Mostly studio time. Your job: circulate, force scope cuts, run the clinic.
60-second recap
The deal
AI for friction, humans for meaning. Raw output is never finished work. Verify before trusting. Disclose where it matters. Skepticism welcome when specific.
Standing recap. Last full recap before Demo Day.
Synthesis
Thirteen weeks of techniques
are worth nothing until one of them lives in your practice.
Today you build that one thing.
This is where the term pays off.
The week in one line
Ship one true thing that gives you leverage without taking your voice.
Both halves matter. Leverage AND voice.
The trap I'll be policing
Automating your ART instead of your FRICTION
A capstone that makes your finished work is a worse portfolio piece than one that buys you two hours a week.
Say it now, repeat it at every desk. The #1 failure mode.
Print it · keep it up all day
The Capstone Quality Bar
Half-sheet for every student. Self-check at the end.
All must be yes
Four questions
Real friction — an actual recurring problem in YOUR practice
Documented human decisions — selection, rejection, judgment, visible
Honest disclosure — what AI did, what you did
Demo-able to an employer — every choice defensible
"Small-and-passes-all-four beats ambitious-and-half-built. Every time."
Deliverable
You will leave with
A capstone draft — a working(ish) thing A 1-page process explanation — proves it's yours
The draft is the bonus. The process explanation is the assessed deliverable.
Demo
Three capstone shapes
~0:10. Three worked examples from DEMOS Week 14.
Shape 1
Workflow / process system
The Week 5 chain, assembled
brief → boards → shot list → AE checklist, reusable.
Friction → AI role → human role → time saved.
Best for motion, animation, multi-stage pipelines.
Shape 2
Prompt system / kit
The Week 8 five prompts, tuned
A reusable kit you run per project.
Best for UX, design, illustration, writing-heavy practices.
Shape 3
Tiny coded tool
The Week 11 generator + Week 12 changelog
Single-file, inspectable, yours.
Best for "I keep retyping/recalculating this" problems.
For each
The question
What would make it FAIL the bar?
No real friction · no documented decisions · no disclosure · couldn't demo it
Collect answers — the room internalizes the bar by hunting failures.
Live
Scope a volunteer's idea
Act as a capstone advisor. My idea: [X]. Interview me briefly, then output a one-week build plan: scope cut to ~6 hours, AI-role/human-role split, riskiest assumption first, and what "done" looks like.
When the model over/under-scopes, correct it out loud. "It's too agreeable to tell you you're over-scoped. I'm not."
Hour 1
Scope lock
~0:35. Every student leaves with three things written.
Write down three things
Friction · Shape · Riskiest assumption
The friction — the real recurring problem, with frequency The shape — which lane The riskiest assumption — what sinks it if it fails
This is where the day is won or lost. Pressure-test scope at every desk.
The shortcut
Pick the shape closest to a draft you already have