caozhiyi

caozhiyi / ai-programming-book

Public

《AI编程的第一性原理》开源书。从大模型的底层物理约束出发，推导 Agent 运行机制与 AI 编程系统的工程化实践。

agentic-coding ai llm programming
14
0
85% credibility
Found May 17, 2026 at 14 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
AI Summary

This is a free online book written in Chinese that teaches the fundamental principles of AI programming. Rather than focusing on specific tools or code snippets, it takes readers through a systematic journey covering how large language models work internally, how to build AI agents that can take actions, techniques for managing conversation context and memory, strategies for choosing the right AI architecture, and the practical challenges of engineering AI systems. The book is structured across five volumes with 17 chapters and practical appendices, designed for developers and technical professionals who want to understand AI programming at a deeper level rather than just copying examples.

How It Works

1
📚 You discover an AI programming book

You find a free online book that promises to teach you the real principles behind AI programming, not just surface-level tricks.

2
🧠 You learn how AI models actually think

The book opens your eyes to understanding what happens inside an AI model when you ask it questions, making the mysterious feel approachable.

3
🤖 You discover AI agents that can take action

Suddenly you understand how AI can go beyond chatting—it can use tools, call functions, and work alongside other AI helpers to get things done.

4
💭 You explore memory and context

You learn the clever techniques to help AI remember your conversations, manage costs, and bring in knowledge from outside sources.

5
You choose your learning path
🏗️
Architecture Path

For those building systems: learn how to choose the right approach and design secure, reliable AI applications.

🔧
Engineering Path

For those working with AI: understand the real challenges of unpredictable systems and how to manage AI teams.

🎓 You gain true understanding

You emerge with a solid mental model of how AI programming really works—from asking questions to getting results—and the confidence to apply it.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 14 to 14 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is ai-programming-book?

This is an open-source technical book in Chinese that teaches AI programming from first principles. Rather than showing you how to copy-paste prompts, it walks through the underlying physics of large language models to explain why agents behave the way they do. The book covers everything from token economics and context engineering to multi-agent collaboration and security alignment, structured across five volumes that take you from understanding LLM internals to building production AI systems.

Why is it gaining traction?

The "first principles" angle is the hook. Most AI programming content teaches you to use tools like GitHub Copilot or Claude without explaining the constraints that make those tools work or fail. This book starts with physical limitations like context window size and inference costs, then derives agent behavior from there. The table of contents reads like a curriculum: agents, MCP protocols, memory systems, RAG versus fine-tuning, and end-to-end architecture blueprints. It fills a gap for developers who want to understand the "why" behind AI engineering patterns.

Who should use this?

Backend engineers building AI-powered features who want to move beyond prompt engineering into actual system design. Technical leads evaluating AI toolchains for their teams. Developers working with GitHub Copilot, Claude, or similar tools who feel like they're fighting the system instead of understanding it. Chinese-speaking developers will get the most value, though the concepts translate.

Verdict

The credibility score sits at 0.85%, and with only 14 stars, this is an early-stage project from an unknown author. The structure is solid and the scope is ambitious, but there's no way to evaluate content quality since the actual book content is missing from this repository. Worth bookmarking and checking back on once content is populated, but not ready for serious study yet.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.