📋 Key Takeaways
- An LLM is a super probability model built on human language — it knows everything, but what you say determines what you get
- LLM vs Agent are two things: LLM = raw power (Wukong), Agent = LLM + system prompt + tools
- Same model, different prompts, drastically different results — prompts are your lever
- Your domain expertise (industry knowledge, client understanding) is your moat — AI gets cheaper, your world gets more valuable
- Six Golden Hoops: Fundamentals → Prompting → API Keys → Build Agent → Compete → Teach
🎬 Video Replay
YouTube: https://youtu.be/UucMswDPsVc
🔧 Materials
✏️ Tarea
HW1 (warm-up): Use AI to revise your resume — compare casual vs structured prompting. HW2 (submit): Answer two questions — ① What do I most want to achieve with AI? ② What was most memorable today & what do I want to learn next?
⏰ Fecha límite: 5/7 周三 9:00 PM PT
→ Enviar tarea (S1-SPARK)▶ 🤖 Criterios de evaluación IA
| Dimensión | Peso | Descripción |
|---|---|---|
| Feedback quality | 40% | Whether course feedback is specific and constructive |
| Effort | 20% | Whether answered with care, substantial and specific rather than perfunctory |
| Authenticity | 15% | Whether these are the student's genuine thoughts, not AI-generated boilerplate. Real: colloquial, specific personal context, emotional, occasional typos. AI: over-polished, many parallel clauses, heavy use of AI cliches. |
| Course comprehension | 15% | Whether shows understanding of class content (mentions LLM, Agent, prompts, Wukong/Master Roshi metaphor, Golden Hoops, tokens, etc.) |
| Goal clarity | 10% | Whether the AI wish is specific, actionable, and personal |
90-100 Highly thoughtful, clearly the student's own work · 70-89 Completed seriously, with the student's own ideas · 50-69 Completed but perfunctory, or partially AI-generated · 30-49 Very perfunctory or mostly AI-generated · 1-29 Nearly empty or entirely AI cliches