Master Hub · Course Documentation
Directing the Machine A three-week intensive · Seoul · ~50 students, ages 15–18
The 2026 AI job market hires on delivery evidence — repos, deployed products, eval suites — not credentials. So every lesson hands the student a real artefact and a named, transferable skill. Three professional seasons: judge (Week 1), make (Week 2), build (Week 3).
0 The aim format: I can use [a new quantifiable skill] to create [an artefact] which is used for [a purpose].
0 The doubling: every lesson teaches one industry skill AND one AI tool/skill that hyperdrives it. The AI is the level-up, not the substitute.
0
3 weeks
14 days
32 lessons
120 stages
261 scenes
Every lesson opens into its stages (sequences) and their child activities (scenes), each with a UI metaphor, a governing dynamic, and a named game-mechanic — plus a clickable panel of reviews and a ready-to-use generation prompt. Click any lesson to open it.
Features ▤ Projects
A shared workspace with a persistent knowledge base + instructions. On Team plans, shareable across the org with can-use / can-edit permissions. Static background context loaded into chats.
"Our team's shared brain."
Best for Shared source material; team continuity across days
◆ Skills (SKILL.md)
A folder with a SKILL.md (YAML name + description, then instructions). Claude auto-invokes it when relevant. Portable, task-activated procedures usable everywhere. Owners can provision org-wide.
"I taught Claude a repeatable job once, and now it just does it."
Best for Packaging a workflow; the day's specialist
⬢ Cowork
A desktop agent that runs on the student's machine, plans, runs code in an isolated VM, and writes real files to disk. On by default for Team. Projects in Cowork are local, not shared.
"I'm directing an AI engineer who actually builds things."
Best for The real build work; the engineer/founder LARP
❰❱ CLAUDE.md
A plain-markdown file of standing rules/conventions Claude reads and obeys for a build. Version-controlled context.
"I wrote the rules and the AI holds them all session."
Best for Teaching context-engineering; enforcing a standard
⌗ Claude Code
The command-line agent for directing builds via Explore→Plan→Code→Commit.
"I ship software by directing, not typing."
Best for The hands-on dev lessons; deploying the shell
Setup A course Project (shared, "can use")
A course plugin / Skill bundle (provisioned org-wide)
⚠ Governance facts to plan around
Week 1
Building AI Fluency What Is AI?
AI Literacy Verdict — Tribunal performance + a truly-paper test
Capstone t2_curriculum.h1_modules.capstone_slug ai-literacy-portfolio
Day 1 Where Do You Fit?
What this machine is, and what each prompt costs you.
Aim Learn to explain what generative AI is and how it predicts text, getting set up on Slack and Claude, to produce your first working hello-world exchanges.
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name ⬚ author → t2_lessonRail.h1_lessons.deliverable_slug
Stage 2 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to use generative AI to find and shape information, framed by the AI 4D framework, to produce a fast group presentation under pressure.
Format shape audit-skills-against-ai · concept
Shape 4 stages · 8 scenes
Artefact Personal Automation Risk Map
▶ Start Lesson 2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Personal Automation Risk Map
Brief t1_stage_larp.a1_artefacts.brief Written positioning: where the student places themselves on the automation spectrum, with evidence for at least two positions on the map.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Day 2 The Source
Treat the machine as a source to interrogate, not an oracle.
Aim Learn to craft effective interrogating questions and recognise AI limits and hallucinations, pressure-testing a source, to produce a sharp question set.
Format shape generate-interview-questions · concept
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Vetted Question Bank
Brief t1_stage_larp.a1_artefacts.brief A sequenced, fact-check-proof set of 8-12 interview questions that would crack open one real source for one real story.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to delegate tasks to AI through prompt engineering, building a piece end to end, to produce a finished article.
Format shape fact-check-media · drill
Shape 4 stages · 8 scenes
Artefact Source-Criticism Brief
▶ Start Lesson 2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Source-Criticism Brief
Brief t1_stage_larp.a1_artefacts.brief A one-page dossier that takes one circulating claim apart on the table — claim, source chain, red flags, and a verdict you would defend out loud.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Day 3 The Tribunal
Argue AI’s harms through lenses, not hot takes.
Aim Learn to break an argument into its lenses and steer AI for audience, style and tone, analysing a real dilemma, to produce a multi-lens breakdown.
Format shape apply-ethics-lenses · case
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name 6-Lens Case Analysis
Brief t1_stage_larp.a1_artefacts.brief A written analysis of one AI case run through all six Markkula lenses.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to choose a legal doctrine and argue a case within it, while fact-checking AI's claims rather than trusting them blindly, to produce a tribunal verdict you can defend on grounds of law, not sympathy.
Format shape argue-a-position-in-frame · performance
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Tribunal Verdict
Brief t1_stage_larp.a1_artefacts.brief A one-page verdict on a live AI controversy — claim, two evidenced reasons, the strongest counter you could find, and your rebuttal — tight enough to read aloud to a hostile room in three minutes.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Day 4 How We Know
You only know what you can write down without the tool.
Aim Learn to design careful prompts that one-shot strong study artefacts, applying revision strategy and TOK, to produce personal revision materials.
Format shape build-ai-study-tools · concept
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name AI Study Kit
Brief t1_stage_larp.a1_artefacts.brief A one-page doc of three battle-tested prompts that turn any AI chatbot into a flashcard deck, a Socratic quizzer, and a devil's-advocate for one topic you have to revise.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to revise with AI tools ethically and in your own interest, drilling under pressure, to produce a measured score on a no-AI paper exam.
Format shape revise-then-perform-solo · drill
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name ⬚ author → t2_lessonRail.h1_lessons.deliverable_slug
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Week 2
Creating a Trailer Make a Trailer
AI Trailer (~60–90s) screened at Seoul Expo
Capstone t2_curriculum.h1_modules.capstone_slug final-trailer
Day 1 The Pitch
Find the story’s engine; promise it in one line.
Aim Learn to identify the dramatic engine of a story, ideating with AI, to produce an approved trailer concept with a named character, forced change, and real stakes.
Format shape Brief, examples and idea lock
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name ⬚ author → t2_lessonRail.h1_lessons.deliverable_slug
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to crystallize a premise into one promise, ideating loglines with AI, to produce a logline one-pager.
Format shape write-a-trailer-logline · decision
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Logline + 1-Pager
Brief t1_stage_larp.a1_artefacts.brief One sayable-in-a-breath logline plus a tight 1-pager that locks the promise your whole trailer will keep.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to sequence beats that build persuasion and rhythm, drafting with AI, to produce a beat sheet blueprint for the trailer.
Format shape structure-trailer-beats · build
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Trailer Beat Sheet
Brief t1_stage_larp.a1_artefacts.brief A timecoded map of your 60-90s trailer's four beats that the edit, voiceover, and shot list all obey.
Stage 1 • arcadearcade
hook
Every Trailer Is The Same Shape
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Day 2 The Look
Lock a world the camera can’t contradict.
Aim Learn to select the optimal AI model on cost versus capability, benchmarking prompts across models, to produce a model budget plan for your trailer.
Format shape operate-models-in-claude-code · decision
Shape 4 stages · 9 scenes
Artefact Model Comparison Brief
▶ Start Lesson 2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Model Comparison Brief
Brief t1_stage_larp.a1_artefacts.brief Structured comparison of 3 AI models on one task: quality, cost, and recommendation with arithmetic.
2 Moodboard & Style Frame Aim Learn to define and constrain a visual world, generating styled frames with image AI, to produce a style frame moodboard.
Format shape define-a-visual-style · build
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Style Frame
Brief t1_stage_larp.a1_artefacts.brief A one-page reference sheet — palette, lighting, lens, texture, signature move, plus one hero frame — that locks the look of your 60-90 second trailer so every later shot matches it.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to maintain world and character consistency, using AI reference and seed control, to produce a world-and-character visual reference.
Format shape engineer-visual-consistency · drill
Shape 4 stages · 7 scenes
Artefact World + Character Ref Sheet
▶ Start Lesson 2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name World + Character Ref Sheet
Brief t1_stage_larp.a1_artefacts.brief A one-page reference sheet — locked world spec, locked character spec, and 4 matching reference frames — that every later shot in your trailer is generated against.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Day 3 The Footage
Turn the style frames into shots that move.
1 Hero Shots & Cinematography Aim Learn to select hero shots that anchor the narrative, generating keyframes with AI, to produce a cinematic shotlist with hero keyframes.
Format shape generate-key-shots · build
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Shotlist + Keyframes
Brief t1_stage_larp.a1_artefacts.brief A 4-6 row shotlist and the matching AI keyframes that, laid side by side, tell your trailer's whole story in stills.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to convert static frames into fluid motion, animating with image-to-video AI, to produce a motion clip set for your sequence.
Format shape animate-stills-to-video · build
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Motion Clip Set
Brief t1_stage_larp.a1_artefacts.brief Four to six animated 4-8 second shots, each with its source still and the prompt that made it move, ready to cut into your trailer.
Stage 1 • arcadearcade
hook
Will Smith Eating Spaghetti
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Day 4 The Cut
Cut, voice and score it into something that lands.
Aim Learn to cut to musical rhythm and emotional beats, using AI editing assist, to produce a trailer rough cut.
Format shape edit-trailer-to-rhythm · drill
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Trailer Rough Cut
Brief t1_stage_larp.a1_artefacts.brief A 60-90 second beat-mapped rough cut of your trailer where every hard cut, reveal and title lands on the music.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to script voiceover that lands over the cut, synthesizing the voice with AI, to produce a final voiceover track.
Format shape direct-trailer-voiceover · build
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Voiceover Track
Brief t1_stage_larp.a1_artefacts.brief A timed, directed voiceover laid over your trailer cut — every line landing on its shot, in a voice that sells the film in 60 to 90 seconds.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to score picture to emotional intention, generating music and sound with AI, to produce a music and SFX bed.
Format shape score-and-sound-design · build
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Music + SFX Bed
Brief t1_stage_larp.a1_artefacts.brief A timecoded audio plan - one music spine and its key SFX accents - mapped beat-by-beat to your locked trailer so an editor could drop it under picture tomorrow.
Stage 1 • arcadearcade
hook
Sound Is Half The Picture
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Day 5 The Premiere
Finish it, screen it cold, own the room.
Aim Learn to finish and evaluate the cut, polishing with AI upscaling, to produce a final screened trailer for a live audience.
Format shape finish-and-screen · performance
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name ⬚ author → t2_lessonRail.h1_lessons.deliverable_slug
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learners examine AI's real impact on media and creative professionals through a structured panel, drawing on their own trailer production, to produce authentic peer critique and a position they can defend.
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name ⬚ author → t2_lessonRail.h1_lessons.deliverable_slug
Week 3
Coding JARVIS Coding JARVIS
Deployed JARVIS agent (persona + knowledge base + tool use), demoed at Expo
Capstone t2_curriculum.h1_modules.capstone_slug korean-ai-penpal
Day 1 The Build
Make the machine write code you can steer.
Aim Learn to design and operate an AI coding assistant, driving real-time interactions through specification and iteration, to produce a working dev environment.
Format shape set-up-ai-dev-loop · concept
Shape 4 stages · 8 scenes
Artefact Working Dev Environment
▶ Start Lesson 2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Working Dev Environment
Brief t1_stage_larp.a1_artefacts.brief A screenshot-backed setup log proving your editor, terminal, runtime, and AI assistant all work together on one running program.
Stage 1 • arcadearcade
hook
The Real Job Is Debugging
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to specify, generate, and correct code with AI, turning a vibe into executed code, to produce a shipped feature.
Format shape build-with-ai-codegen · build
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Shipped Feature
Brief t1_stage_larp.a1_artefacts.brief One small feature that actually runs in the browser, built by steering an AI codegen tool, with the working prompt and the fix log kept beside it.
Stage 1 • arcadearcade
hook
When Coding Got A New Name
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Day 2 The Brain
Give the build a character and something true to say.
Aim Learn to author and constrain an agent's persona, refining its voice through system prompts, to produce a tuned persona spec.
Format shape design-agent-persona · build
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Persona Spec
Brief t1_stage_larp.a1_artefacts.brief A one-page contract that fixes an AI agent's job, voice, boundaries, and refusal behaviour so tightly that its character survives contact with strangers.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
2 Building a Knowledge Base Aim Learn to ground an agent in curated facts, indexing your knowledge and wiring retrieval, to produce a live knowledge base.
Format shape ground-an-agent · build
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Agent Knowledge Base
Brief t1_stage_larp.a1_artefacts.brief A tight 8-12 entry knowledge base — the exact facts, FAQs, and refusal boundaries your agent quotes from so it stops making things up.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Day 3 The Hands
Give the character hands, and a plan to use them.
Aim Learn to connect external actions to a model, wiring tools via function calling, to produce a tool-using agent.
Format shape wire-agent-tools · build
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Agent with Tools
Brief t1_stage_larp.a1_artefacts.brief A working agent that, handed a task it would otherwise botch, reaches for the right tool with valid arguments and finishes the job.
Aim Learn to sequence dependent agent steps in a plan-act-observe loop, enabling autonomous task execution, to produce a multi-step agent loop.
Format shape orchestrate-agent-steps · build
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Working Agent Loop
Brief t1_stage_larp.a1_artefacts.brief A documented loop that takes one messy real-world goal, drives an AI agent through act-observe-decide turns until a stated stop condition fires, and returns a finished result you would actually use.
Stage 1 • arcadearcade
hook
The Loop That Would Not Stop
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Day 4 The Guardrails
Break your own agent before a stranger does.
Aim Learn to draw and defend refusal boundaries, configuring guardrails under pressure, to produce a refusal policy.
Format shape design-refusal-boundaries · case
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Refusal Policy
Brief t1_stage_larp.a1_artefacts.brief A one-page operating policy that tells an AI agent exactly when to help, when to redirect, and when to refuse — with one worked edge case where you defend the call.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to probe a system for failure, attacking it with adversarial prompts, to produce a hardened agent.
Format shape red-team-an-agent · drill
Shape 4 stages · 9 scenes
Artefact Hardened Agent + Test Log
▶ Start Lesson 2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Hardened Agent + Test Log
Brief t1_stage_larp.a1_artefacts.brief A patched system prompt plus a test log of ten attacks you threw at it, each with the input, the response, and a before-and-after verdict.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Day 5 Ship It
Put it online and run it live for a stranger.
Aim Learn to move an agent to a public endpoint, deploying it as a live web chat widget, to produce a deployed assistant.
Format shape deploy-an-agent · build
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name Live Deployment
Brief t1_stage_larp.a1_artefacts.brief A public HTTPS URL where anyone can chat with your agent, the API key hidden server-side, ready to paste into the group chat.
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to stage a working build into a live moment, polishing the demo for real strangers, to produce a 90-second live presentation.
Format shape ship-and-demo · performance
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name ⬚ author → t2_lessonRail.h1_lessons.deliverable_slug
Stage 1 • arcadearcade
hook
Just a Minute: AI Edition
Stage 2 • dossierdossier
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Stage 3 • workbenchworkbench
Scenes ⬚ author → t2_lessonRail.h3_lessonScenes
Aim Learn to translate hands-on AI experience into a clear-eyed view of your own future work, reflecting after shipping a real product, to produce an articulated personal stance on a changed professional horizon.
2 Skill · quantified
Skill — ⬚ author → t2_lessonRail.h1_lessons.skill_slug
Measured by — ⬚ author → —
Data-type family — ⬚ author → t2_lessonRail.h1_lessons.nativeDataShape
Rubric — ⬚ author → —
3 AI × skill → tasks
AI feature t1_stage_workbench.a1_workbench_index.aim ⬚ author → t1_stageAi.h1_aiCapabilities
Handoff — ⬚ author → t2_lessonRail.h2_lessonStages.ai_handoff
4 Artefact → capstone
Deliverable t1_stage_larp.a1_artefacts.name ⬚ author → t2_lessonRail.h1_lessons.deliverable_slug
Stage 1 • arcadearcade
~45m
Spotlight: Interview Quality