Package the capstone, clarify disclosure, prep a 3-minute demo.
All tracks — 4 hours · Demo Day & close
Demo Day, Future-Proofing, and the Artist's AI Operating System
Present what you built, rewrite your boundaries now that you actually know things, and leave with a maintainable personal system for using AI — and staying current — without drowning or losing your voice.
Prepared demo artifact
Instructor backup / finished example: ai-os.html. Use it if the live tool run fails, or to reveal the finished structure after critique.
01This Week at a Glance
You will make: your final capstone (presented) + AI Use Manifesto v2 + a personal tool-tracker, assembled into your AI Operating System one-pager.
Present; feedback on clarity, value, and artistic integrity.
The evaluation pattern, the instructor's real system, your one-pager + Manifesto v2.
The week in one line
Systems outlive tools — leave with a system, not a tool list. Sora's consumer app was retired in April 2026. If your AI skill is "I know how to use [tool]," you're already expiring. If it's a system, you never fall behind — because the system absorbs whatever's new.
02Part 1: Demo Day
Bring the finished(ish) capstone + 1-page process explanation from Week 14. The 3-minute presentation:
- The friction (30s) — the real recurring problem, with its weekly cost.
- The thing (60s) — show it working; the one decision you're proudest of.
- The split (45s) — what AI did, what you did, the honest disclosure.
- The verdict (15s) — does it pass the quality bar? What's next?
- Questions (30s) — the room probes usefulness and integrity.
What you're being seen for
Not polish — judgment. A small honest tool you can defend beats an ambitious thing you can't explain. The presentation is itself a portfolio rehearsal (Week 13 skills, live).
03Part 2: Future-Proofing (the evaluation system)
New AI products arrive constantly. The skill is evaluating them fast and calmly instead of fear-adopting or fear-refusing. Run any new tool through these six questions in five minutes:
| Question | Why it matters |
|---|---|
| What's the friction it actually removes? | If you can't name it, you don't need the tool |
| What's the training-data / rights situation? | The Week 4 question — values decision, not default |
| Control, continuity, cost? | The Week 5 video pattern, generalized |
| What does it touch / where does my data go? | The Week 11–12 inspection habit |
| Does it want my friction or my art? | The Week 14 line |
| What would I disclose if I used it? | If you can't write the disclosure, reconsider |
The hype filter
Every tool is announced as revolutionary. Wait two weeks; read a skeptic; try it on one real task; check what it actually removed. Most tools are a feature, not a future — the few that matter will still matter in a month.
04Part 3: The Artist's AI Operating System
Your AI OS is a one-pager — honest, not aspirational — with six sections:
- Tools I actually pay for — and why. (And what I cancelled — the cancellations are the credibility.)
- My folder / naming system — for prompts, process screenshots, iteration archives.
- My 5 most-used prompts — the real ones (the critique loop, the brief skeleton, the extraction interview).
- My automations — the workflows/tools I trust, and the checkpoint each must pass.
- My refusals — what I keep fully human, stated plainly.
- My review rhythm — a monthly tool check, one trusted source, what I ignore.
This is the document that makes you durable. Maintain these six sections and you have a working relationship with AI that survives any individual tool dying.
The tool-tracker (part of section 6)
| Tool | What I use it for | Tier/cost | Last checked | Keep / watch / drop |
|---|---|---|---|---|
Five minutes a month beats five hours of panic when something changes.
05Part 4: Manifesto v2 (close the loop)
Find your Week 1 Manifesto v1. Read it. You knew less then. Now rewrite it — 8–12 bullets — informed by thirteen weeks of actually doing the work:
- What boundary held? What boundary moved, and why?
- What did you learn you'd refuse that you were neutral on in Week 1?
- What did you learn you'd embrace that you were afraid of?
- Where does your authorship actually live, now that you've tested it?
The measure of the whole course
The v1→v2 distance is what this course measured. Keep both versions. In a year, write v3 — the habit of revisiting your boundaries is the most durable thing you'll take from here.
06This Week's Words
AI Operating SystemYour six-section system: tools, folders, prompts, automations, refusals, rhythm.
Evaluation patternThe six questions you run any new tool through in five minutes.
Hype filterWait, read a skeptic, try on one real task, check what it removed.
Tool-trackerThe monthly five-minute table: what you use, cost, keep/watch/drop.
Review rhythmYour sustainable cadence for staying current without doom-scrolling.
RefusalsWhat you keep fully human, stated plainly — the matured manifesto.
Manifesto v2The Week-1 manifesto rewritten with thirteen weeks of evidence.
Future-proofingBuilding skill in the system, not the tool, so tool-death doesn't set you back.
Slop resistanceThe habit of staying distinctive as generic floods every feed.
Durable skillJudgment, taste, accountability — what no tool update obsoletes.
07Prompt Recipes
The OS assemblerHelp me turn these rough notes about how I actually use AI into a clean one-page "AI Operating System": [paste your notes on tools, prompts, automations, refusals, rhythm]. Six sections: tools I pay for, folder system, top 5 prompts, automations + checkpoints, refusals, review rhythm. Keep it honest, not aspirational — flag anything that sounds like a system I wish I had rather than one I actually use.
The new-tool evaluatorA new AI tool just launched: [name / what it claims]. Run it through my evaluation pattern: what friction does it actually remove, what's the training-data/rights situation, what about control/continuity/cost, where does my data go, does it want my friction or my art, and what would I disclose if I used it? End with a keep / watch / drop call and your honest reasoning — don't sell it to me.
The Manifesto v2 reflectorHere is my AI Use Manifesto v1 from 13 weeks ago: [paste]. Interview me about what's changed: which boundaries held, which moved and why, what I'd now refuse that I was neutral on, what I'd now embrace that I feared. Don't rewrite it for me — ask the questions that help me write v2 myself.
The hype debunkerThis AI announcement claims to be revolutionary: [paste]. Strip the hype: what can it actually do today (not "will soon"), what are the named limitations, who's the skeptical take, and what's the smallest real task I could test it on this week? Be the calm friend, not the launch livestream.
08Lab: AI OS Checklist + Manifesto v2
Demo Day runs through the middle of the session; the OS/manifesto work bookends it. The deliverables:
- AI OS one-pager — the six sections, honest, findable.
- Manifesto v2 — v1 rewritten with evidence; keep both versions.
- Tool-tracker — the monthly table, started with your current tools.
The AI OS Checklist (your staying-current decision)
- My monthly 5-minute tool-check date.
- My ONE trusted source for AI news (not the firehose).
- What I deliberately ignore.
- My re-read-the-manifesto cadence.
- My "try one real task before adopting" rule — confirmed.
Zero-AI track (fully equal)
The OS, manifesto, and tracker are reflective writing — they need no AI at all, and the most honest versions are often written by hand.
09Integrity Check (the last one)
This week's four
- Is your AI OS honest — the tools you actually use and refuse, not the impressive version?
- Does your Manifesto v2 reflect what you genuinely learned, including where you changed your mind?
- Can you still name, plainly, where your authorship lives — and defend it?
- Is your system built to keep you distinctive as the feeds fill with slop — or just efficient?
10Resources
- NewsletterLinkedOne Useful Thing — Ethan MollickThe model "one trusted source" — hype-free, practical, sustainable. A good default for your review rhythm.
- Course docLinkedStudent AI Integrity Pact (this course)The disclosure language and boundary framing for your OS — the course's spine, to keep.
- Course docLinkedTools & Access Notes (this course)The "never teach one tool as permanent; teach the decision pattern" doc — the model for your own tool-tracker.
- ReferenceLinkedThe EU AI Act explorerDisclosure and synthetic-media rules are becoming law — your OS's disclosure habits are future-proofing in the legal sense too.
- RoundupVerify before teachingA current, credible "state of AI" roundupFor the future-proofing framing; these date fast — pick the current best as a model of a good source, not gospel.
- Course docLinkedGlossary (this course)Keep it — when a new tool uses a word you half-know, this is your reference. The vocabulary outlives the tools.
The point of Week 15
This entire course's tool-specific content has a shelf life. The systems, habits, ethics, and judgment do not. That asymmetry is the whole point.
After the course
- Schedule your first monthly tool-check now — actually put it in your calendar.
- Share your AI OS one-pager with one classmate; steal one section idea from theirs.
- In a year: write Manifesto v3. The habit of revisiting your boundaries is the most durable thing you'll take from here.
You're done
You came in possibly skeptical, anxious, or overwhelmed. You leave with a working vocabulary, a briefing skill, an iteration discipline, discipline-specific workflows, a tool or system you built, career materials in your own voice, a capstone — and, most durably, a system and a set of boundaries you can defend. You don't have to love AI. You understand the machine your industry is using, and you've decided, on your own terms, how you'll work with it. Keep your manifesto where you can find it.
← Back to the full 15-week workshop