Class 20, hybrid format

The exercise · 6:32 PM onward

Mining the capstone brief live

The brief

Run the Coca-Cola holiday brief through all four stages.

The night's material was the real final assignment: invent the next holiday shorthand for Coca-Cola . Each stage ran thirty minutes: individual work in the GPT first, then group elevation on Miro.

Nothing was due that night. "You're building to your final assignment."

Every group converged on the same ideas

Red thread, exhale, pause: the share-outs kept producing the same concepts in different wrappers. The cause got named in the room: same source material, same model, same guardrails. It is the course's multi-order lesson demonstrated live, and the reason the next class's winning concept came from human forced connections instead.

The archive gap

One wish from the room: a collaborator with a creative archive, so it can say "someone already did this." Angella's honest answer: no system for it exists yet, because AI presents everything as if it has never been done. The current workaround, credited to Sydney's class: upload a board of your own work as knowledge and make the tool study it before continuing.

The fight for something physical

Every generated concept was a feeling. The push that mattered was for something tangible, "recognizable, that someone will still remember what it looked like five minutes later." The gap between a feeling and an object turned out to be the real creative problem of the night.

Methods and prompts

Five methods to take with you

METHOD 01 · STAGE 1 part of: structure your ask

X-ray the brief

Interrogate any brief for the tension it avoids stating, before a single idea gets generated. The output format does half the work: insight, enemy, role of the brand, why it matters.

Where it came fromThis was stage one of Angella Tape's Creative Collaborator GPT, taught on July 29 when the class ran the capstone brief through it live, asking what tension the brief is too polite to state and producing five to seven spark ideas in a table.Use it whenUse it at the very start of working a brief, before generating any ideas, to surface the tension the brief itself is avoiding.

Same idea elsewhereGive it two keyframes instead of one description (class 2) · Speak the language of the internet (class 3) · Write it as a GOLD block (class 4)

Working prompt

Here is my brief: [paste]. X-ray it: what is the tension this brief is too polite to say? Give me 5 to 7 spark ideas as a table, each with the insight, the enemy, the role of the brand, and why it matters.

X-ray a brief of your own →

You will know it worked whenthe table lists five to seven ideas, each with its own insight, enemy, and role for the brand, not one idea repeated with different words.

METHOD 02 · STAGE 2 part of: go wide then narrow

The 2 AM confession

Escape the category by forcing the brand and the idea through frames the category never uses. The confession question breaks the polite surface fastest.

Where it came fromThis was stage two of Angella's process on July 29, escaping the category by asking what the brand would confess at 2 AM and then forcing the strongest idea through five unrelated lenses like art, architecture, and choreography.Use it whenUse it once a brief feels stuck in familiar category territory and an idea needs to be pushed through unrelated frames to break loose.

Same idea elsewhereThe lateral reframe (class 2) · Start a fresh chat instead of one long thread (class 3) · Force the creative partner deeper (class 4)

Working prompt

If this brand was a person at 2 AM, what would it confess? Then take my strongest idea and force it through five unrelated lenses: art, architecture, music, choreography, medical devices. Show me what survives the trip.

You will know it worked whenwhat survives the five lenses looks genuinely changed by at least one of them, not the same idea restated in new dress.

METHOD 03 · STAGE 3 part of: reject the first draft part of: keep the judgment human

Spar with your rebel strategist

Score every idea against a ladder whose top three rungs are contagious, cultural phenomenon, legendary. The kill threshold is the point: most ideas do not survive it, and that is the tool working.

Where it came fromThis was stage three of the July 29 session, scoring every idea on a one-to-ten creative ladder that tops out at contagious, cultural phenomenon, and legendary, with anything below eight killed outright.Use it whenUse it once a batch of ideas exists and needs a hard cut, keeping only what clears a genuinely high bar.

Same idea elsewhereJudge AI output with your own trained eye (class 1) · The sea-of-sameness check (class 2) · Pick the relationship on purpose (class 2)

Working prompt

Act as my rebel strategist. Score each idea above on a 1 to 10 creative ladder where 8 is contagious, 9 is a cultural phenomenon, and 10 is legendary. Kill everything below 8 and tell me exactly why the survivors deserve to live.

You will know it worked whenmost of the ideas actually get killed for scoring below eight, and the survivors get a specific reason, not just a passing grade.

METHOD 04 · THE ARCHIVE GAP part of: context beats prompts

Make it study your archive

AI presents every idea as if it has never been done. The current fix: upload your own body of work as knowledge and make the tool study it before ideating, so "someone already did this" becomes possible.

Where it came fromAngella raised this on July 29 in response to a classmate's wish for an AI that could recognize when an idea had already been done, explaining that no such system exists yet and that the workaround, credited to an earlier class on prompt libraries, is to upload your own past work as knowledge before ideating.Use it whenUse it before a new ideation session, when you want a tool to check new ideas against your own existing body of work rather than presenting everything as untried.

Same idea elsewhereControl the input and build gradually (class 2) · Show it your voice and not just the output (class 3) · Evolve the reference instead of restarting (class 5)

Working prompt

Before we ideate: here is a collection of my past work [upload]. Study it. Tell me the patterns you see. When we generate ideas, flag anything that repeats what I or this category have already done.

You will know it worked whenthe new ideas get flagged specifically when they repeat something in your uploaded work or something the category has already done.

METHOD 05 · THE BUILD RULES part of: build tools not chats

Stage-gate your own GPT

Angella's design constraints, portable to any assistant you build: capped questions, approval gates, pre-formatted outputs, reference material uploaded as knowledge. The user owns the pace.

Where it came fromAngella described these as her own design constraints for the Creative Collaborator GPT on July 29: capped questions per section, an approval gate before advancing stages, pre-formatted table outputs, and uploaded reference material, all meant to keep the user in control of the pace.Use it whenUse it when building your own custom assistant, to set the same pacing and approval rules so the tool cannot rush ahead of your thinking.

Same idea elsewhereStand up a reusable Project or Gem (class 4) · Weight your references on purpose (class 5) · Let one model write the prompt for another (class 6)

Working prompt

You work in stages. Ask me at most two questions per stage. Deliver every output as a table I can copy. Never advance to the next stage until I approve. I own the pace of this conversation.

You will know it worked whenit asks no more than two questions per stage and stops for your approval before moving to the next one.