Cheat sheet · keep this open

Frameworks & shorthand

Everything the cards reference, spelled out. Spine of the whole talk: Idea → Create → Grow — “I help founders build products and grow audiences with AI.”

1 · Walk & Talk
the intake — pull a real product out of your own head

Talk the idea out loud (literally on a walk). The rule: if you can't verbalize it, you can't build it. This is where a messy thought becomes something you can actually define.

2 · Dump → Define → Build
the core loop — the step everyone skips is the whole game
Dumpbrain-dump it all→ Defineturn it into a PRD/spec→ BuildAI ships it, you babysit

“I stopped trying to code, and learned to define.” The Define step on a real idea: 1. Dump → 2. Problem → 3. User → 4. Features (and the NOT-features). That is a PRD → prd.midcurved.com. If you can define it, you can build it.

3 · Owned → Earned → Paid
the growth ladder — how you get your first users
Ownedyour content + list→ Earnedothers share it for you→ Paidads — only after 1 & 2 work

In 2026 building is the easy part — growing is the game. Don't pay for reach until owned + earned are working.

Shorthand key (what the labels mean)

A–EThe 5 question blocks (themes): A make money · B cost · C tools · D real engineering · E the leap.
#1–#10The individual panel question cards (see Panel tab). #1 opens the whole panel.
O→E→POwned → Earned → Paid (framework 3).
D→D→BDump → Define → Build (framework 2).
[LUCA FILLS]A spot in the warmup where Luca's real story / numbers go — get these beforehand.

Who's who — colour = throw to first

RikHost / moderator. The “define, don't code” thesis. O→E→P, D→D→B.
ChelseaEducator (CoLab Lisbon). Non-technical founders, learning, craft.
SethProduct / marketing (Squidgy.ai). Distribution, finding users, agents.
DanielDeep engineer (ex-Google / ex-Alexa). Your skeptic anchor on hard cards.

Closing CTA: one path only — Banskonnect community QR or the Define-to-Ship Kit opt-in. Soft-mention the Ship Sprint, don't hard-sell.

🎙 GMP cold open · 5–10 min

Luca: Idea → Revenue in 10 min

You interview Luca, podcast-style. Prove the premise is real, then bring up the panel.

The spine Luca tells
Spotted a real gap → moved fast → vibecoded to revenue → now it runs itself.
[LUCA FILLS] = get from Luca beforehand.
0:00~30s · Rik hook

“Before I bring the panel up — quick one. Luca found a thing, built it himself with AI, and now it makes money while he does basically nothing. Ten minutes, idea to revenue. Clock's running — go.”

0:30~2 min · The opportunity
How did you spot it?
  • “What was the gap? How did you spot it?” → [LUCA: the pain / niche]
  • Key beat: how he knew it was real — someone already paying, obvious pain, a thing he needed.
push“What made it a real opportunity and not just a cool idea?”
2:30~2 min · Moving fast
Idea → first version: how fast, what did you skip?
  • → [LUCA: timeline] · speed = ruthless scope → ties to Define.
push“What did you deliberately NOT build?”
4:30~2 min · Vibecode → revenue
When did it stop being a project and start being a business?
  • build → first paying user → [LUCA: stack + first-dollar moment]
push“What was the first dollar? How did money start flowing?”
6:30~1.5 min · Punchline
Today — how much do you actually touch it?
  • → [LUCA: current state + revenue shape]
Let it land: “…so it makes money while you do basically nothing.”
8:00~1 min · Rik handoff
One lesson → tee up the panel
  • “One thing everyone should take from this?” → [LUCA: one line]

“That's the whole game — find something real, define it, ship it fast, let it run. Everything my panel is about to argue over is HOW. Let's bring them up.”

Get from Luca beforehand

  1. The opportunity + one sentence on why it was real (not hype).
  2. Timeline — idea → first version → first paying user.
  3. Revenue shape he'll say on stage (MRR / range / “it pays”).
  4. The one lesson, pre-thought so it's crisp.
  • Real numbers > big numbers. “$400/mo, I touch it twice a month” beats vague hype.
  • If MRR is light, frame on speed + real opportunity.
  • Don't run past 10 min — it's the appetizer.
Moderated by Rik

Panel question cards

2–3 cards per block, not all 10 in order. Read it, name the asker, throw to the OPEN WITH person, use the push if they agree too fast.

★ Opener — read first

1“Define, don't code” — is that really the whole game?
openDanielthen Chelsea → Rik defends
push“Where's the line? Name the moment defining stops being enough.”
Why: engineer pushes back first → sparks the night.

Block A — Will it make money? cards 2–4 · lead Chelsea + Rik

2How do you know it'll make money — before you build it?
openSeththen Rik (O→E→P)
push“Pre-sell vs build-then-see — who's charged before they built?”
Asked by: Alina.
3Should you even talk to users first?
openChelseathen Daniel/Seth disagree
push“AI made building cheap — did that kill the customer interview or make it matter more?”
Asked by: Juliette.
4In a sea of vibed apps, how do you stand out?
openRikthen Seth · Chelsea
push“Is the moat the product, the audience, or the founder? Pick one.”
Asked by: Arthur.

Block B — Building without burning cash cost card · lead Daniel + Rik

BYour single best cost / token-saving habit?
openDanielthen Rik (sub > API, KISS)
push“Where do beginners waste the most money without realizing?”
Asked by: Manuel, Oscar, Andrew.

Block C — What do I actually use? cards 6–7 · lead Rik + Daniel

6Does the tool matter? Claude Code vs Codex vs the rest.
openRikthen Daniel · Seth
push“One tool for a total beginner tomorrow — no hedging.”
Asked by: Maria, Torben (Ralph), Tony.
7How much should you plan before you build?
openChelseathen Rik (D→D→B) · Daniel
push“Can you spec a vibe-coded app to death? What's the cost of over-planning?”
Asked by: Tony, André.

Block D — Is this real engineering? cards 5 + 8 · Daniel carries it

5Where does vibecoding actually break?
openDanieleveryone: one war story
push“Is the ceiling getting lower every month, or are we just dodging it better?”
The “keep it real” card — earns trust.
8Can you trust a vibe-coded app with real user data?
openDanielthen Seth (risk vs paralysis)
push“The one security thing you'd never let AI decide?”
Asked by: Ligia (OWASP), Luciano, Anastasia, Luc.

Block E — Making the leap cards 9–10 · lead Rik · honest closer

9Building got cheap — so where do your first users come from?
openSeththen Rik (O→E→P) · Chelsea
push“Build-in-public vs launch-quiet-and-sell — which actually worked?”
Asked by: Emily.
10The leap: quit the job, or keep the safety net?
openRikthen Chelsea + Seth
push“Is solo AI-building a real living yet — or are we selling the dream?”
Honest closer. Asked by Arthur, Alina, Alexandra.

Backup cards

Users who don't know what they want.
openChelseathe Mom Test · Andrew
Best AI news source — one follow, one habit.
lightningall fourAlexandra
Always connect projects to GitHub?
openDanielthen Rik · Oscar

Moderator reminders

  • Throw to ONE named person. Never “anyone?” — silence kills panels.
  • Name the asker — they're in the room.
  • If they all agree, use the push line. Disagreement is the show.
  • Daniel = skeptic anchor — open hard/technical cards with him.