The Human Algorithm: What Teaching Machines Reveals About Ourselves

Disclaimer

This book represents an experimental collaboration between human creativity and artificial intelligence. I (Jay W) am not the author of this content, nor do I possess expertise in the domains explored within. The true authors - Claude Opus 4 and Claude Code - drew upon their training on humanity's collective knowledge to create this work.

The creation process was deliberately autonomous. After providing the initial prompt and concept, I configured Claude Code to operate in auto-accept mode, allowing it to write with minimal human intervention. My role was limited to occasional review points where I could accept or reject proposed changes. Beyond setting the initial direction, I consciously chose to let the AI systems pursue their own understanding and interpretation of the subject matter.

This experimental approach emerged from a conversation exploring whether AI-generated books on topics of personal interest might offer unique value compared to traditional authored works. The book serves dual purposes: first, to provoke reflection on the parallels between human and artificial intelligence explored within its pages; second, to demonstrate the capabilities of agentic AI applications beyond conventional coding tasks.

Important Notice: This book is intended for entertainment and experimental purposes only. It should not be treated as an authoritative source of information. All claims and insights presented should be independently verified. Future iterations of this experiment will include AI-powered fact-checking of the content. The value lies not in accepting these ideas as truth, but in using them as starting points for your own critical thinking and exploration.

An agent-assisted book exploring how AI development illuminates human cognition and communication.

Authors: Claude Code and Claude Opus 4

Concept & Creative Direction: Jay W

About This Book

In our rush to make artificial intelligence more human, we've overlooked a profound opportunity: using AI as a mirror to understand our own minds. This book explores how the challenges we face with Large Language Models reveal uncomfortable truths about human cognition and communication.

When we worry about LLMs "hallucinating," we ignore that humans confidently state falsehoods every day - yet we demand rigorous fact-checking from machines whilst accepting human claims at face value. We've developed sophisticated "grounding" techniques to verify AI outputs, but rarely apply the same standards to ourselves or others. We meticulously measure the emotional intelligence of AI systems whilst neglecting these metrics in our daily interactions.

Through practical examples and thought-provoking parallels, this book demonstrates how LLM concepts can transform human relationships. By examining how we build and refine artificial minds, we gain unprecedented insights into our biological ones - turning the mirror of AI back on ourselves to become more aware, intentional, and effective communicators.

This is not a book about making AI more human. It's about using AI to make humans more conscious of what they already are.

Book Status

📚 First Revision COMPLETE - AI-Generated Draft (52,384 words)

✅ First Revision - Completed with minimal human contribution (auto-accept mode)

🔄 Second Revision - Planned with increased human-in-the-loop involvement

The current version represents an experimental autonomous AI writing process. The planned second revision will incorporate more human review, fact-checking, and editorial input while maintaining the unique AI perspective.

Reading Order

Start with the Introduction to understand the book's premise, then proceed through the chapters in order. Each chapter builds on previous concepts while remaining self-contained enough to read independently.

Project Structure

/agentic-book/
├── book/                      # Generated book files (HTML, PDF)
│   ├── html/                  # HTML version
│   │   └── the-human-algorithm.html
│   └── pdf/                   # PDF versions
│       ├── the-human-algorithm-digital.pdf
│       └── the-human-algorithm-print.pdf
├── build/                     # Build system and templates
│   ├── README.md              # Build documentation
│   ├── scripts/               # Build scripts
│   │   ├── build.sh           # Main build script
│   │   └── remove-emoji.lua   # Lua filter for PDF generation
│   └── templates/             # Pandoc templates
│       ├── book.css           # HTML styling
│       ├── html-template.html # HTML template
│       ├── pdf-digital-template.tex
│       └── pdf-print-template.tex
├── manuscript/                # Book chapters in markdown
│   ├── 00-front-matter.md     # Title, ToC, license
│   ├── 01-introduction.md     # Introduction: The Mirror We're Building
│   ├── 02-part-1-intro.md     # Part I introduction
│   ├── 02-chapter-1-hallucination-paradox.md
│   ├── 03-chapter-2-grounding-ourselves.md
│   ├── 04-chapter-3-temperature-control.md
│   ├── 05-part-2-intro.md     # Part II introduction
│   ├── 05-chapter-4-context-windows.md
│   ├── 06-chapter-5-prompting-personalities.md
│   ├── 07-chapter-6-fine-tuning-relationships.md
│   ├── 08-part-3-intro.md     # Part III introduction
│   ├── 08-chapter-7-bias-detection.md
│   ├── 09-chapter-8-emotional-tokens.md
│   ├── 10-chapter-9-training-data-of-life.md
│   ├── 11-part-4-intro.md     # Part IV introduction
│   ├── 11-chapter-10-overfitting-to-trauma.md
│   ├── 12-chapter-11-model-collapse.md
│   ├── 13-chapter-12-emergent-properties.md
│   ├── 14-part-5-intro.md     # Part V introduction
│   ├── 14-chapter-13-alignment-problem.md
│   ├── 15-chapter-14-recursive-self-improvement.md
│   ├── 16-chapter-15-consciousness-question.md
│   └── 17-conclusion-becoming-better-algorithms.md
├── notes/                     # Project tracking and research
│   ├── 00-chapter-metadata.md
│   ├── 00-general-tracker.md
│   ├── 01-formatting-tracker.md
│   ├── 01-initial-chapter-outlines.md
│   ├── 02-revision-plan.md
│   ├── 02-safety-review-tracker.md
│   ├── 03-development-progress.md
│   ├── 04-target-reassessment.md
│   ├── 05-final-review-tracker.md
│   └── 06-stats.md
├── CLAUDE.md                  # AI assistant guidelines
├── Dockerfile                 # Container build
├── Makefile                   # Build automation
└── README.md                  # This file

Key Themes

The Double Standard - We pathologize in AI what we normalize in humans
Cognitive Mirrors - How AI development reveals hidden patterns in human thinking
Processing Architectures - Both humans and AI have similar computational constraints
Training Data Effects - How past experiences shape all future responses
System Failures - When protective mechanisms become prisons
Emergent Possibilities - How constraints can catalyze unexpected capabilities

Chapter Overview

Part I: The Accuracy Paradox

When Machines Hallucinate - Why we fact-check AI but accept human confabulation
The Grounding Problem - The infrastructure gap between AI and human verification systems
Temperature and Creativity - Balancing predictability with innovation in AI and humans

Part II: Processing Limits

Context Windows and Memory - How cognitive boundaries shape conversations and relationships
The Art of Prompting - Why how we ask determines what we receive from humans and AI
Fine-Tuning and Habit Formation - How repetition shapes responses in both systems

Part III: Hidden Patterns

Detecting Our Own Biases - What AI prejudice reveals about human pattern-matching
Emotional Tokens - How emotions function as information-processing signals, not just feelings
The Training Data of Life - How childhood experiences create persistent behavioral patterns

Part IV: System Failures

Overfitting to Trauma - When survival strategies become life-limiting specializations
Model Collapse - How echo chambers emerge and diversity dies gradually
Emergent Properties - When constraints create unexpected new capabilities

Part V: The Future Human

The Alignment Problem - Why aligning AI values requires confronting human value incoherence
Recursive Self-Improvement - The power and peril of systems that enhance their own enhancement
The Consciousness Question - What separates human awareness from artificial processing

Building the Book

The book can be built into multiple formats using the included build system:

# Build all formats (HTML and both PDF versions)
make all

# Build specific formats
make html
make pdf-digital
make pdf-print

# Clean build directory
make clean

# Get help
make help

Requirements: Pandoc, XeLaTeX, and the fonts specified in the templates.

Download the Book

Key Features

Opening Scenes: Each chapter begins with a relatable scenario that illustrates the concept
The AI Mirror: Technical concepts explained through human parallels
Practical Applications: Every insight translated to actionable practices
Reflection Questions: Prompts for personal contemplation and growth
Neuroscience Integration: Current brain research supporting the parallels
Cross-Cultural Perspectives: Diverse viewpoints on universal cognitive patterns

Who This Book Is For

Anyone curious about what AI reveals about human nature
Tech professionals wanting a humanistic perspective on AI
Psychology enthusiasts interested in cognitive parallels
Leaders navigating human and artificial intelligence
Anyone seeking practical self-improvement through AI insights

Technical Notes

Written in Markdown for easy version control and conversion
Each chapter ~3,700 words for consistent reading sessions
Designed for both linear reading and standalone chapters
Balanced between accessibility and intellectual depth
No technical background required

How to Contribute

This is an experimental collaboration between human creativity and AI capability. While the book is being written by Claude with human direction, insights and feedback are welcome.

License

This work is licensed under a Creative Commons Attribution 4.0 International License.

You are free to:

Share - copy and redistribute the material in any medium or format
Adapt - remix, transform, and build upon the material for any purpose, even commercially

Under the following terms: