TL;DR
A persona that survived the audition loop isn't a chat trick anymore — it's a digital asset, and in Uzu you can publish it to the Trending marketplace, where other users spend DAG-Credits to run it and you earn a 50% revenue share on every single use. This is the final exam no prompt-engineering course can simulate: real users, real credits, real feedback — the market grading your craft.
The Analogy
At some point, every film director faces the same threshold moment.
For two years, the film has existed in the safest environment imaginable: private screenings, trusted colleagues, controlled feedback. Every note can be implemented, every scene re-cut, every ending re-shot. It's comfortable — and it isn't a film yet. A film becomes a film the night it reaches strangers who paid for tickets, because only they will tell you the truth: the boring scene your friends were too kind to mention, the joke that only the crew laughed at. The director's craft is complete at the final cut; the director's knowledge begins at the box office.
And here's the part creators get backwards: releasing the film isn't abandoning it. The box office isn't applause or rejection — it's data. Which audiences came. Which scenes made them stay. What they told their friends. The great directors of any era are the ones who shipped, watched the receipts, and made the next film smarter — not the ones still re-cutting the first one in a private room, perfecting a movie nobody has seen.
Your persona is at that threshold. Through the casting process you built it in the safest environment imaginable: your own chats, your own hostile scenes, your own eyes. It survived your auditions — which is the prerequisite, not the destination. Keeping it private is the perpetual private screening: comfortable, unfalsifiable, and quietly worthless to everyone but you.
Uzu's Trending marketplace is the theatrical release. You publish; strangers decide; the credits they spend are the ticket sales. And like the box office, it pays: every time another user runs your engineered persona, you earn a share of what they spend. The market doesn't grade your intentions or your cleverness. It grades the asset.
How It Works & The Uzu Revenue Share
Preparing for Launch
Before you publish, the asset needs three things the audition loop didn't require:
| Launch requirement | Why strangers need it | Where it came from |
|---|---|---|
| A name that says the job | In Trending, the name is the trailer — 3 seconds to communicate the value | Your persona's role, sharpened |
| A pitch that promises one outcome | Buyers skim descriptions; they commit to promises | Your constraints, translated into a benefit |
| A V3 that survives strangers | Your hostile scenes ≠ their hostile scenes; ship final cuts only | The audition loop |
That second row is where most launches die, so take it seriously: your description is a one-shot prompt aimed at a human — verb, value, format. "The Editor: paste any draft, get a PUBLISH / REVISE / REJECT verdict with three numbered findings and zero flattery." A stranger reads that and already knows the experience. Compare it to "Marcus, a witty helper for your writing needs!" — a costume with no job.
The 50% Revenue Share
Now the economics of the theatrical release. In Uzu, personas run on DAG-Credits — the premium currency, transmutable at 1 DAG ≈ 10 G, because a persona holds a standing rulebook on the model's context window every single turn. When you publish to Trending:
- Other users browse Trending and choose your persona. They're spending their premium credits on your engineering — the constraints you wrote, the refusals you hardened, the tone that never drifts.
- Every use pays you: a 50% revenue share of what each user spends running it. Half of every ticket goes to the director.
- The loop compounds. Earnings arrive in G-Credits — the everyday currency you can spend on your own drafting, audits, and chains — which means a good persona funds the tools that build your next persona.
Read that third point again, because it's the real mechanic: the revenue share turns your skill into a budget. The discipline you learned across this whole course — precise one-shots, tight scopes, few-shot examples only when required — keeps your costs lean; a published persona keeps your balance growing. Skill compounds in both directions.
And so does the feedback. Trending is a live audit that never stops: usage is approval, silence is a finding, and a one-star experience is a hostile scene you forgot to write. Ship, watch the receipts, revise the asset, ship again — the director's loop, forever.
Before & After (The Launch)
The selfish launch:
Name: "Helper V2" Description: "A useful assistant for many tasks. Try it!" The persona: a V1 with vague constraints, never auditioned against strangers.
Nobody browsing Trending can tell what this does, for whom, or why it costs premium credits. Even if a curious user tries it once, the vague constraints drift by message ten — and a persona that breaks its promise in front of a paying stranger doesn't get a second chance. Zero shares, zero data, zero growth.
The marketplace launch:
Name: "The Commissioning Editor" Description: "Paste any draft — blog post, memo, chapter. Get a PUBLISH / REVISE / REJECT verdict in 30 seconds: three numbered findings, each quoting your weakest line and what it fails to do. Blunt by design. No rewrites, no flattery." The persona: the V3 that survived off-topic bait, author arguments, and tone-drift marathons.
Same marketplace, opposite fate. The name states the job; the description is a demo the buyer can feel before purchasing; and the persona behind the pitch actually delivers the promise — because every phrase in that description is a constraint that was auditioned. Strangers convert, come back, and tell others. That's the difference between publishing a costume and publishing an asset.
Common Pitfalls
- Terrible names and descriptions. "Helper V2" and "try it!" are box-office poison: strangers skim in seconds, and a name that hides the job guarantees they never meet the persona. Write the name as the job title and the description as a 30-second demo — the exact experience, the exact output, the exact promise.
- Shipping a V1 to learn "what people want." The market is a grader, not a focus group. An unhardened persona breaking in front of paying strangers doesn't gather feedback — it burns first impressions, and Trending has a long memory. Audition first; the market grades final cuts only.
- Pricing by ego instead of value. The question isn't "how clever is my persona?" but "does it save the buyer more than it costs them?" A persona that delivers a 30-second editorial verdict earns its DAG price daily; a persona that delivers vibes never does. Price the outcome, not the effort.
- Launching and vanishing. Publishing is the start of the loop, not the end: watch which promises attract users, which scenes still break in the wild, and revise the asset. The 50% share rewards maintained assets; abandoned ones decay quietly down Trending.
- Keeping the asset private forever. The comfortable trap: re-cutting the film in the private room. An unpublished persona pays exactly zero, teaches exactly zero, and improves exactly zero — no matter how brilliant. The threshold exists to be crossed.
FAQ
Can I make money creating AI prompts?
Yes — if you build assets, not strings of text. In Uzu, a hardened persona published to the Trending marketplace earns a 50% revenue share every time another user spends DAG-Credits to run it. The bar is real: the persona must survive strangers, not just its author — precise constraints, honest boundaries, a pitch that promises one outcome and delivers it. Prompt engineering becomes income when the market can verify it.
How does the Uzu Trending marketplace work?
Trending is where published personas meet the public. Creators list hardened personas with a name, a pitch, and a price; users browse, spend DAG-Credits to run them, and the creator receives a 50% revenue share of every use, paid in G-Credits. Usage drives visibility, silence buries weak assets — it's a live, continuous audit of your engineering, with receipts.
What makes a persona worth publishing?
One test: does it deliver the same promise to a stranger as it did to you? That means a V3 that survived hostile auditions — off-topic bait, refusals, arguments, drift — plus a name that states the job and a description that demonstrates the outcome. If a stranger can read your pitch and predict the experience, and the persona keeps that prediction under pressure, it's an asset. Otherwise it's still a draft.
The End
Twelve lessons ago, you met an over-qualified intern who knew nothing about you. Since then, you've learned to hand that intern complete work orders — to aim its attention with words and images, to keep its desk clean, to rent its doubt, to chain its stations, to show it your patterns, and finally to build characters that hold under pressure. You are no longer someone who talks to AI. You engineer behavior. That skill has exactly one graduation ceremony: the moment something you built works for a stranger.
Somewhere in your chats right now is a persona that survived your auditions — or one five minutes from existing. Name it. Pitch it. Publish it. Let strangers spend their credits on your craft, collect your half, read the market like a box office, and build the next one smarter.
The course is over. The director's chair is yours — and the theater is open.
Final drill — the only mandatory one: open Uzu, publish one persona to Trending today. Not next week. Today. Your first listing won't be your best work; every director's first release wasn't either. But zero releases teach you nothing, and the market is already seating the audience.