ai llm ai/agenticai agents ai/books/anillustratedguidetoaiagents
Preface
Intuition-First Philosophy
build a strong mental picture first, then add the technical details.
The intuition-first philosophy is a way of teaching technical ideas by helping the reader first understand how something works conceptually, before focusing on detailed implementations, frameworks, or code.
For AI agents, this means learning the basic mental model first: an agent has an LLM that reasons, memory that preserves useful information, tools that let it act, planning that organizes steps, and reflection that helps it improve its approach.
The philosophy also relies heavily on visual explanations. Diagrams and illustrations make abstract processes easier to picture, so readers can understand the relationships between different parts of an agent instead of just memorizing definitions.
Most importantly, it focuses on fundamental ideas that are unlikely to change quickly. Specific agent frameworks and technologies may come and go, but if you understand the underlying principles, you can more easily understand new systems as they appear.
Part 1. The Anatomy of an AI Agent
1. Introduction to AI Agents
2. LLMs and AI Agents
3. Resoning LLMs and AI Agents
4. Memory in AI Agents
5. Tool Usage, Learning, and Protocols
6. Planning and Reflection
7. Evaluating Agents
Part 2. Specialized Agents
8. Multi-Agent Systems
9. Multi-Model Understanding
10. Code Agents and Code LLMs