AI 原生开放教材

外观

Vibe Coding 101 for Software Engineers系统课程·英文原文 ↗原件 ↓仓库 ↗

Week 1: Slides Outline ​

"There's a new kind of coding I call 'vibe coding', where you fully give in to the vibes, embrace exponentials, and forget that the code even exists." — Andrej Karpathy, February 3, 2025

Slide 1: Title ​

Vibe Coding 101: Week 1The History of Vibe Coding & PRD-First Development

Slide 2: Pulse Check (Start) ​

Before we begin, rate your confidence (1-5):

  1. How comfortable are you using AI coding tools?
  2. Have you ever written a PRD before?

Slide 3: The Tweet That Changed Everything ​

Show Karpathy's original tweet (February 3, 2025)

"There's a new kind of coding I call 'vibe coding', where you fully give in to the vibes, embrace exponentials, and forget that the code even exists."

— Andrej Karpathy

Source: x.com/karpathy/status/1886192184808149383

Slide 4: Who is Andrej Karpathy? ​

  • Former OpenAI researcher
  • Former Director of AI at Tesla (Autopilot)
  • PhD from Stanford (computer vision)
  • Created popular deep learning courses
  • 2023: "The hottest new programming language is English"
  • 2025: Coined "vibe coding"

Slide 5: The Problem with Pure Vibe Coding ​

Risks ​

  • Hallucinations: AI can invent packages or APIs that don't exist
  • AI Slop: Code that looks right but doesn't work
  • Security: Slopsquatting attacks (malicious packages with hallucinated names)
  • Tech Debt: "Wide but shallow" code—hairballs that are hard to maintain

Slide 6: The Spectrum ​

PURE VIBE CODING              DISCIPLINED VIBE CODING
(Weekend Projects)            (This Course)

• "Accept All" always         • Review diffs critically
• Skip reading code           • Understand what's generated
• Copy-paste errors           • Verify with tests & PRD
• Works until it doesn't      • Production-ready mindset

Slide 7: Our Solution: PRD-First Development ​

Before You Write a Single Prompt... ​

Create a Product Requirements Document (PRD)

Why?

  • Clarity for YOU
  • Guidance for the AI
  • Prevents scope creep
  • Catches hallucinations early

Slide 8: The PRD Structure ​

  1. Overview - What/Who/Why
  2. Core Features - MVP only
  3. Non-Goals - CRITICAL for AI!
  4. Technical Constraints - Language, deps, standards
  5. Success Criteria - How do we know it works?
  6. Phases - Sequential, verifiable chunks

Slide 9: Why Non-Goals Matter ​

"AI cannot infer from omission. If the specification does not explicitly state 'do not implement user authentication in this phase,' an agent might reasonably add it."

Without Non-Goals:

  • AI adds features you didn't ask for
  • Scope keeps expanding
  • You spend time removing unwanted code

Slide 10: The Six Essential Areas ​

From analysis of many agent configurations:

  1. Commands - Full commands with flags
  2. Testing - Framework, coverage, locations
  3. Project Structure - Directory organization
  4. Code Style - Real examples, not just rules
  5. Git Workflow - Branches, commits, PRs
  6. Boundaries - What agents should NEVER do

Source: Addy Osmani

Slide 11: Tool Setup ​

Claude Code (macOS/Linux):

bash
curl -fsSL https://claude.ai/install.sh | bash
claude
/init       # Generate CLAUDE.md

Gemini CLI (macOS/Linux):

bash
npm install -g @google/gemini-cli
gemini

Windows PowerShell:

powershell
# Claude Code
winget install Anthropic.Claude
claude

# Gemini CLI
npm install -g @google/gemini-cli
gemini

Slide 12: The Workflow Cycle ​

1. EXPLORE → 2. PLAN → 3. CODE → 4. COMMIT
     ↑                              |
     └──────────────────────────────┘
              (Iterate)

Slide 13: PRD to Code ​

  1. Show the greeting-cli PRD
  2. Start AI session
  3. Share PRD with AI
  4. Ask for plan (don't code yet!)
  5. Review plan together
  6. Implement Phase 1 only
  7. Verify it works
  8. Continue to Phase 2...

Slide 14: Pulse Check (End) ​

Reflect on today's session:

  1. What's your biggest takeaway about disciplined vibe coding?
  2. What project will you build your first PRD for?

Slide 15: What's Next ​

Week 2: How AI Coding Actually Works

  • Tokenization, context windows, temperature
  • The AI tool landscape
  • Prompt engineering fundamentals

Slide 16: Questions? ​

Resources:

  • Course repo: [link]
  • Claude Code docs: code.claude.com
  • Gemini CLI docs: ai.google.dev/gemini-cli

Instructor: Goker Ezberci | gokerez@gmail.com