← Back to courses
T1ENS6 · 8/29 (Sat)

Demo Day — Open Your Dojo

★★★★★★ Mastery

📋 Key Takeaways

  • You made it — graduation night · four real graduates demo live · six weeks ago they were you · the only gap between audience and stage is six weeks and a room
  • Tonight's four builders: Jeffrey (Five-Elements Relationship Coach) · Jinxia (Holistic Financial Planning) · Minzi (CoolArch Designer) · Grace (Trading Platform Selector) — real students, real agents
  • The student capability ladder S1→S5 — every step a real student who did it: see what AI is · summon the hero · write your own script · the three-layer shuttle · ship a working agent
  • The agent market is a blue ocean — value isn't handed to you, you find it in your own world · $400 per booked date proves it · polish one agent that truly delivers and it monetizes
  • Academy meets platform: the Academy teaches you to build · the platform carries and monetizes it · the members ship real projects together — this graduation is one proof, yours is next
  • S6 homework · publish · share · let 1 real person use it · interview them · write a short report — an agent is not a document, it is a product meant to be used. AI is here. Just build it.

🎬 Video Replay

Demo 1 11:02
Demo 2 11:54
Demo 3 7:20
Demo 4 16:14
Closing 25:52

📝 Course Notes

Expand full notes
# Session 6 · Demo Day — Open Your Dojo

**Trait: Mastery (出师).** Graduation night. Six weeks ago the four builders on stage were
sitting where you are. Tonight your Wukong stops being a script you polish alone and
becomes a product a real person actually uses.

## Part 0 · You made it

Tonight's arc: **Intro** (meet the four graduates) → **Demos** (four builders, real
agents, live) → **Closing** (what's next · Q & A). No lecture after the peak — the demos
*are* the peak, and everything after descends to one ask.

## Part 1 · Tonight's four graduates

Four real students, four real agents:

- **Jeffrey · Five-Elements Relationship Coach** — turns abstract relationship labels into
practical actions you can try this week.
- **Jinxia · Holistic Financial Planning Agent** — sees income, debt, savings, insurance,
and goals as one picture, not one product at a time.
- **Minzi · CoolArch Designer** — helps engineers generate and evaluate cooling-system
design choices in minutes, not days.
- **Grace · Trading Platform Selector** — picks the right brokerage from a real, dated
dataset: verdict first, always cited.

**Six weeks ago, they were you.** The only thing between the audience and the stage is
six weeks and a room.

## Part 2 · The student capability ladder (S1 → S5)

Every step is a real student who actually did it — not an arrow on a slide:

- **S1 · See what AI really is** — the LLM is Origin Wukong: no memory, not magic.
- **S2 · Learn to summon the hero** — call a specific specialist; stop asking ChatGPT
vaguely.
- **S3, S4 · Write your own script** — my model, my data, my authority.
- **S3.5 · See the three-layer shuttle** — Symptom → Essence → Philosophy: solve your own
problems.
- **S5 · Ship a working agent** — write the script, feed the KB, plan the fallback, go live.

## Part 3 · The agent market is a blue ocean

The value isn't handed to you — you find it in your own world, work, and experience.
Silicon Valley, mid-2026: an agent that chats for a guy and lands a real date is priced at
**$400 per booked date**. We don't teach that one, but it proves the rule — polish an agent
that truly delivers value and it monetizes.

So we asked: what agent would a high schooler build? Three ideas from a hackathon we're
designing — **Get My Evening Back** (plans homework so they get their night back) ·
**Flip** (a trivia duel bot on the idols they love) · **Vibe Relief** (breathes with them
through exam panic). Useful agents, found in their own lives. Find the value in *your*
domain — polish one until it delivers, and a blue ocean pays.

## Part 4 · Academy meets platform

We don't stop at teaching. The **Academy** teaches you to build; the **platform** carries
and monetizes it; the **members** ship real projects together. This graduation is one
proof. Yours is next.

**Join to be a member** — three tiers:

- **Start · Free** — free courses, ship your first agent.
- **Charter · $499 one-time** ( → **$449** with code `SILICONROSHI`) — one payment, no repeat fee for as long as the program runs; build your website,
your channel, and the agent your customers use.
- **AI Principal** — run AI programs with the school.

The person next to you today may be your co-founder tomorrow. Your talent is not meant to
be buried.

## Part 5 · Learn anything super fast · 10 prompts (handout)

Open ChatGPT or your agent, paste, and learn faster than school ever taught you:

1. **Explain** — "Explain this like I'm 10, then test me."
2. **Visualize** — "Turn this topic into 8 visual sticky notes."
3. **80/20** — "Teach me the 20% that matters most."
4. **Map** — "Make a visual map connecting the key ideas."
5. **Quiz me** — "Quiz me one question at a time."
6. **Deep dive** — "Do a deep research on this topic."
7. **Best practice** — "Show me the best practice of the industry."
8. **Pitfalls** — "Show me the top 3 mistakes beginners make."
9. **Mental models** — "Give me 3 mental models to understand this."
10. **Opposing view** — "Tell me the adversary facts and views."

## Homework

- **HW (submit · `ENS6-SHIP` · due Sat 9/2 9:00 AM PT):** PUBLISH your agent to WukongGPT,
SHARE the link with at least 1 real person (**not you**), let them USE it, INTERVIEW them
with the **5-question template** below, and write a short user-feedback report in three
sections: **who + how** (background, duration, what they did) · **what they said** (paste
exact quotes) · **what I learned + next step** (most surprising finding + one specific
change). Record exact words — their quotes are 100× more valuable than your summary.

**5-question interview template** (ask exactly these, in order):
1. What did you use this agent for? (let them describe in their own words)
2. Was there a moment you thought "oh, it gets me"? Which line?
3. Was there a moment you thought "it's rambling / dodging the question"? Which line?
4. If you'd use it again, for what? If not, why not?
5. Describe this agent in one word.

See the S6 Homework Guide in materials for the full report template and grading rubric.

An agent is not a document — it is a product meant to be used.

**AI is here. Just build it.**

✏️ Tarea

HW (submit · Golden Hoop 6 · Mastery): PUBLISH your agent to WukongGPT, SHARE the link with at least 1 real person, let them USE it, INTERVIEW them, write a short user-feedback report. ⚠️ That person CANNOT be you. Self-testing ≠ real user. ⚠️ Record exact quotes. Their words are 100x more valuable than your summary. ① WHO + HOW: background, duration, what they used the agent for. ② WHAT THEY SAID: interview with the 5-question template, paste exact quotes. ③ WHAT I LEARNED + NEXT STEP: most surprising finding + what to change.

⏰ Fecha límite: 9/2 Sat 9:00 AM PT

→ Enviar tarea (ENS6-SHIP)
🤖 Criterios de evaluación IA
Dimensión Peso Descripción
Publish + Share 20% Agent is published on WukongGPT + link shared with at least 1 real person + that person actually used it. Not published = low. Self-testing does not count.
Interview record 35% Interviewed a real user with the 5-question template, **pasted exact quotes**. Exact words beat summaries. "She thought it was fine" is not a quote; "how did it know what I was thinking" is. Not all 5 need to shine but at least 3 should have substantive answers.
Surprising finding 25% At least 1 finding you did NOT expect — user reaction differed from your design intent. "Matched my expectations" suggests soft scenarios or shallow observation. The valuable insight is the gap between your design focus and their actual reaction.
Next step 20% Based on user feedback, name the SPECIFIC change (which .md / step / feature). "Keep polishing" = low. Next step should logically follow the surprising finding.

90-100 Published + real user used it + interview has quotes + deep surprise + concrete next step · 60-89 Published + real user + interview occasionally vague + next step slightly generic · 40-59 Published but no real user / user used it but no interview / interview is all summary no quotes · 20-39 Not published / only self-tested · 1-19 Nearly empty