How the Book Is Organized
The Book of AI is a practical guide to agentic engineering: understanding the parts of an AI system, choosing them deliberately, and assembling them into useful, controllable stacks. It is written for engineers and technical practitioners who want more than a catalog of products or model rankings.
The chapters move from models to inference, agent harnesses, and complete working systems. They emphasize practical trade-offs—including capability, cost, privacy, local control, and operational safety. Because tools and model capabilities change quickly, the book’s core chapters provide durable concepts while the monthly digest and resource lists track the moving landscape.
How to Use the Book
- Read the chapters in order for a guided progression from model fundamentals to agentic systems.
- Use the chapters as a reference when making a specific design or deployment decision.
- Treat product names, benchmarks, and dated examples as snapshots; check the linked sources and digest for current details.
- In Chapter 5, follow the modular blocks, labs, and microlabs to build a working stack step by step.
Chapter Roadmap
| Chapter | Focus | What it helps you do |
|---|---|---|
| 1. Introduction | Agentic engineering and the layers of an agentic system | Understand the book’s scope, audience, and practical orientation. |
| 2. Models | Model types, capabilities, access, context, evaluation, and training data | Select models based on the work an agent must perform, not just leaderboard position. |
| 3. Running a Model | Local inference, hardware, hosted services, and hybrid deployments | Put models behind usable interfaces and understand deployment trade-offs. |
| 4. Agents and Harnesses | Agent loops, tools, MCP, memory, and execution environments | Understand the runtime and safeguards that turn a model into an agent. |
| 5. Building a Working Stack | Models, harnesses, tools, and operational practices assembled together | Build and operate a practical stack through blocks, labs, and microlabs. |
| 6. Miscellaneous Tools and Supporting Infrastructure | Additional tools, model selection and modification, and open-source resources | Explore supporting capabilities that do not belong to one layer alone. |
| 7. Resources | Curated articles, videos, publications, books, and people | Continue learning from selected external sources. |
Appendices and Living References
The appendices provide a glossary, source material and references, general notes, and prompt-engineering notes. The digest is maintained in its own folder and included as a Digests part of the book. Use it as a periodic snapshot of developments in agentic engineering, not as a replacement for the foundational chapters.
Repository File Structure
The Quarto project configuration defines the book’s chapter and appendix order. The chapter files contain the main narrative; underscore-prefixed files under chapters/blocks/ are reusable Chapter 5 sections, labs, and microlabs included by the chapter.
AI/
├── _quarto.yml # Book configuration and reading order
├── README.md # Project overview and progress
├── LICENSE
├── index.qmd # Book landing page
├── chapter-structure.qmd # This overview
├── references.bib
├── ieee.csl
├── scratch.md # Working notes
├── chapters/
│ ├── 01-introduction.qmd
│ ├── 02-models.qmd
│ ├── 03-running-a-model.qmd
│ ├── 04-agents-and-harnesses.qmd
│ ├── 05-building-a-working-stack.qmd
│ ├── 06-miscellaneous-tools.qmd
│ ├── 07-resources.qmd
│ ├── _05.1-nvidia-working-stack.qmd
│ ├── _07.1-youtube-xml-feeds.qmd
│ └── blocks/
│ ├── _block-0-pre-flight.qmd
│ ├── _block-0.1-pre-flight-gnu-linux-distributions.qmd
│ ├── _block-1-models.qmd
│ ├── _block-2-running-models.qmd
│ ├── _block-3-agents-harnesses.qmd
│ ├── _block-4-orchestrating-swarm.qmd
│ ├── _block-5-misc-tools-utilities.qmd
│ └── labs/
│ ├── _lab-0.1-linux-distribution-choices.qmd
│ ├── _lab-0.2-vscode.qmd
│ ├── _lab-0.3-runtime-isolation.qmd
│ └── microlabs/
│ ├── _0.1.1-rhel.qmd ... _0.1.14-alpine.qmd
│ ├── _0.2.1-vscode-installation.qmd
│ ├── _0.2.2-vscode-extensions.qmd
│ ├── _0.2.3-vscode-theme-settings.qmd
│ ├── _0.3.1-mise.qmd
│ └── _0.3.2-uv-python.qmd
├── appendices/
│ ├── glossary.qmd
│ ├── references.qmd
│ ├── SOURCES.qmd
│ ├── notes.qmd
│ └── prompt-engineering-notes.qmd
├── digest/
│ ├── digest_index.qmd
│ ├── 2026-10-October-digest.qmd
│ └── 2026-11-November-digest.qmd
├── assets/
├── configuration/
├── examples/
├── scripts/
├── tests/
├── _extensions/
├── archive/
├── _book/ # Generated book artifacts
└── docs/ # Generated website output
The source files are the maintained material; _book/ and docs/ are generated outputs. The monthly digest and linked resources keep the book’s map current as tools, models, and practices evolve.