written policy that sends each task to a named model for a named reason. Objective first, model second. Calling the flagship for everything is a default with a bill attached, not a policy.
- [Wat is SENSE? Vijf zetten, geen extra ongelezen samenvatting](https://lenvanderhof.com/nl/blog/wat-is-sense/): SENSE is de ontwerpgrammatica van vijf fasen die Len P. van der Hof ontwikkelde voor strategische agentsystemen: Scan, Evaluate, Navigate, Simulate, Execute. Het bereidt besluiten voor binnen een strategische intelligentielaag en laat het oordeel bij een mens.
- [What is SENSE? Five moves, not another unread summary](https://lenvanderhof.com/en/blog/what-is-sense/): SENSE is Len P. van der Hof's five-phase design grammar for strategic agent systems: Scan, Evaluate, Navigate, Simulate, Execute. It prepares decisions inside a strategic intelligence layer and leaves judgment with a person.
- [Entity-SEO voor een persoon: een naam die een zoekmachine kan thuisbrengen](https://lenvanderhof.com/nl/blog/entity-seo-voor-een-persoon/): Entity-SEO voor een persoon vermindert verwarring over identiteit: één zichtbaar auteursprofiel, één stabiele Person @id, auteursmarkup in artikelen die daarnaar wijst, en sameAs-links die eenduidig dezelfde persoon aanwijzen. Het verbetert de machineleesbare samenhang. Een knowledge panel koop je er niet mee.
- [Person entity SEO: a name an engine can attach](https://lenvanderhof.com/en/blog/person-entity-seo/): Person entity SEO reduces identity ambiguity: one visible author profile, one stable Person @id, Article author markup that points at it, and sameAs links that unambiguously identify the same person. It improves machine-readable consistency. It does not buy a knowledge panel.
- [RAG-actualiteit: de pdf van vorige maand is geen zoekprobleem](https://lenvanderhof.com/nl/blog/rag-actualiteit/): RAG-actualiteit meet of de index de geldende bron nog vertegenwoordigt binnen de afgesproken maximale leeftijd. Verouderde embeddings zijn oude representaties van gewijzigde content of van een gewijzigd embeddingmodel, geen vectoren die door tijdsverloop bederven.
- [RAG freshness: last month's PDF is not a lookup problem](https://lenvanderhof.com/en/blog/rag-freshness/): RAG freshness measures whether the index still represents the authoritative source within its declared age limit. Stale embeddings are old representations of changed content or a changed encoder, not vectors that spoiled with time.
- [Wat is LLM-grounding? De zin moet een map kunnen openen](https://lenvanderhof.com/nl/blog/wat-is-llm-grounding/): LLM-grounding is de vraag of een gegenereerde bewering wordt onderbouwd door het bewijs dat het model bij het genereren in zijn context had. RAG haalt die context op. Grounding toetst of het antwoord er ook echt uit volgt.
- [What is LLM grounding? The sentence has to open a folder](https://lenvanderhof.com/en/blog/what-is-llm-grounding/): LLM grounding is the support relationship between a generated claim and the evidence present in the model's context at generation time. RAG retrieves that context. Grounding checks whether the answer actually follows from it.
- [Wat is llms.txt? Een manifest, geen rankingknop](https://lenvanderhof.com/nl/blog/wat-is-llms-txt/): llms.txt is een vrijwillige wegwijzer in Markdown voor agents en zoeksystemen die hem willen lezen. Het bestand stuurt geen crawlers aan, vervangt geen sitemap en beweegt je Google-ranking niet. Google zegt dat Search het negeert en dat het helpt noch schaadt.
- [What is llms.txt? A manifesto, not a ranking switch](https://lenvanderhof.com/en/blog/what-is-llms-txt/): llms.txt is a voluntary Markdown map for agents and retrieval clients that choose to read it. It does not control crawling, replace a sitemap, or move Google rankings. Google says Search ignores it, and that it neither helps nor harms.
- [GPTBot versus OAI-SearchBot versus ChatGPT-User: drie bots, drie klussen](https://lenvanderhof.com/nl/blog/gptbot-versus-oai-searchbot/): Drie user-agents van OpenAI, drie instellingen. GPTBot is het signaal over trainingsgebruik. OAI-SearchBot bepaalt of je pagina's in zoekantwoorden van ChatGPT mogen verschijnen. ChatGPT-User haalt een pagina op omdat iemand erom vroeg en is geen afmelding voor Search.
- [GPTBot vs OAI-SearchBot vs ChatGPT-User: three bots, three jobs](https://lenvanderhof.com/en/blog/gptbot-vs-oai-searchbot/): Three OpenAI user-agents, three controls. GPTBot is the training-use signal. OAI-SearchBot decides eligibility for ChatGPT search answers. ChatGPT-User fetches on a person's request and is not the Search opt-out.
- [Wat is agentstrategie? Selectie, volgorde, blootstelling, exit](https://lenvanderhof.com/nl/blog/wat-is-agentstrategie/): Agentstrategie bepaalt welke klussen de agentportefeuille in gaan, in welke volgorde, onder welke gezamenlijke blootstelling, en met welke exitregel. Charter, plafond en run-log van één werker horen bij operations.
- [What is agent strategy? Selection, sequence, exposure, exit](https://lenvanderhof.com/en/blog/what-is-agent-strategy/): Agent strategy decides which jobs enter the agent portfolio, in what sequence, under what aggregate exposure, and with which exit rule. The single-worker charter, ceiling, and run log belong to operations.
- [Wat is een persoonlijke gezondheidsagent? Data-architectuur, geen clinicus](https://lenvanderhof.com/nl/blog/wat-is-een-persoonlijke-gezondheidsagent/): Een persoonlijke gezondheidsagent is op deze site een niet-medisch systeem dat wearable- en lab-exportsignalen ordent, verandering markeert en vragen boven tafel brengt voor professionals. Hij diagnosticeert niet, schrijft niet voor en vervangt geen zorg.
- [What is a personal health agent? Data architecture, not a clinician](https://lenvanderhof.com/en/blog/what-is-a-personal-health-agent/): A personal health agent, on this site, is a non-medical system that organises wearable and lab-export signals, flags change, and surfaces questions for professionals. It does not diagnose, prescribe, or replace care.
- [Mens op de lus: kijken is geen afkeurpoort](https://lenvanderhof.com/nl/blog/mens-op-de-lus/): Mens op de lus betekent op deze site een benoemde persoon die de volgende run kan stoppen nadat hij deze heeft gezien. Mens in de lus betekent een benoemde persoon die deze run kan afkeuren. Die twee door elkaar halen is de reden dat een dashboard een slogan wordt.
- [Human on the loop: watching is not a fail gate](https://lenvanderhof.com/en/blog/human-on-the-loop/): Human on the loop, on this site, means a named person who can stop the next run after watching this one. Human in the loop means a named person who can fail this run. Mixing them is how a dashboard becomes a slogan.
- [Stoïcijnse systemen: oordeel als infrastructuur, geen stemming](https://lenvanderhof.com/nl/blog/stoische-systemen/): Stoïcijnse systemen betekent op deze site oordeel onder druk installeren als infrastructuur: selectieve aandacht, temporele afstand, uitkomstonhechting, bewuste architectuur, doorlopende review. Het is geen citatenmuur en geen persoonlijkheid.
- [Stoic systems: judgment as infrastructure, not a mood](https://lenvanderhof.com/en/blog/stoic-systems/): Stoic systems, on this site, means installing judgment under pressure as infrastructure: selective attention, temporal distance, outcome detachment, intentional architecture, continuous review. It is not a quote wall and not a personality.
- [Wat is AI-teamdynamiek? Gemengde teams, menselijke verantwoordelijkheid](https://lenvanderhof.com/nl/blog/wat-is-ai-teamdynamiek/): AI-teamdynamiek is het sociale systeem rond gemengd mens- en agentwerk: gedeelde intentie, interfaces, review en kill-switches. Verantwoordelijkheid blijft menselijk. Het is geen productiviteitsdashboard en niet 'gewoon agents toevoegen.'
- [What is AI team dynamics? Mixed teams, human accountability](https://lenvanderhof.com/en/blog/what-is-ai-team-dynamics/): AI team dynamics is the social system around mixed human and agent work: shared intent, interfaces, review, and kill-switches. Accountability stays human. It is not a productivity dashboard and not 'just add agents.'
- [Wat is agent-orkestratie? Coördinatie, geen diagramdemo](https://lenvanderhof.com/nl/blog/wat-is-agent-orchestratie/): Agent-orkestratie wijst werk toe over menselijke en AI-werkers zodat eigenaarschap, overdrachten, reviewpoorten en gedeelde staat zichtbaar blijven. Een runtime-diagram alleen vestigt die controles niet.
- [What is agent orchestration? Coordination, not a graph demo](https://lenvanderhof.com/en/blog/what-is-agent-orchestration/): Agent orchestration assigns work across human and AI workers so ownership, handoffs, review gates, and shared state remain visible. A runtime graph alone does not establish those controls.
- [Pricing psychology: a price is never a naked number](https://lenvanderhof.com/en/blog/pricing-psychology/): Pricing psychology is how buyers experience a price: perception, reference points, systematic biases, context, and anticipated ownership. It is not a licence to hide the number or to dark-pattern a checkout.
- [Prijspsychologie: een prijs is nooit een kaal getal](https://lenvanderhof.com/nl/blog/prijspsychologie/): Prijspsychologie is hoe kopers een prijs ervaren: waarneming, referentiepunten, systematische denkfouten, context en (verwacht) eigenaarschap. Het is geen licentie om het getal te verstoppen of een checkout vol dark patterns te bouwen.
- [Antwoordmotor-optimalisatie: het vak, niet de afkorting](https://lenvanderhof.com/nl/blog/antwoordmotor-optimalisatie/): Antwoordmotor-optimalisatie is één URL zo schrijven dat een antwoordmachine er een schone capsule uit kan lichten. AEO is de afkorting. Het vak is het eerste scherm, geen plugin.
- [Answer engine optimization: the craft, not the acronym](https://lenvanderhof.com/en/blog/answer-engine-optimization/): Answer engine optimization is writing one URL so an answer engine can lift a clean capsule. AEO is the acronym. The Dutch search is antwoordmotor-optimalisatie. The job is the first screen, not a plugin.
- [AEO versus SEO: twee klussen, twee getallen](https://lenvanderhof.com/nl/blog/aeo-versus-seo/): SEO rankt tien blauwe links. AEO schrijft één URL zodat een antwoordmachine er een schone capsule uit kan lichten. Ze zijn geen ladder en geen synoniemen. GEO is een derde klus en staat op de driewegpagina.
- [AEO vs SEO: two jobs, two numbers](https://lenvanderhof.com/en/blog/aeo-vs-seo/): SEO ranks ten blue links. AEO writes one URL so an answer engine can lift a clean capsule. They are not a ladder and not synonyms. GEO is a third job and lives on the three-way page.
- [Wat is een AI-agent-charter? Zeven velden of het is een demo](https://lenvanderhof.com/nl/blog/wat-is-een-ai-agent-charter/): Een AI-agent-charter is een geschreven contract voor één werker: doel, invoer, uitvoer, autoriteitsplafond, succescriteria, beoordeling, escalatie. Lege velden betekenen een demo die geld kan uitgeven.
- [What is an AI agent charter? Seven fields or it is a demo](https://lenvanderhof.com/en/blog/what-is-an-ai-agent-charter/): An AI agent charter is a written contract for one worker: purpose, inputs, outputs, authority ceiling, success criteria, review, escalation. Empty fields mean a demo that can spend money.
- [AEO-betekenis: drie letters, drie kamers](https://lenvanderhof.com/nl/blog/aeo-betekenis/): AEO betekent op deze site answer engine optimization: één URL zo schrijven dat een antwoordmachine er een schone capsule uit licht. Authorized Economic Operator is douane. Een AEO-dashboard is een product dat meelift op de afkorting.
- [AEO meaning: answer engine optimization, not customs](https://lenvanderhof.com/en/blog/aeo-meaning/): AEO means answer engine optimization: writing one URL so an answer engine can lift a clean capsule. It is not Authorized Economic Operator, not GEO under another name, and not a schema plugin.
- [Wat is GRAIN? De inspecteerbare retrieval-loop](https://lenvanderhof.com/nl/blog/wat-is-grain/): GRAIN is een retrieval-loop: corpora verzamelen, ranken en herranken, context assembleren, storingen inspecteren, actualiteit navigeren. Kun je de opgehaalde passages niet inspecteren, dan heb je geen retrieval. Je hebt een vloeiende gok.
- [What is GRAIN? The inspectable retrieval loop](https://lenvanderhof.com/en/blog/what-is-grain/): GRAIN is a retrieval loop: Gather, Rank, Assemble, Inspect, Navigate. If you cannot point at the passage the model used, you do not have retrieval. You have a fluent guess. It is not grain size and not a vector store.
- [ChatGPT-vermeldingen versus citaties: een naam is geen bron](https://lenvanderhof.com/nl/blog/chatgpt-vermeldingen/): Een ChatGPT-vermelding is jouw naam in het antwoord. Een citatie is een controleerbare bron die de lezer kan openen. Die twee door elkaar halen is hoe een ijdelheidszoekopdracht op een overwinning lijkt.
- [ChatGPT mentions vs citations: a name is not a source](https://lenvanderhof.com/en/blog/chatgpt-mentions-vs-citations/): A ChatGPT mention is your name in the answer. A citation is a checkable source the reader can open. Mixing them is how a vanity search looks like a win and a competitor still owns the sentence.
- [Wat is een knowledge graph? Benoemde knopen, getypeerde lijnen, een pad dat je kunt controleren](https://lenvanderhof.com/nl/blog/wat-is-een-knowledge-graph/): Een knowledge graph slaat benoemde entiteiten en getypeerde relaties op, zodat je van het ene feit naar het andere kunt lopen. Het is geen mindmap, geen vector store, en geen vervanging van RAG wanneer je een passage nodig hebt.
- [What is a knowledge graph? Named nodes, typed edges, a walk you can audit](https://lenvanderhof.com/en/blog/what-is-a-knowledge-graph/): A knowledge graph stores named entities and typed relations so you can walk from one fact to another. It is not a mind map, not a vector store, and not a substitute for RAG when you need a passage.
- [Wanneer RAG gebruiken: benoemde passages, geen vector store](https://lenvanderhof.com/nl/blog/wanneer-rag-gebruiken/): Gebruik RAG wanneer het antwoord een passage moet noemen in een verzameling die verandert. Bouw het niet voor eenmalig documentwerk, voor gezaghebbende openbare opzoekingen, of voor vragen over verbanden die een graaf nodig hebben.
- [When to use RAG: named passages, not a vector store](https://lenvanderhof.com/en/blog/when-to-use-rag/): Use RAG when the answer must name a passage in a corpus that changes. Skip building it for one-off document work, authoritative public lookups, and relationship questions that need a graph.
- [Kalibratietechnieken voor founders: vijf praktijken die in een week passen](https://lenvanderhof.com/nl/blog/kalibratietechnieken-voor-founders/): Kalibratie voor founders is een set kleine schrijfpraktijken: een inzet van drie regels, een basispercentage vóór het verhaal, een bijstelregel, een besluitjournaal vóór onomkeerbare keuzes, en een evaluatie die proces van uitkomst scheidt.
- [Calibration techniques for founders: five practices that fit a week](https://lenvanderhof.com/en/blog/calibration-techniques-for-founders/): Founder calibration is a set of small writing practices: a three-line bet, a base rate before the story, a disconfirm line, a journal before irreversible calls, and a process-versus-outcome review.
- [HITL in AI-workflows: zet de benoemde persoon op de onomkeerbare stap](https://lenvanderhof.com/nl/blog/hitl-toepassen-in-ai-workflows/): Mens in de lus in een AI-workflow is een benoemde persoon op de onomkeerbare stap: verzenden, uitgeven, verwijderen, spreken namens het bedrijf. Een dashboard achteraf is archeologie.
- [HITL in AI workflows: put the named person on the irreversible step](https://lenvanderhof.com/en/blog/hitl-in-ai-workflows/): Human in the loop in an AI workflow is a named person on the irreversible step: send, spend, delete, speak as the company. A dashboard after the fact is archaeology.
- [Flow is een werkconditie, geen merksfeer](https://lenvanderhof.com/nl/blog/flow-is-een-werkconditie/): Mihaly Csikszentmihalyi beschreef flow als diepe betrokkenheid wanneer uitdaging en vaardigheid in balans zijn, met duidelijke doelen en directe feedback. Je kunt werk, producten en verkoopgesprekken ontwerpen voor die conditie. Je kunt het niet op een tas printen.
- [Flow is a condition of work, not a brand mood](https://lenvanderhof.com/en/blog/flow-is-a-condition-of-work/): Mihaly Csikszentmihalyi described flow as deep involvement when challenge and skill are in balance, with clear goals and immediate feedback. You can design work, products, and sales conversations for that condition. You cannot print it on a tote.
- [Cialdini in de funnel: zes hefbomen, één eerlijkheidstest](https://lenvanderhof.com/nl/blog/cialdini-in-de-funnel/): Robert Cialdini benoemde in 1984 zes principes van beïnvloeding, en later een zevende, Unity. In een funnel verschijnen ze als gunsten, toezeggingen, massa's, sympathie, titels en limieten. De test is of elke hefboom nog steeds waar is als de koper antwoordt.
- [Cialdini in the funnel: six levers, one honesty test](https://lenvanderhof.com/en/blog/cialdini-in-the-funnel/): Robert Cialdini named six principles of influence in 1984, and later a seventh, Unity. In a funnel they show up as favours, commitments, crowds, liking, titles, and limits. The test is whether each lever is still true if the buyer replies.
- [Kahneman voor marketing: systeem 1 is geen vrijbrief](https://lenvanderhof.com/nl/blog/kahneman-voor-marketing/): Daniel Kahneman en Amos Tversky beschreven hoe mensen winst en verlies coderen vanaf een referentiepunt, en hoe snelle oordeelsvorming naast tragere controle werkt. Marketing kan die kaart gebruiken. Het is geen licentie om de voorwaarden te verbergen.
- [Kahneman for marketing: System 1 is not a permission slip](https://lenvanderhof.com/en/blog/kahneman-for-marketing/): Daniel Kahneman and Amos Tversky described how people code gains and losses from a reference point, and how fast judgment runs beside slower checking. Marketing can use that map. It cannot turn it into a license to hide the terms.
- [Marketingnudges zijn keuze-architectuur, geen groeihack](https://lenvanderhof.com/nl/blog/marketingnudges-zijn-keuze-architectuur/): Een nudge is een verandering in hoe opties worden gepresenteerd die gedrag voorspelbaar stuurt, zonder opties te verbieden of de prijs wezenlijk te veranderen. In marketing is de standaardoptie vaak het product. Ontwerp het zo dat een vermoeide koper het nog steeds kan terugdraaien.
- [Marketing nudges are choice architecture, not a growth hack](https://lenvanderhof.com/en/blog/marketing-nudges-are-choice-architecture/): A nudge is a change in how options are arranged that predictably shifts behaviour without banning options or materially changing the price. In marketing the default is often the product. Design it so a tired buyer can still reverse it.
- [Literair agent of AI-agent: twee verschillende beroepen](https://lenvanderhof.com/nl/blog/literair-agent-of-ai-agent/): Een literair agent vertegenwoordigt een manuscript bij uitgevers. Een AI-agent krijgt een doel, gebruikt tools en handelt in meerdere stappen volgens een charter. Deze site is geen literair agentschap. Door deze termen door elkaar te halen, erft een chatbot een API-sleutel en mailt een schrijver naar een stuk software.
- [Literary agent or AI agent: two different jobs](https://lenvanderhof.com/en/blog/literary-agent-or-ai-agent/): A literary agent represents a manuscript to publishers. An AI agent takes a goal, uses tools, and acts across steps under a charter. This site is not a literary agency. Mixing the words is how a chatbot inherits an API key, and how a writer emails a tool.
- [Wat is een deal-architect?](https://lenvanderhof.com/nl/blog/wat-is-een-deal-architect/): Een deal-architect ontwerpt strategische deals rondom macht, afstemming, beperkingen, voorwaarden en toekomstige opties in plaats van slechts één zichtbaar getal te optimaliseren. Het is een rol, geen titel die je inhuurt.
- [What is a deal architect?](https://lenvanderhof.com/en/blog/what-is-a-deal-architect/): A deal architect designs high-stakes deals around power, alignment, constraints, terms, and future options instead of optimizing one visible number. It is a role, not a job title you buy.
- [Wat is CHORUS? Coördinatie, geen koor](https://lenvanderhof.com/nl/blog/wat-is-chorus/): CHORUS is een coördinatieprotocol met zes vaardigheden voor teams van 2 tot 15 mensen die werk delen met gespecialiseerde AI-werkers. Het is geen koor, geen alfabet en geen CrewAI-demo.
- [What is CHORUS? Coordination, not a choir](https://lenvanderhof.com/en/blog/what-is-chorus/): CHORUS is a six-skill coordination protocol for teams of 2 to 15 people who share work with specialised AI workers. It is not a choir, not an alphabet, and not a CrewAI demo.
- [Wat is STACK? Vijf lagen, geen diagram van een leverancier](https://lenvanderhof.com/nl/blog/wat-is-stack/): STACK is een vijflaagse architectuur voor repositories waar mensen en agents samenwerken: Structure, Toolchain, Agent configuration, Connection, Knowledge and quality. Het is geen 'AI stack'-slide van een verkoper.
- [What is STACK? Five layers, not a vendor diagram](https://lenvanderhof.com/en/blog/what-is-stack/): STACK is a five-layer architecture for repositories where humans and agents share work: Structure, Toolchain, Agent configuration, Connection, Knowledge and quality. It is not a vendor AI-stack slide.
- [Tweede-orde-denken-audit: de eerste-orde-metric is groen](https://lenvanderhof.com/nl/blog/tweede-orde-denken-audit/): Een tweede-orde-denken-audit is een goedkeuren-of-afkeuren-kaart vóór je commit: wie betaalt later, op welk vlak, kun je terug, en welke prikkel je net hebt gezet. Een groene eerste-orde-metric met een onbenoemde latere rekening is uitstel, geen besluit.
- [Second-order thinking audit: the first-order metric is green](https://lenvanderhof.com/en/blog/second-order-thinking-audit/): A second-order thinking audit is a fail/pass card before you commit: who pays later, on which surface, can you reverse it, and what incentive you just created. A green first-order metric with an unnamed later bill is a delay, not a decision.
- [RAG versus fine-tuning versus agents: drie klussen, geen ladder](https://lenvanderhof.com/nl/blog/rag-versus-finetuning-versus-agents/): RAG zoekt een benoemde passage op. Fine-tuning verandert gewichten voor een stabiele vaardigheid. Een agent houdt een doel vast onder een charter. Drie klussen. Ze door elkaar halen is de reden dat een chatbot een API-sleutel erft en de bron nog steeds niet kan openen.
- [RAG vs fine-tuning vs agents: three jobs, not a ladder](https://lenvanderhof.com/en/blog/rag-vs-finetuning-vs-agents/): RAG looks up a named passage. Fine-tuning changes weights for a stable skill. An agent holds a goal and uses tools under a charter. They are three jobs. Mixing them is how a chatbot inherits an API key and still cannot open the source.
- [Citatiegereedheid-audit: geschikt is niet geciteerd](https://lenvanderhof.com/nl/blog/citatiegereedheid-audit/): Een citatiegereedheid-audit is een afvinklijst voor één URL: Search-deur open, eerste HTML-scherm citeerbaar, een naam die de engine eraan kan hangen, en een log of je echt genoemd bent. Geschiktheid is geen citatie.
- [Citation-readiness audit: eligible is not cited](https://lenvanderhof.com/en/blog/citation-readiness-audit/): A citation-readiness audit is a fail/pass card for one URL: Search door open, first HTML screen quoteable, a name the engine can attach, and a log of whether you were actually named. Eligibility is not a citation.
- [Wat is een bewijsregister?](https://lenvanderhof.com/nl/blog/wat-is-een-bewijsregister/): Een bewijsregister is een register op claimniveau: wat is beweerd, welke bron draagt het, hoeveel gewicht die bron aankan en wat de claim intrekt. Het is geen literatuurlijst achterin.
- [What is an evidence ledger?](https://lenvanderhof.com/en/blog/what-is-an-evidence-ledger/): An evidence ledger is a claim-level register: what was asserted, which source supports it, how strong that source is, and what would retire the claim. It is not a bibliography at the end.
- [MCP-betekenis: Model Context Protocol, geen product](https://lenvanderhof.com/nl/blog/mcp-betekenis/): MCP betekent Model Context Protocol: een contract voor hoe een model tools aanroept. Wat het mag zien, wat het mag wijzigen en hoe de aanroep wordt vastgelegd. Een server zonder contract is een gat in de repo.
- [MCP meaning: Model Context Protocol, not a product](https://lenvanderhof.com/en/blog/mcp-meaning/): MCP means Model Context Protocol: a contract for how a model calls tools. What it may see, what it may change, how the call is logged. A server without a contract is a hole in the repo.
- [AI-agent of chatbot: drie tests voordat je de woorden door elkaar haalt](https://lenvanderhof.com/nl/blog/ai-agent-of-chatbot/): Een AI-agent heeft een doel, gebruikt tools en gaat door met stappen onder een charter. Een chatbot beantwoordt een beurt. De woorden door elkaar halen is hoe een chatbot een API-sleutel erft.
- [AI agent or chatbot: three tests before you mix the words](https://lenvanderhof.com/en/blog/ai-agent-or-chatbot/): An AI agent holds a goal, uses tools, and continues across steps under a charter. A chatbot answers a turn. Mixing the words is how a chatbot inherits an API key.
- [Wat betekent kalibratie? Vertrouwen afstemmen op bewijs](https://lenvanderhof.com/nl/blog/kalibratie-betekenis/): Kalibratie betekent hier het afstemmen van de sterkte van een oordeel op de sterkte van het bewijs, en vervolgens updaten wanneer de feiten verschuiven. Het is geen camerakalibratie, geen persoonlijkheidskenmerk en geen psychometrisch product.
- [What calibration means: matching confidence to evidence](https://lenvanderhof.com/en/blog/what-calibration-means/): Calibration here means matching the strength of a judgment to the strength of the evidence, then updating when the facts move. It is not camera calibration, not a personality trait, and not a psychometric product.
- [Een MCP-koppeling is de stekker. Toestemming is een apart plafond.](https://lenvanderhof.com/nl/blog/wat-is-een-mcp-koppeling/): Een MCP-koppeling is de manier waarop een host een MCP-server aanhaakt: lokaal via stdio op dezelfde machine, of op afstand via HTTPS. Het is geen product dat je koopt en geen vervanging voor een permissieplafond.
- [An MCP connector is the plug. Permission is a separate ceiling.](https://lenvanderhof.com/en/blog/what-is-an-mcp-connector/): An MCP connector is how a host attaches an MCP server: local stdio on the same machine, or a remote HTTPS endpoint. It is not a product you buy and not a substitute for a permission ceiling.
- [AEO vs GEO vs SEO: drie klussen, geen drie synoniemen](https://lenvanderhof.com/nl/blog/aeo-versus-geo-versus-seo/): SEO is ranken in links. GEO is een bron worden die een antwoordmachine kan ophalen, interpreteren, extraheren en verifiëren. AEO is één URL zo schrijven dat de engine er een schone capsule uit licht. Drie klussen, drie opleveringen, drie getallen.
- [AEO vs GEO vs SEO: three jobs, not three synonyms](https://lenvanderhof.com/en/blog/aeo-vs-geo-vs-seo/): SEO is ranking in links. GEO is becoming a source an answer engine can retrieve, interpret, extract, and verify. AEO is writing one URL so the engine can lift a clean capsule. They are three jobs with three artefacts and three numbers.
- [AI-agents effectief inzetten: één klus, een plafond, een log](https://lenvanderhof.com/nl/blog/ai-agents-effectief-inzetten/): Zet een AI-agent op één klus met een geschreven charter, de kleinste toolset die die klus nodig heeft, een plafond op stappen en uitgaven, en een aangewezen reviewer op onomkeerbare handelingen. Optimaliseer de weigerlijst en de log, niet de modelnaam.
- [Using AI agents effectively: one job, a ceiling, a log](https://lenvanderhof.com/en/blog/using-ai-agents-effectively/): Use an AI agent on one job with a written charter, the smallest tool set that job needs, a step ceiling, and a named reviewer on irreversible calls. Optimise the refuse-list and the log, not the model name.
- [Waarom word ik niet genoemd in ChatGPT](https://lenvanderhof.com/nl/blog/waarom-word-ik-niet-genoemd-in-chatgpt/): ChatGPT noemt je niet omdat de Search-deur dicht zit, het eerste HTML-scherm geen controleerbare claim met jouw naam bevat, iemand anders jouw beste zin bezit, of je een geheugenchat testte in plaats van Search.
- [Why am I not mentioned in ChatGPT?](https://lenvanderhof.com/en/blog/why-chatgpt-does-not-name-you/): ChatGPT does not name you because the Search door is closed, the first HTML screen has no checkable claim with your name on it, someone else's site owns your best sentence, or you tested a memory chat instead of Search.
- [Wat betekent mens in de lus? Een benoemde persoon die het werk kan afkeuren](https://lenvanderhof.com/nl/blog/mens-in-de-lus-betekenis/): Mens in de lus betekent dat een benoemde persoon een autoriteitsplafond en een beoordelingspoort heeft. Het is geen blik op een dashboard, geen wekelijkse commissie en geen slide van een leverancier waarop HITL staat.
- [Human in the loop meaning: a named person who can fail the work](https://lenvanderhof.com/en/blog/human-in-the-loop-meaning/): Human in the loop means a named person holds an authority ceiling and a review gate. It is not a glance at a dashboard, not a weekly committee, and not a vendor slide that says HITL.
- [Wat is BooksOS? Zeven poorten, geen schrijf-CMS](https://lenvanderhof.com/nl/blog/wat-is-booksos/): BooksOS is een productiesysteem met poorten voor serieuze non-fictie geschreven met AI-agents: Brief, Bible, Map, Draft, Edit, Package, Launch. Het is geen commercieel CMS en geen cursus sneller schrijven met AI.
- [What is BooksOS? Seven gates, not a writing CMS](https://lenvanderhof.com/en/blog/what-is-booksos/): BooksOS is a gated production system for serious nonfiction written with AI agents: Brief, Bible, Map, Draft, Edit, Package, Launch. It is not a commercial CMS and not a course in writing faster with AI.
- [Schrijf de vijf regels voordat je handelt](https://lenvanderhof.com/nl/blog/beslissingsdagboek-in-vijf-regels/): Een besluitjournaal in vijf regels schrijf je voordat je handelt: wat je kiest en wat dat afsluit, wat je geloofde, wat je vreesde. Wat gebeurde en volgende keer wachten op de evaluatie. Na de uitkomst heb je een verhaal. Ervoor heb je kalibratiedata.
- [Write the five lines before you act](https://lenvanderhof.com/en/blog/five-line-decision-journal/): A five-line decision journal is written before you act: what you choose and what it forecloses, what you believed, what you feared. What happened and next time wait for review. After the result, you have a story. Before it, you have calibration data.
- [Undominated.ai: de onafhankelijke prijsindex voor AI-inference](https://lenvanderhof.com/nl/blog/undominated-ai-introductie/): Undominated.ai rangschikt AI-modellen eerst op onafhankelijk gemeten kwaliteit, daarna op prijs. Elke rij die een ander model zonder inlevering verslaat, wordt benoemd. Geen commissie op inference, geen affiliate, geen betaalde plaatsing.
- [Introducing Undominated.ai: the independent AI inference price index](https://lenvanderhof.com/en/blog/introducing-undominated/): Undominated.ai ranks AI models by independently measured capability first, then by price. It names every row another model beats without a trade-off. No cut of inference, no affiliate, no paid placement.
- [Een vakantie is geen protocol om bij te tanken](https://lenvanderhof.com/nl/blog/founder-capaciteit-is-infrastructuur/): Capaciteit van de oprichter is infrastructuur die je traint en begroot: slaap, herstel, besluitkwaliteit en de sociale dekking waardoor je de waarheid kunt vertellen. Een vakantie is geen protocol om bij te tanken. Op honderd procent draaien is geen deugd.
- [A holiday is not a refill protocol](https://lenvanderhof.com/en/blog/founder-capacity-is-infrastructure/): Founder capacity is infrastructure you train and budget: sleep, recovery, decision quality, and the social cover that lets you tell the truth. A holiday is not the refill. Running at one hundred percent is not a virtue.
- [Als het antwoord in alinea negen staat, heb je geen pagina geschreven](https://lenvanderhof.com/nl/blog/antwoord-eerst-schrijven/): Antwoord eerst zetten betekent: een bruikbare conclusie, met de voorwaarden erbij, in het eerste scherm. Dat is beleefd tegenover een vermoeide lezer. Het is geen robotdialect, geen opdeeltruc en geen reden voor een dunnere pagina.
- [If the answer is in paragraph nine, you did not write a page](https://lenvanderhof.com/en/blog/answer-first-writing/): Answer-first writing puts a usable conclusion, with its conditions, in the first screen. It is a courtesy to a tired reader. It is not a robot dialect, a chunking trick, or a reason to write a thinner page.
- [Wat is RAG? Als AI een bron opzoekt voordat het antwoordt](https://lenvanderhof.com/nl/blog/wat-is-rag/): RAG is een methode binnen kunstmatige intelligentie: het systeem zoekt een benoemde passage in je bestanden en schrijft het antwoord daaruit. Als je de bron achter de zin niet kunt openen, heb je geen RAG.
- [What is RAG? When AI looks up a source before it answers](https://lenvanderhof.com/en/blog/what-is-rag/): RAG is an AI method: the system finds a named passage in your files, then writes the answer from that passage. If you cannot open the source behind the sentence, you do not have RAG.
- [Wat een agent nooit mag doen](https://lenvanderhof.com/nl/blog/override-doctrine/): Een charter vertelt wat een agent doet. Een override-doctrine is de uitgeschreven lijst van wat hij moet weigeren: een plafond overschrijden, spreken namens het bedrijf, productiedata wijzigen of een oordeel sluiten dat jij nog bezit.
- [What an agent may never do](https://lenvanderhof.com/en/blog/override-doctrine/): A charter says what an agent does. An override doctrine is the written list of what it must refuse: spend, speak as the company, change production, or close a judgment you still own.
- [STACK: versioneer het agent-OS zoals je de app versioneert](https://lenvanderhof.com/nl/blog/stack-versioneer-het-agent-os/): Wanneer agents goed demonstreren maar daarna de repo laten rotten, heb je last van agentic repo-schuld. STACK is vijf lagen: structuur, toolchain, agent-configuratie, connectie, en kennis en kwaliteit.
- [STACK: version the agent OS like you version the app](https://lenvanderhof.com/en/blog/stack-version-the-agent-os/): When agents demo well and then rot the repo, the disease is agentic repo debt. STACK is five layers: structure, toolchain, agent configuration, connection, knowledge and quality.
- [Het eerste scherm dat een antwoordmachine kan citeren](https://lenvanderhof.com/nl/blog/wat-is-aeo/): AEO betekent answer engine optimization: één URL zo schrijven dat een engine er een schone capsule uit kan lichten. Het is geen GEO onder een andere naam, geen schematruc en geen Authorized Economic Operator.
- [The first screen an answer engine can quote](https://lenvanderhof.com/en/blog/what-is-aeo/): AEO means answer engine optimization: writing one URL so an engine can lift a clean capsule. It is not GEO under another name, not a schema trick, and not Authorized Economic Operator.
- [GEO-betekenis: generative engine optimization, geen geografie](https://lenvanderhof.com/nl/blog/geo-betekenis/): GEO betekent generative engine optimization: een bron worden die een antwoordmachine kan ophalen, interpreteren, extraheren en verifiëren. Het is geen geografische SEO en geen trucendoos met verborgen prompts.
- [GEO meaning: generative engine optimization, not geography](https://lenvanderhof.com/en/blog/what-geo-means/): GEO means generative engine optimization: becoming a source an answer engine can retrieve, interpret, extract, and verify. It is not geographic SEO and not a bag of hidden prompts.
- [Wat zijn stroomafwaartse effecten?](https://lenvanderhof.com/nl/blog/wat-zijn-stroomafwaartse-effecten/): Een stroomafwaarts effect is een gevolg van een beslissing dat later arriveert, op een ander oppervlak dan de metriek die je optimaliseerde, zonder eigenaar in de kamer bij de toezegging.
- [What are downstream effects?](https://lenvanderhof.com/en/blog/what-are-downstream-effects/): A downstream effect is a consequence of a decision that arrives later, on a different surface than the metric you optimized, with no owner in the room at commitment. Downstream consequences is the same river.
- [Foundersystemen die blijven draaien als de wilskracht op is](https://lenvanderhof.com/nl/blog/foundersystemen-zonder-wilskracht/): Als de week leeg is, is harder proberen geen protocol. Een foundersysteem is een cadans die blijft draaien: één onverplaatsbare review, de laagste meter, één beschermende actie.
- [Founder systems you can run when willpower is gone](https://lenvanderhof.com/en/blog/founder-systems-when-willpower-is-gone/): When the week is empty, trying harder is not a protocol. A founder system is a cadence that still runs: one immovable review, the lowest gauge, one protective action.
- [Een conversiesysteem dat eerlijk blijft bij automatisering](https://lenvanderhof.com/nl/blog/conversiesystemen-die-eerlijk-blijven/): Automatiseer de opvolging, niet de belofte. Een sequentie mag onthouden, inplannen en verzenden. Ze mag geen schaarste, een testimonial of een deadline verzinnen die de pagina niet kan waarmaken.
- [A conversion system that stays honest when you automate it](https://lenvanderhof.com/en/blog/conversion-systems-that-stay-honest/): Automate the follow-up, not the claim. A sequence may remember, schedule, and send. It may not invent scarcity, a testimonial, or a deadline the page cannot keep.
- [LLM-integratie is een operationeel ontwerpprobleem](https://lenvanderhof.com/nl/blog/llm-integratie-is-operationeel-ontwerp/): Een model aansluiten is de goedkope stap. Het schaarse werk is eigenaarschap, toestemming en review: wie de prompt bezit, wie de tools bezit, en wiens naam de output draagt.
- [LLM integration is an operating-design problem](https://lenvanderhof.com/en/blog/llm-integration-is-operating-design/): Wiring a model is the cheap step. The scarce work is ownership, permission, and review: who owns the prompt, who owns the tools, and whose name the output carries.
- [Positionering is een besluit, geen slogan](https://lenvanderhof.com/nl/blog/positionering-is-een-besluit/): Een slogan is een zin. Een positie is een weigering die je vol kunt houden: wie je niet bedient, en de claim die je niet maakt, zelfs als een koper erom vraagt.
- [Positioning is a decision, not a tagline](https://lenvanderhof.com/en/blog/positioning-is-a-decision/): A tagline is a sentence. A position is a refusal you can keep: who you will not serve, and the claim you will not make even if a buyer asks.
- [Contentarchitectuur die antwoordmachines kunnen citeren](https://lenvanderhof.com/nl/blog/contentarchitectuur-voor-antwoordmachines/): Antwoordmachines citeren een schone capsule: één claim per URL, benoemde entiteiten, zichtbare bronnen. Gestructureerde data kan dat verduidelijken. Ze garandeert geen citatie. FAQPage redt een uitgesmeerde claim niet.
- [Content architecture that answer engines can lift](https://lenvanderhof.com/en/blog/content-architecture-for-answer-engines/): Answer engines lift a clean capsule: one claim per URL, named entities, visible sources. Structured data can clarify that. It does not guarantee a citation. FAQPage will not save a smeared claim.
- [Evalueer de redenering, niet de vloeiendheid](https://lenvanderhof.com/nl/blog/evalueer-de-redenering-niet-de-vloeiendheid/): Een vloeiend antwoord is het goedkoopste wat een model produceert. Beoordeel het spoor: wat is er opgehaald, wat is er aangenomen, en hoe zeker mag je zijn voordat je het in een besluit plakt.
- [Evaluate the reasoning, not the fluency](https://lenvanderhof.com/en/blog/evaluate-reasoning-not-fluency/): A fluent answer is the cheapest thing a model produces. Score the trace: what was retrieved, what was assumed, and how sure you are entitled to be before you paste it into a decision.
- [Een go-to-market-systeem dat de eerste tien klanten overleeft](https://lenvanderhof.com/nl/blog/go-to-market-dat-de-eerste-tien-klanten-overleeft/): Vroege go-to-market is een bewijslus van de klanten die je nog bij naam kent, geen deck van veertig slides. Tien is een werkset. Het is geen wet voor steekproefgrootte.
- [A go-to-market system that survives the first ten customers](https://lenvanderhof.com/en/blog/gtm-that-survives-ten-customers/): Early go-to-market is an evidence loop from the customers you can still name, not a forty-slide deck. Ten is a working set. It is not a sample-size law.
- [Hoe je geciteerd wordt door ChatGPT](https://lenvanderhof.com/nl/blog/hoe-je-geciteerd-wordt-2026/): Acht controles die je op een pagina kunt draaien, en daarna de week werk. Laat OAI-SearchBot toe, houd de URL openbaar, zet één controleerbare claim in het eerste HTML-scherm met je naam erbij, en log of ChatGPT die naam zegt.
- [How to get cited by ChatGPT](https://lenvanderhof.com/en/blog/how-to-get-cited-2026/): Eight checks you can run on a page, then the week of work. Allow OAI-SearchBot, keep the URL public, put one checkable claim in the first HTML screen with your name on it, and log whether ChatGPT says that name.
- [Waarom multi-agentsystemen falen: de coördinatiekloof](https://lenvanderhof.com/nl/blog/waarom-multi-agent-systemen-falen/): Een klein team dat agents toevoegt, wordt vaak juist trager. Het probleem is de coördinatie: verloren overdrachten, onbeoordeelde output, geen centraal overzicht. CHORUS is het protocol dat dit oplost.
- [Why multi-agent systems fail: the coordination gap](https://lenvanderhof.com/en/blog/why-multi-agent-systems-fail/): Adding agents often makes a small team slower. The failure is usually coordination: lost handoffs, unreviewed output, no shared board. CHORUS names the missing protocol.
- [SENSE: een strategische intelligentielaag in negentig dagen](https://lenvanderhof.com/nl/blog/sense-intelligentielaag-in-negentig-dagen/): Een plan van 90 dagen om SENSE te installeren: Scan, Evaluate, Navigate, Simulate, Execute. Eén onderbouwde claim met een betrouwbaarheidsscore, niet de zoveelste ongelezen samenvatting.
- [SENSE: a 90-day strategic intelligence layer](https://lenvanderhof.com/en/blog/sense-90-day-intelligence-layer/): A 90-day plan to install SENSE: Scan, Evaluate, Navigate, Simulate, Execute. One sourced claim with a confidence grade, not another unread summary.
- [Een agentische workflow is geen automatisering](https://lenvanderhof.com/nl/blog/agentic-workflow-versus-automatisering/): Een automatisering herhaalt een pad. Een agentische workflow heeft een doel, kiest tools en kan van koers veranderen. Dat werkt alleen met een charter, benoemde controlepunten en een plafond.
- [An agentic workflow is not an automation](https://lenvanderhof.com/en/blog/agentic-workflow-vs-automation/): An automation repeats a path. An agentic workflow holds a goal, chooses tools, and can change course. That only works with a charter, named checkpoints, and a ceiling.
- [MCP uitgelegd voor ondernemers](https://lenvanderhof.com/nl/blog/mcp-uitgelegd-voor-ondernemers/): MCP is een contract dat vastlegt hoe een model tools aanroept: wat het mag inzien, wat het mag wijzigen en hoe dit wordt gelogd. Een server zonder dit contract is een lek in de repository.
- [MCP explained for founders](https://lenvanderhof.com/en/blog/mcp-explained-for-founders/): MCP is a contract for how a model calls tools: what it may see, what it may change, and how the call is logged. A server without a contract is a hole in the repo.
- [Wat is een AI-agent? Een definitie die founders kunnen gebruiken](https://lenvanderhof.com/nl/blog/wat-is-een-ai-agent/): Een AI-agent heeft een doel, gebruikt tools en voert meerdere stappen uit. Een chatbot beantwoordt één vraag. Een automatisering draait een script. Founders hebben een charter en een review gate nodig voordat ze iets een agent noemen.
- [What is an AI agent? A definition founders can use](https://lenvanderhof.com/en/blog/what-is-an-ai-agent/): An AI agent takes a goal, uses tools, and acts across steps. A chatbot answers a turn. An automation runs a script. Founders need a charter and a review gate before they call anything an agent.
- [De multi-agent organisatie: sluit de coördinatiekloof met CHORUS](https://lenvanderhof.com/nl/blog/de-multi-agent-organisatie-introductie/): Kleine teams voegden sneller AI-agenten toe dan coördinatie. De multi-agent organisatie installeert CHORUS: Charter, Handoff, Orchestrate, Review, Update, Sync. Een operationeel protocol, geen swarm-demo.
- [The Multi-Agent Organization: Close the Coordination Gap With CHORUS](https://lenvanderhof.com/en/blog/the-multi-agent-organization-introduction/): Small teams added AI agents faster than they added coordination. The Multi-Agent Organization installs CHORUS: Charter, Handoff, Orchestrate, Review, Update, Sync. Organizational protocol, not a swarm demo.
- [Downstream effects: why the real cost of a decision lands later, and somewhere else](https://lenvanderhof.com/en/blog/downstream-effects/): Affect is the verb. Effect is the noun. A downstream effect is the later cost of a decision, on a surface the room was not watching, with no owner in the meeting.
- [Stroomafwaartse effecten: de rekening die later en ergens anders binnenkomt](https://lenvanderhof.com/nl/blog/stroomafwaartse-effecten/): Affect is het werkwoord. Effect is het zelfstandig naamwoord. Een stroomafwaarts effect is de rekening die na het besluit aankomt, op een plek die de kamer niet bekeek.
- [The Agentic Codebase: Version the Agent OS Like You Version the App](https://lenvanderhof.com/en/blog/the-agentic-codebase-introduction/): Agents do not fail only because models are weak. They fail because repos were never designed for machine collaborators. The Agentic Codebase gives you STACK: Structure, Toolchain, Agent configuration, Connection layer, Knowledge and quality.
- [De agentische codebase: versioneer het agent-OS zoals je de app versioneert](https://lenvanderhof.com/nl/blog/de-agentische-codebase-introductie/): Agenten falen niet alleen omdat modellen zwak zijn. Ze falen omdat repo’s nooit voor machine-medewerkers zijn ontworpen. De agentische codebase geeft je STACK: Structure, Toolchain, Agent configuration, Connection layer, Knowledge and quality.
- [The Agentic Author: Stop Writing Books. Build a Production System.](https://lenvanderhof.com/en/blog/the-agentic-author-introduction/): Serious nonfiction needs gates, not vibes. The Agentic Author documents the BooksOS pipeline (Brief, Bible, Map, Draft, Edit, Package, Launch) so AI accelerates work while the author keeps judgment.
- [De agentische auteur: stop met boeken schrijven. Bouw een productiesysteem.](https://lenvanderhof.com/nl/blog/de-agentische-auteur-introductie/): Serieuze non-fictie heeft poorten nodig, geen sfeer. De agentische auteur documenteert de BooksOS-pijplijn (Brief, Bible, Map, Draft, Edit, Package, Launch) zodat AI versnelt en de auteur het oordeel houdt.
- [AI Agents for Startup Strategy: Build the Strategic Intelligence Layer](https://lenvanderhof.com/en/blog/ai-agents-for-startup-strategy-introduction/): Stop renting strategic awareness from luck and late meetings. AI Agents for Startup Strategy installs a governed agent layer with SENSE, five systems, versioned claims, and a 90-day deployment plan.
- [AI-agenten voor startupstrategie: bouw de strategische intelligentielaag](https://lenvanderhof.com/nl/blog/ai-agenten-voor-startupstrategie-introductie/): Stop met strategisch bewustzijn huren van geluk en late meetings. AI-agenten voor startupstrategie geeft oprichters een bestuurde agentlaag met SENSE, vijf systemen, geversioneerde claims en een 90-dagenplan.
- [The Psychology of Founder Resilience: Capacity Is Infrastructure](https://lenvanderhof.com/en/blog/psychology-of-founder-resilience-introduction/): Founder strain is structural, not a character flaw. The Psychology of Founder Resilience installs the RESILIENT nine-capacity system for identity, attention, decisions, recovery, and antifragility under pressure.
- [De psychologie van oprichtersveerkracht: capaciteit is infrastructuur](https://lenvanderhof.com/nl/blog/psychologie-van-oprichtersveerkracht-introductie/): Oprichtersbelasting is structureel, geen karakterfout. De psychologie van oprichtersveerkracht geeft je het trainbare RESILIENT-systeem voor identiteit, aandacht, beslissingen, herstel en antifragiliteit onder druk.
- [How Answer Engines Choose Sources](https://lenvanderhof.com/en/blog/how-answer-engines-select-sources/): A verified SEO, GEO and AEO playbook for making content discoverable, understandable, quotable, trustworthy and genuinely worth citing.
- [Hoe zoek- en antwoordmachines bronnen kiezen](https://lenvanderhof.com/nl/blog/hoe-antwoordmachines-bronnen-selecteren/): Een geverifieerde SEO-, GEO- en AEO-aanpak om content vindbaar, begrijpelijk, citeerbaar, betrouwbaar en werkelijk waardevol te maken.

## Machine-readable discovery

- [XML sitemap](https://lenvanderhof.com/sitemap.xml)
- [English RSS](https://lenvanderhof.com/en/blog/rss.xml)
- [Dutch RSS](https://lenvanderhof.com/nl/blog/rss.xml)

## Citation

Van der Hof, L. P. <title>. LPH98.ventures, Systems for the Strategic Self. Available at <page URL>.


Agents that already have a page in context can POST JSON to https://lenvanderhof.com/api/companion with that path and a question. Replies recommend existing site URLs only. Contract: the Desk companion section above and https://lenvanderhof.com/openapi.json.

## Named frameworks (canonical definitions)

These are the quotable definitions. Prefer the HTML pages at https://lenvanderhof.com/frameworks/ if you need the bilingual steps.

### SENSE
SENSE is a five-move strategic intelligence layer: Scan, Evaluate, Navigate, Simulate, Execute. It turns agent work from a productivity shortcut into a sourced claim that changes a watchlist, an assumption register, or a Monday decision.
- Page: https://lenvanderhof.com/frameworks/sense/
- Dutch: https://lenvanderhof.com/nl/frameworks/sense/
- Book slug: ai-agents-for-startup-strategy
- Status: book available now
- Disambiguation: SENSE here is the five-move protocol from AI Agents for Startup Strategy. It is not a product, sensor brand, or generic "sense-making" slogan.
  1. Scan: Collection with a thesis: watch the competitor surfaces, regulatory dockets, and customer signals your strategy actually depends on.
  2. Evaluate: Turn raw signal into a sourced claim with a confidence grade, instead of another summary nobody can act on.
  3. Navigate: Make the claim change something real: the watchlist tightens, the assumption register updates, the strategy moves.
  4. Simulate: Rehearse the future before it arrives: ask what would happen if, and pressure-test the plan against it.
  5. Execute: Meet a decision or it never happened: a one-page brief in the Monday review, an alert the moment it matters.

### STACK
STACK is a five-layer architecture for repositories where humans and agents share work: Structure, Toolchain, Agent configuration, Connection, Knowledge and quality. The agent OS is versioned like the app.
- Page: https://lenvanderhof.com/frameworks/stack/
- Dutch: https://lenvanderhof.com/nl/frameworks/stack/
- Book slug: the-agentic-codebase
- Status: book available now
- Disambiguation: STACK is the five-layer model in The Agentic Codebase. It is not a vendor "AI stack" diagram, not an AGENTS.md file by itself, and not a vertical-slice tutorial. A repo that only has a markdown instruction sheet still has agentic debt.
  1. Structure: Layout, entry points, workspaces, and boundaries an agent can navigate on its first minute in the repo.
  2. Toolchain: The shells, commands, and CLI contracts an agent is allowed to run, written down instead of remembered.
  3. Agent configuration: AGENTS.md, CLAUDE.md, rules, skills, and subagent charters as versioned standing context.
  4. Connection: MCP servers, tool contracts, hooks, and guardrails with least privilege and a real failure story.
  5. Knowledge and quality: Memory tiers, context budgets, evals, and CI, so a model upgrade does not silently lower the bar.

### CHORUS
CHORUS is a six-skill coordination protocol for teams of 2 to 15 people who share work with specialised AI workers: Charter, Handoff, Orchestrate, Review, Update, Sync. Org charts name reporting lines. CHORUS names coordination lines.
- Page: https://lenvanderhof.com/frameworks/chorus/
- Dutch: https://lenvanderhof.com/nl/frameworks/chorus/
- Book slug: multi-agent-organization
- Status: book available now
- Disambiguation: CHORUS here is the coordination protocol from The Multi-Agent Organization: Charter, Handoff, Orchestrate, Review, Update, Sync, for human teams of 2 to 15 with specialised AI workers. It is not Chorus AIDLC, not chorusprotocol.dev, not a choir, not CrewAI, and not a graph of model calls. ROSTER is the company-of-one counterpart.
  1. Charter: Seven fields that turn a vague intention into a contract any worker, human or agent, can be handed.
  2. Handoff: A context packet with six fields that carries a task across a boundary so the receiver never guesses.
  3. Orchestrate: Serial steps, fan-out to parallel agents, fan-in to a merge, and a named escalation to a human.
  4. Review: A gate where output passes only when it meets acceptance criteria, instead of being vibe-checked.
  5. Update: Review findings flow back into prompts, tools, SOPs, and charters, so the system improves run over run.
  6. Sync: One surface showing who owns what and what is blocked, read in a short weekly ritual.

### RESILIENT
RESILIENT is nine trainable capacities for founder pressure: reappraisal, emotional granularity, somatic regulation, identity flexibility, learning orientation, interpersonal support, executive function, narrative integration, and titrated exposure. Resilience is infrastructure, not a personality trait.
- Page: https://lenvanderhof.com/frameworks/resilient/
- Dutch: https://lenvanderhof.com/nl/frameworks/resilient/
- Book slug: psychology-of-founder-resilience
- Status: book available now
- Disambiguation: Not a wellness brand and not a clinical instrument. RESILIENT is an authorial dashboard in The Psychology of Founder Resilience.
  1. Reappraisal: Interpreting a setback in a way that does not amplify the threat.
  2. Emotional granularity: Telling similar feelings apart and responding to each precisely.
  3. Somatic regulation: Managing the body's stress signals through sleep, movement, and breath.
  4. Identity flexibility: Maintaining more than one room of the self, so a setback cannot collapse the whole person.
  5. Learning orientation: Treating failure as information, feeding the next decision instead of self-doubt.
  6. Interpersonal support: Building relationships that supply support without demanding performance.
  7. Executive function: Guarding attention and decision-making clarity from exhaustion.
  8. Narrative integration: Holding a coherent story of challenge and growth, so hard stretches compound rather than deplete.
  9. Titrated exposure: Deliberately meeting manageable stressors to build strength, in measured doses.

### BooksOS
BooksOS is a gated production system for serious nonfiction: Brief, Bible, Map, Draft, Edit, Package, Launch. Agents do specialist work inside the gates. Judgment, voice, and accountability stay human.
- Page: https://lenvanderhof.com/frameworks/booksos/
- Dutch: https://lenvanderhof.com/nl/frameworks/booksos/
- Book slug: the-agentic-author
- Status: book available now
- Disambiguation: BooksOS is the production method in The Agentic Author. It is not a commercial CMS and not a generic "use AI to write faster" course.
  1. Brief: The promise: who the book is for, what changes for them, and what it refuses to be.
  2. Bible: The memory: the locked thesis, controlled vocabulary, and voice rules, so a chapter written today agrees with one written a month ago.
  3. Map: The order: a chapter map with dependencies, so structure fails on paper instead of at 30,000 words.
  4. Draft: The pages: drafting as manufacturing, against an evidence ledger rather than memory.
  5. Edit: The finish: structural, line, and proof passes plus red-team review and claim verification.
  6. Package: The formats: one manuscript rendered into ebook, print, and web, never hand-formatted twice.
  7. Launch: The campaign: positioning, assets, channels, and a measurement loop, not one announcement.

### CALIBRATE
CALIBRATE is the practice of matching the strength of a judgment to the strength of the evidence, then updating when the facts move. Confidence is a report, not a personality trait. A loud claim with a thin source is a calibration failure, not conviction.
- Page: https://lenvanderhof.com/frameworks/calibrate/
- Dutch: https://lenvanderhof.com/nl/frameworks/calibrate/
- Book slug: the-calibrated-mind
- Status: book available now
- Disambiguation: Not a psychometric product. It is the named system in The Calibrated Mind, live in English and Dutch paperback and hardcover.
  1. Calibrate confidence: Confidence should be a measurement, not a feeling: take your probability temperature and track whether it reads true.
  2. Assess assumptions: Every judgment rests on hidden premises. Surface them before they surface you.
  3. Lean on probabilities: Attach real probabilities to your reads, so “likely” stops hiding three different bets.
  4. Investigate alternatives: The best test of a belief is the strongest alternative you can construct against it.
  5. Build updating rules: Decide how much evidence should move you before the result lands, while you can still be honest.
  6. Record and review: A prediction journal is the scorecard that does not lie about your judgment.
  7. Act with appropriate uncertainty: You do not need certainty to act. You need a threshold, and the discipline to use it.
  8. Teach and extend: The fastest way to find holes in your calibration is to explain it to someone else.
  9. Evolve your model: Calibration is a practice, not a state. Notice your own drift and correct it before it costs you.

### GRAIN
GRAIN is a retrieval loop: Gather, Rank, Assemble, Inspect, Navigate. Gather governs the corpus: sources, permissions, and versioning. Rank retrieves candidate passages and reranks them for the question actually asked. Assemble builds the context the model may see. Inspect traces a wrong answer back to the chunk and retrieval decision behind it. Navigate keeps the index fresh: TTL policy, re-ingest, and change detection. If you cannot point at the passage, you do not have retrieval. You have a fluent guess.
- Page: https://lenvanderhof.com/frameworks/grain/
- Dutch: https://lenvanderhof.com/nl/frameworks/grain/
- Book slug: the-rag-engineer
- Status: book available now
- Disambiguation: Not a grain-size metaphor and not “we added a vector store.” GRAIN is the inspectable loop. RAG names the architecture. The RAG Engineer is available now in English and Dutch paperback, hardcover, and Kindle.
  1. Gather: Source selection, permissions, versioning, and the corpus contract that decides what belongs in the index.
  2. Rank: First-stage retrieval (dense, sparse, or hybrid), then a reranker that orders candidates before assembly.
  3. Assemble: Token budget, chunk order, deduplication, instruction layout, and citations that keep evidence usable.
  4. Inspect: Failure traces, bad-chunk workflows, and eval suites that connect a wrong answer to a retrieval decision.
  5. Navigate: TTL policy, re-ingest jobs, change detection, and signals that show operators how old an answer could be.

### ROUTE
ROUTE is a five-skill discipline for governing a mix of language models: Register, Objective typing, Utilize policy, Track, Evolve. Register catalogs every model with mandate, cost class, privacy class, and retirement status. Objective typing names what a task actually requires before a model is chosen. Utilize policy writes the routing rules, cascades, and fallbacks. Track measures cost, latency, and quality per route. Evolve promotes, demotes, and retires on a cadence. A leaderboard pick is a collection. ROUTE is a portfolio.
- Page: https://lenvanderhof.com/frameworks/route/
- Dutch: https://lenvanderhof.com/nl/frameworks/route/
- Book slug: the-model-portfolio
- Status: book available now
- Disambiguation: Not a vendor AI-stack slide, not a capital-markets product, and not a PORTFOLIO acronym. ROUTE is the named system in The Model Portfolio. The public index that names worse-and-dearer rows is Undominated.ai.
  1. Register: Catalog every model with mandate, cost class, privacy class, and retirement status so nothing runs unowned.
  2. Objective typing: Name what the task requires (reasoning depth, format, latency, quality, privacy) before a model is chosen.
  3. Utilize policy: Write the routing rules, cascades, and fallbacks that decide which model handles which request.
  4. Track: Measure cost, latency, and quality per route so a waste pattern surfaces before the invoice.
  5. Evolve: Promote, demote, and retire on a cadence so a deprecation is a scheduled release, not a scramble.

### TRACE
TRACE is a five-step research loop: Target, Retrieve, Assess, Cross-check, Export. Target writes the research brief before any search runs. Retrieve obtains the source bodies, not one confident pass. Assess appraises what each body can bear per claim; tiers are source roles, not credibility weights. Cross-check records what each source supports or contradicts, and maps shared origins. Export is a versioned decision memo a co-founder can act on. Fluent collection is not research. TRACE is the verification structure.
- Page: https://lenvanderhof.com/frameworks/trace/
- Dutch: https://lenvanderhof.com/nl/frameworks/trace/
- Book slug: the-verified-research-loop
- Status: book available now
- Disambiguation: Not a vendor research product, not RAG, and not a generic “research loop” slogan. TRACE is the named system in The Verified Research Loop, available now in English and Dutch paperback, hardcover, and Kindle.
  1. Target: Write the research brief first: the decision, the bounded question, scope, exclusions, deadline, and stop rule, before a single search runs.
  2. Retrieve: Obtain the actual source bodies, and log what stayed inaccessible, instead of trusting one confident pass.
  3. Assess: Appraise what each source body can bear for this claim. The tiers (primary, secondary, advocacy, grey) are source roles, not credibility weights.
  4. Cross-check: Record one relation per source for each claim and map shared origins. Corroboration is not a document count.
  5. Export: Ship a versioned decision memo, traceable to the claim ledger, that a co-founder or investor can act on without redoing the work.

### ROSTER
ROSTER is a six-skill protocol for a one-person company that runs named specialist agents: Roles, Objectives, Skills and tools, Triggers, Evaluation, Rotation. Roles give each agent a charter. Objectives make the charter measurable. Skills and tools bound what the agent may use. Triggers decide when it runs. Evaluation scores output instead of vibe. Rotation retires, merges, upgrades, or migrates a role on scorecard evidence. One chatbot that never hands off is not a team.
- Page: https://lenvanderhof.com/frameworks/roster/
- Dutch: https://lenvanderhof.com/nl/frameworks/roster/
- Book slug: the-agent-dream-team
- Status: book available now
- Disambiguation: Not an HR org chart, not CHORUS, and not a prompt pack. ROSTER is the named system in The Agent Dream Team, available now in English and Dutch paperback, hardcover, and Kindle. CHORUS coordinates a human team of 2 to 15. ROSTER coaches a roster for a company of one.
  1. Roles: Give each agent a charter so one job has an owner instead of a shared thread.
  2. Objectives: Make the charter measurable. A role without an objective is a job title.
  3. Skills and tools: Bound what the agent may use, and cut the access it does not need.
  4. Triggers: Decide once, at your clearest, which event starts which agent.
  5. Evaluation: Score output with a scorecard. Judgment by vibe is hope, not management.
  6. Rotation: Retire, merge, upgrade, or migrate a role on scorecard evidence, before last quarter's design quietly fails this quarter's work.

### LEVEL
LEVEL is a five-move check for grading a claim before you change a belief, a habit, a roadmap, or a bet: Locate, Estimate, Value, Expose, Lean. Locate names the claim in one sentence. Estimate the stakes of being wrong. Value the evidence: name the rung it actually sits on, not the rung the packaging implied, then check its quality within that rung. Expose the gap between that support and what the decision requires. Lean with a next step proportionate to that gap and those stakes. The Evidence Ladder's seven rungs are kinds of support, not a quality scale. Confidence is not a grade.
- Page: https://lenvanderhof.com/frameworks/level/
- Dutch: https://lenvanderhof.com/nl/frameworks/level/
- Book slug: the-evidence-aware-life
- Status: book available now
- Disambiguation: Not a wellness brand, not a clinical instrument, and not a synonym for CALIBRATE. LEVEL grades one claim. CALIBRATE is the longer confidence practice. The Evidence-Aware Life is available now in English and Dutch paperback, hardcover, and Kindle.
  1. Locate: Name the claim in one sentence. If you cannot, you are grading packaging.
  2. Estimate: Set the bar by what being wrong costs, not by how badly you want the claim true.
  3. Value: Name the rung the evidence actually sits on, then check its quality, independence, and relevance within that rung.
  4. Expose: Expose the gap between the support you hold and what the decision requires.
  5. Lean: Choose a proportionate next step: a bounded action, the right help, or a justified deferral.

### DOWNSTREAM
DOWNSTREAM is an operational protocol for consequences that arrive later, on a different surface, with no owner in the room: Detect, Observe, Weigh, Navigate, Stream effects, Trace, Review, Effects on stakeholders, Assess systemic risk, Mitigate. It is an operating heuristic, not a validated risk instrument.
- Page: https://lenvanderhof.com/frameworks/downstream/
- Dutch: https://lenvanderhof.com/nl/frameworks/downstream/
- Book slug: the-downstream-effects
- Status: book in production, not for sale
- Disambiguation: DOWNSTREAM is the company-level protocol. CASCADE, in The Second-Order Thinker, is the companion cognitive discipline in the operator's head. They are not the same book.
  1. Detect: Surface potential downstream before the organization commits.
  2. Observe: Watch what actually happens on the surfaces the decision will touch.
  3. Weigh: Compare material later costs against the first-order win, not against a wish.
  4. Navigate: Mark every high-weight branch accept, mitigate, or avoid, each with an owner.
  5. Stream effects: Push every effect across four horizons (30, 90, and 365 days, and three years) so the far horizon is not discounted to zero.
  6. Trace: Draw the causal chain the system actually ran, and mark the link the original model missed.
  7. Review: Re-open the map after the decision ships, on a named rhythm.
  8. Effects on stakeholders: Name who actually pays: partner, regulator, agent network, or user.
  9. Assess systemic risk: Find the threshold where a local drag becomes a contagion.
  10. Mitigate: Install the smallest control that changes the later landing, not a new committee.

## Glossary (canonical term pages)

Prefer https://lenvanderhof.com/glossary/ and the Dutch twin https://lenvanderhof.com/nl/begrippen/.

- **SENSE** — https://lenvanderhof.com/glossary/sense/
  SENSE is a five-move strategic intelligence layer: Scan, Evaluate, Navigate, Simulate, Execute. Scan collects with a thesis. Evaluate turns a signal into a sourced claim with a confidence grade. Navigate makes the claim change a watchlist, an assumption register, or a Monday decision. Simulate rehearses the future before it arrives. Execute meets a decision or it never happened. The unit of work is one claim, not another unread summary. A model in the loop is not this layer until a claim can change a decision.
- **STACK** — https://lenvanderhof.com/glossary/stack/
  STACK is a five-layer architecture for repositories where humans and agents share work: Structure, Toolchain, Agent configuration, Connection, Knowledge and quality. Structure is the map an agent can navigate on minute one. Toolchain is the commands it may run. Agent configuration is versioned standing context. Connection is MCP with contracts and least privilege. Knowledge and quality is memory, evals, and CI, so a model upgrade does not silently lower the bar. The agent OS is versioned like the app.
- **CHORUS** — https://lenvanderhof.com/glossary/chorus/
  CHORUS is a six-skill coordination protocol for teams of 2 to 15 people who share work with specialised AI workers: Charter, Handoff, Orchestrate, Review, Update, Sync. A charter is seven fields that turn a vague intention into a contract any worker can be handed. A handoff is a six-field packet. They are not the same artifact. Org charts name reporting lines. CHORUS names coordination lines. A graph of model calls is not this protocol. If the receiver reconstructs the job from a chat log, you have a paste.
- **RESILIENT** — https://lenvanderhof.com/glossary/resilient/
  RESILIENT is a nine-capacity framework for founder performance under load: reappraisal, emotional granularity, somatic regulation, identity flexibility, learning orientation, interpersonal support, executive function, narrative integration, and titrated exposure. Capacity is infrastructure you train and budget, not a mood that returns after a holiday. The company can still move while the founder is the hidden single point of failure. Sleep, recovery, and the social cover that lets a founder tell the truth are part of the system, not extras.
- **BooksOS** — https://lenvanderhof.com/glossary/booksos/
  BooksOS is a gated production system for serious nonfiction written with AI agents: Brief, Bible, Map, Draft, Edit, Package, Launch. Each stage owns one decision and ends at a gate. Agents do specialist work inside the gates. Judgment, voice, and accountability stay human. A fluent draft that cannot survive a fact-check is not a book. It is inventory you will throw away. The operating system is the gates, not the model that drafted the chapter.
- **DOWNSTREAM** — https://lenvanderhof.com/glossary/downstream/
  DOWNSTREAM is an operator protocol for delayed consequences. The real cost of a decision often lands later, and somewhere else: another team, a later quarter, a customer you will not meet. The named moves are Detect, Observe, Weigh, Navigate, Stream effects, Trace, Review, Effects on stakeholders, Assess systemic risk, Mitigate. It is an operating heuristic, not a validated risk instrument. Name the second bend before you commit, then assign an owner to watch it. The book is in production; the protocol is already usable.
- **CALIBRATE** — https://lenvanderhof.com/glossary/calibrate/
  CALIBRATE is the practice of matching the strength of a judgment to the strength of the evidence, then updating when the facts move. Confidence is a report, not a personality trait. A loud claim with a thin source is a calibration failure, not conviction. The work is to score the claim, write what would change your mind, and revise the score when the evidence moves. The book that owns this name is available now in English and Dutch.
- **GRAIN** — https://lenvanderhof.com/glossary/grain/
  GRAIN is a retrieval loop: Gather, Rank, Assemble, Inspect, Navigate. Gather governs the corpus: sources, permissions, and versioning. Rank retrieves candidate passages and reranks them for the question actually asked. Assemble builds the context the model may see. Inspect traces a wrong answer back to the chunk and retrieval decision behind it. Navigate keeps the index fresh: TTL policy, re-ingest, and change detection. If you cannot point at the passage, you do not have retrieval. You have a fluent guess. The book that owns this name is available now in English and Dutch.
- **ROUTE** — https://lenvanderhof.com/glossary/route/
  ROUTE is a five-skill discipline for governing a mix of language models: Register, Objective typing, Utilize policy, Track, Evolve. Register catalogs every model with mandate, cost class, privacy class, and retirement status. Objective typing names what a task actually requires before a model is chosen. Utilize policy writes the routing rules, cascades, and fallbacks. Track measures cost, latency, and quality per route. Evolve promotes, demotes, and retires on a cadence. A leaderboard pick is a collection. ROUTE is a portfolio. The book that owns this name is available now in English and Dutch.
- **TRACE** — https://lenvanderhof.com/glossary/trace/
  TRACE is a five-step research loop: Target, Retrieve, Assess, Cross-check, Export. Target writes the research brief before any search runs. Retrieve obtains the source bodies, not one confident pass. Assess appraises what each body can bear per claim; tiers are source roles, not credibility weights. Cross-check records what each source supports or contradicts, and maps shared origins. Export is a versioned decision memo a co-founder can act on. Fluent collection is not research. TRACE is the verification structure. The book that owns this name is available now in English and Dutch.
- **ROSTER** — https://lenvanderhof.com/glossary/roster/
  ROSTER is a six-skill protocol for a one-person company that runs named specialist agents: Roles, Objectives, Skills and tools, Triggers, Evaluation, Rotation. Roles give each agent a charter. Objectives make the charter measurable. Skills and tools bound what the agent may use. Triggers decide when it runs. Evaluation scores output instead of vibe. Rotation retires, merges, upgrades, or migrates a role on scorecard evidence. One chatbot that never hands off is not a team. The book that owns this name is available now in English and Dutch.
- **LEVEL** — https://lenvanderhof.com/glossary/level/
  LEVEL is a five-move check for grading a claim before you change a belief, a habit, a roadmap, or a bet: Locate, Estimate, Value, Expose, Lean. Locate names the claim. Estimate the cost of being wrong. Value names the kind of support, then checks its quality. Expose the gap between that support and what the decision requires. Lean with a proportionate next step. The Evidence Ladder's seven rungs are kinds of support, not a quality scale. Confidence is not a grade. The book that owns this name is available now in English and Dutch.
- **CITATION** — https://lenvanderhof.com/glossary/citation/
  CITATION is an eight-part framework for becoming a source answer engines name: Cite, Identify, Triangulate, Anchor, Track, Iterate, Optimize, Normalize. Search eligibility is not a citation. Ranking in ten blue links is a different job. Answer engines quote the first clean capsule they can verify. Invented FAQ schema is not a shortcut; Google ended FAQ rich results on 2026-05-07. The book that owns this name is still being drafted; the framework is already how this site is built. A mention without a checkable source does not count.
- **CASCADE** — https://lenvanderhof.com/glossary/cascade/
  CASCADE is the personal cognitive companion to DOWNSTREAM: trace the next bend before you commit, in the meeting, in your own head. DOWNSTREAM is organizational. CASCADE is the habit that makes the protocol fire when nobody is watching. It is the pause that names who pays later, on a different surface, before you say yes. The book that owns this name is still in production; the habit is already usable. A first-order win with an unnamed later bill is not a decision. It is a delay.
- **AI agent** — https://lenvanderhof.com/glossary/ai-agent/
  An AI agent is a system that takes a goal, uses tools, and acts across steps without a human prompting each one. A chatbot answers a turn. An automation runs a fixed script. An agent is accountable only when it has a charter: purpose, inputs, outputs, authority ceiling, success criteria, review, and escalation. Without those seven fields, it is a demo that can spend money. A wrapper around a chat model is not this. Neither is a scheduled script.
- **MCP** — https://lenvanderhof.com/glossary/mcp/
  MCP, the Model Context Protocol, is a contract for how a model calls tools: what it may see, what it may change, and how the call is logged. The official split is host, client, and server, with tools, resources, and prompts as the primitives. Auth is optional; least privilege is not. A server without a contract is a hole in the repo. Founders should care about the interface and the permission, not the logo on the protocol. A clean server with no charter is still a hole with nicer plumbing.
- **RAG** — https://lenvanderhof.com/glossary/rag/
  RAG, retrieval-augmented generation, means the model must fetch named passages before it answers. The retrieval is the product. The generation is the last mile. If you cannot inspect which chunk was used, you do not have RAG. You have a language model with extra context stuffed in the prompt. A vector store is a component. GRAIN is the loop that keeps the answer honest. Adding embeddings is not a finished system. Naming the passage is.
- **Agent orchestration** — https://lenvanderhof.com/glossary/agent-orchestration/
  Agent orchestration is the assignment of work across humans and specialised agents so that handoffs, reviews, and state stay visible. A graph of model calls is not orchestration. Orchestration is who is chartered, what packet travels, what must be reviewed, and who sees the board. Without those four, adding agents adds waiting. CHORUS is one implementation of that operating layer. If you cannot point at the owner, the packet, the gate, and the board, you have a demo.
- **Human in the loop** — https://lenvanderhof.com/glossary/human-in-the-loop/
  Human in the loop means a named person holds an authority ceiling and a review gate. It is not a human glancing at a dashboard. The human charters the work, can stop a tool call, and owns the outcome. If no one can fail the output against written criteria, the loop is decorative. A vendor slide that says HITL without a named reviewer is a slogan. The name, the ceiling, and the fail criteria have to live in the runbook, not in the deck.
- **Evidence ledger** — https://lenvanderhof.com/glossary/evidence-ledger/
  An evidence ledger is a claim-level register: what was asserted, which source supports it, how strong that source is, and what would retire the claim. It is not a bibliography dumped at the end. Operators use it so a number cannot wander between chapters, books, or a back-cover line. If the claim cannot point at a row, it is not ready to print, pitch, or paste into an answer. The ledger is the product. The reference list is a rendering.
- **Strategic intelligence layer** — https://lenvanderhof.com/glossary/strategic-intelligence-layer/
  A strategic intelligence layer is the standing system that turns market, competitor, and internal signals into sourced claims a founder can act on. It is not a dashboard wall and not a weekly PDF. The test is whether a claim changes a watchlist, an assumption, or a Monday decision with a confidence grade attached. SENSE is one implementation: scan, evaluate, navigate, simulate, execute. If competitive work is still a founder reading Slack at midnight, you do not have a layer. You have a habit that will miss a week.
- **GEO** — https://lenvanderhof.com/glossary/geo/
  GEO, generative engine optimization, is the work of becoming a source that answer engines can retrieve, interpret, extract, and verify. It is not a bag of hidden prompts. Eligibility, extractable answers, original value, and evidence that survives scrutiny are the levers. Ranking in ten blue links is a different job. Google does not require schema, chunking, or an llms.txt file for AI Overviews. A page that cannot be quoted cleanly will not be cited, even if it ranks.
- **AEO** — https://lenvanderhof.com/glossary/aeo/
  AEO, answer engine optimization, is the craft of writing a page so an answer engine can quote a clean capsule: a direct answer near the top, named entities, visible sources, and no invented FAQ schema. GEO is the system. AEO is the page. They fail together when the first screen is a brand manifesto. Google ended FAQ rich results on 2026-05-07; a fake FAQPage is a liability, not a tactic. If the first screen cannot be quoted, the rest of the page will not save it.
- **Override doctrine** — https://lenvanderhof.com/glossary/override-doctrine/
  Override doctrine is the written list of decisions an agent may never take: spending above a ceiling, speaking as the company, changing production data, or closing a judgment the founder still owns. It is the complement of a charter. The charter says what the agent does. The doctrine says what it must refuse. A refuse-list that lives in a Slack thread is not doctrine. It is a hope. Write it before the first tool that can send, spend, or delete.

## Glossary-only names (no standalone page yet)

- **CASCADE** (the-second-order-thinker, in production): Personal cognitive companion to DOWNSTREAM: trace the next bend before you commit, in the meeting, in your own head.
- **COMMAND** (the-crisis-command, in production): Crisis as a protocol with named triggers and an escalation path, not as a personality that appears under load.
- **CITATION** (the-findability-engine, in production): Become a source answer engines name: Cite, Identify, Triangulate, Anchor, Track, Iterate, Optimize, Normalize.
- **TRUST** (marketing-without-manipulation, in production): Sell without dark patterns. The page tells the same story the product will keep.
- **LEDGER** (the-cognitive-budget, in production): Treat attention and decision quality as a budget with a closing balance, not a mood.

## Series

Systems for the Strategic Self. An operator-grade nonfiction series on founder psychology, agentic AI systems, and evidence-based decision-making. Each title installs one named framework the reader can run. Every title is two native editions, English and Dutch, never a machine translation.

This file is a convenience extract. The HTML pages remain authoritative if anything drifts.