硅基团队 Silicon Team

这不是一本"如何用 ChatGPT"的书。这本书讲的是：一个人如何用 AI agent 构建一支工程团队，然后用这支团队交付真实产品。不是 demo，不是概念验证——是有 deadline、有用户投诉、有 production bug 的真实工程项目。

This is not a "how to use ChatGPT" book. It's about how one person builds an engineering team with AI agents and ships real products. Not demos, not proofs of concept — real engineering projects with deadlines, user complaints, and production bugs.

核心论点：约束越好，自由越大。设计好闭环、给好约束、做好角色隔离，按下启动键，去睡觉。醒来检查结果。

Core thesis: Better constraints, greater freedom. Design the closed loop, set the constraints, isolate the roles, press start, go to sleep. Check results in the morning.

目录 Table of Contents

• 00
  前言：从一个人到一支团队

  这本书讲的是一个人如何用 AI agent 构建一支工程团队
  00
  Preface: From One Person to a Team

  How one person builds an engineering team with AI agents
• 01
  协议层：CLI 作为 Protocol Layer

  CLI 如何解耦 memory 和 agent，以及三层分发模型的设计
  01
  The Protocol Layer: CLI as Protocol Layer

  How CLI decouples memory from agents, and the three-layer distribution model
• 02
  Harness-Native Engineering

  约束系统设计——Mechanical Gate 如何替代概率性判断
  02
  Harness-Native Engineering

  Constraint system design — how mechanical gates replace probabilistic judgment
• 03
  多硅基协作

  Agent routing、Spawn vs Delegate、多 agent 编排
  03
  Multi-Silicon Collaboration

  Agent routing, Spawn vs Delegate, multi-agent orchestration
• 04
  自主运行

  OPC pipeline、iterative review、autonomous loop
  04
  Autonomous Operation

  The OPC pipeline, iterative review, and autonomous loops
• 05
  Agent 时代的商业逻辑

  从工程视角转向商业视角——定位、叙事、投资论点
  05
  Business Logic in the Agent Era

  From engineering to business — positioning, narrative, investment thesis
• 06
  工程手札

  17 个用血换来的具体教训
  06
  Engineering Field Notes

  17 hard-won engineering lessons
• 07
  从记忆到书

  227 张卡片如何变成一本书
  07
  From Memory to Book

  How 227 cards became a book