Every stage has the prompt pre-filled for all three. Hit Copy and paste into your chatbot. Comparing how the same loop plays out across very different ideas is half the lesson.
Thread A · ESCAPED-BUG FINDER — an AI agent that finds the bugs that escaped to production and writes the prompt to fix each. Buyer: on-call engineers paged by bugs their tests missed.
Thread B · NEW DATING SERVICE — one curated intro a week, no feed, no swiping. Buyer: the chronically single, burned out on the swipe treadmill.
Thread C · 1-FOOT ROBOTS — a 1-foot-tall desk humanoid that runs AI skills you can watch. Buyer: makers and AI fans who buy gadgets to feel the future.
Your own: each stage also has a bracketed template — fill in your idea and run the exact same prompt.
Generate cheap, stress-test brutally, keep 2 ideas that survive an obituary. (In the room you swap with a pod-mate before attacking — your own ideas you'll flinch on.)
You are a serial founder who has killed more ideas than you've shipped. Attack — do not flatter. Domain: AI software testing / developer tools Unfair advantage: 20 years finding bugs at Google + Microsoft Hair-on-fire problem: teams ship fast with AI and bugs escape to production — on-call engineers find out from angry users 1. Generate 20 angles around "find the bugs that escaped to prod + write the AI fix." Be specific (wedge, buyer, exact offering). 2. Score 1–10 on wedge sharpness, buyer urgency, AI defensibility, distribution. 3. Kill 18. Defend 2. 4. For the 2, write the OBITUARY — what kills it in 18 months? Refuse to say "interesting" or "promising." Use scores.
You are a serial founder who has killed more ideas than you've shipped. Attack — do not flatter. Domain: dating / relationships Unfair advantage: strong product taste + one tight local community to seed Hair-on-fire problem: the chronically single are burned out on swiping and dating-app fatigue 1. Generate 20 angles around "a dating service that sends one curated intro a week — no feed, no swiping." Be specific (wedge, buyer, exact offering). 2. Score 1–10 on wedge sharpness, buyer urgency, defensibility, distribution. 3. Kill 18. Defend 2. 4. For the 2, write the OBITUARY — what kills it in 18 months? Refuse to say "interesting" or "promising." Use scores.
You are a serial founder who has killed more ideas than you've shipped. Attack — do not flatter. Domain: consumer AI hardware Unfair advantage: maker + AI background; I can ship a small device Hair-on-fire opportunity: people want the magic of a humanoid robot, but full-size is $20k and years away — a tiny one is cute, cheap, shippable now 1. Generate 20 angles around "a 1-foot-tall desk humanoid that runs AI skills you can watch." Be specific (wedge, buyer, exact offering). 2. Score 1–10 on wedge sharpness, buyer urgency, defensibility, distribution. 3. Kill 18. Defend 2. 4. For the 2, write the OBITUARY — what kills it in 18 months? Refuse to say "interesting" or "promising." Use scores.
You are a serial founder who has killed more ideas than you've shipped. Attack — do not flatter. Domain: [your domain] Unfair advantage: [skill / access you have] Hair-on-fire problem or opportunity: [who is bleeding, and why now] 1. Generate 20 angles. Be specific (wedge, buyer, exact offering). 2. Score 1–10 on wedge sharpness, buyer urgency, AI defensibility, distribution. 3. Kill 18. Defend 2. 4. For the 2, write the OBITUARY — what kills it in 18 months? Refuse to say "interesting" or "promising." Use scores.
Ask AI for a single HTML page that animates your wow moment, then iterate on the look & feel. HTML is how you iterate with AI — not Docs or PPT. Out of ideas? Say "make it better."
Build a SINGLE HTML page that ANIMATES the wow: I point it at an app, the bugs that escaped to production stream in with severity, and each shows a one-click "AI fix" prompt. One file, real-ish data, no auth. Then I'll iterate on look & feel; if I stall I'll say "make it better." BONUS — a fun logo: a bug caught in a net (or a shield with a bug), flat vector, bold simple shapes, 3 options, transparent-background PNG.
Build a SINGLE HTML page that ANIMATES the wow: I answer 5 quick honest questions, then ONE curated intro card flips over with a short "why you two" reason. No feed, no swiping. One file, no auth. Then I'll iterate on look & feel; if I stall I'll say "make it better." BONUS — a fun logo: two interlocking speech bubbles (or a single spark), flat vector, 3 options, transparent-background PNG.
Build a SINGLE HTML page that ANIMATES the wow: a 1-foot desk robot turns to face me, reacts to a typed command ("celebrate", "remind me"), and does a little animation. One file, no auth.
Then I'll iterate on look & feel; if I stall I'll say "make it better."
BONUS — a fun logo: a friendly minimalist 1-foot android mascot, rounded, 2 colors, flat vector, 3 options, transparent-background PNG.Build a SINGLE HTML page that ANIMATES my wow moment: [what the user sees in 30 seconds]. One file, real-ish data, no auth. Then I'll iterate on look & feel; if I stall I'll say "make it better." BONUS — a fun logo for [product]: flat vector mascot, bold simple shapes, 3 options, transparent-background PNG.
One sentence that displaces a specific alternative and names the failure mode in the buyer's words. Generate 20, keep 1.
Generate 20 positioning lines for an agent that finds the bugs that escaped to prod and writes the fix. ICP: on-call engineers at fast-shipping startups. Alternatives today: Sentry, Datadog, "just read the error log." Each must name a specific alternative I displace + the failure mode in the buyer's words. BANNED: AI-powered, seamless, revolutionary, next-gen, leverage. Score specificity, urgency, memorability. Show top 3, then pick THE winner.
Generate 20 positioning lines for a dating service that sends one curated intro a week — no feed, no swiping. ICP: late-20s/30s professionals who've deleted Hinge/Tinder repeatedly. Alternatives: the big apps, giving up. Each must name a specific alternative I displace + the failure mode in the buyer's words. BANNED: AI-powered, seamless, revolutionary, next-gen, leverage. Score specificity, urgency, memorability. Show top 3, then pick THE winner.
Generate 20 positioning lines for a 1-foot-tall desk humanoid that runs AI skills you can watch. ICP: makers and AI enthusiasts. Alternatives: a $20k full-size android, a smart speaker, nothing. Each must name a specific alternative I displace + the failure mode in the buyer's words. BANNED: AI-powered, seamless, revolutionary, next-gen, leverage. Score specificity, urgency, memorability. Show top 3, then pick THE winner.
Generate 20 positioning lines for [product, one sentence]. ICP: [specific buyer]. Alternatives they use today: [3 things]. Each must name a specific alternative I displace + the failure mode in the buyer's words. BANNED: AI-powered, seamless, revolutionary, next-gen, leverage. Score specificity, urgency, memorability. Show top 3, then pick THE winner.
AI builds the scaffold; you write the conviction. A pitch is about de-risking — why now, why you, why this wins.
You are a Khosla Ventures seed partner. Zero patience for buzzwords. Positioning: "The bugs your tests missed — already fixed." Traction: 22 repos scanned, 3 paying teams. Unfair advantage: 20 years finding bugs at Google + Microsoft. Build a simple seed deck: what we do · problem · solution+magic · why now · market · why we win · model · go-to-market · traction · team+ask. Per slide: 5 headline options, 3 bullets max, the ONE RISK it must de-risk. Then a 60-second narrative in MY voice. Then a skeptic pass as a tired VC — name the 3 weakest slides. BONUS: which VC firms + partners fit a dev-tools seed? Re-tailor the deck to the top firm + draft the intro email.
You are a Khosla Ventures seed partner. Zero patience for buzzwords. Positioning: "One real intro a week. No swiping." Traction: 40 on the first-city waitlist. Unfair advantage: product taste + a seed community. Build a simple seed deck: what we do · problem · solution+magic · why now · market · why we win · model · go-to-market · traction · team+ask. Per slide: 5 headline options, 3 bullets max, the ONE RISK it must de-risk. Then a 60-second narrative in MY voice. Then a skeptic pass as a tired VC — name the 3 weakest slides. BONUS: which firms + partners fit a consumer-social seed? Re-tailor the deck to the top firm + draft the intro email.
You are a Khosla Ventures seed partner. Zero patience for buzzwords. Positioning: "The future, one foot tall — on your desk today." Traction: 120 pre-orders on the landing page. Unfair advantage: maker + AI; can ship a small device. Build a simple seed deck: what we do · problem · solution+magic · why now · market · why we win · model · go-to-market · traction · team+ask. Per slide: 5 headline options, 3 bullets max, the ONE RISK it must de-risk. Then a 60-second narrative in MY voice. Then a skeptic pass as a tired VC — name the 3 weakest slides. BONUS: which firms + partners fit a hardware/consumer seed? Re-tailor the deck to the top firm + draft the intro email.
You are a Khosla Ventures seed partner. Zero patience for buzzwords. Positioning: [your winning sentence]. Traction: [any one number]. Unfair advantage: [why you]. Build a simple seed deck: what we do · problem · solution+magic · why now · market · why we win · model · go-to-market · traction · team+ask. Per slide: 5 headline options, 3 bullets max, the ONE RISK it must de-risk. Then a 60-second narrative in MY voice. Then a skeptic pass as a tired VC — name the 3 weakest slides. BONUS: which VC firms + partners fit? Re-tailor the deck to the top firm + draft the intro email.
Not "the market." The 10 who bleed most, the one channel they live in, and a price they'll say yes to.
My product: an agent that finds the bugs that escaped to prod and writes the fix. Wedge buyer: on-call engineers at fast-shipping AI startups. 1. My 10 most likely FIRST buyers — segment, role, the trigger that makes them need this now. 2. The ONE channel they already live in, ranked by how fast I reach them with $0. 3. Pricing: 3 options + the number they'd actually pay; how to package it. 4. One shareable launch moment + a weekly engine I can run solo. 5. Draft the reach-out that opens on THEIR pain (a prod incident), not my product.
My product: a dating service that sends one curated intro a week. Wedge buyer: the chronically single in ONE city/scene. 1. My 10 most likely FIRST buyers — segment, role, the trigger (just deleted the apps again). 2. The ONE channel they already live in, ranked by how fast I reach them with $0. 3. Pricing: 3 options + the number they'd actually pay; how to package it. 4. One shareable launch moment + a weekly engine I can run solo. 5. Draft the reach-out that opens on THEIR pain (swipe fatigue), not my product.
My product: a 1-foot desk humanoid that runs AI skills. Wedge buyer: makers and AI enthusiasts. 1. My 10 most likely FIRST buyers — segment, role, the trigger (wants the coolest desk gadget). 2. The ONE channel they already live in, ranked by how fast I reach them with $0. 3. Pricing: 3 options + the number they'd actually pay (pre-order vs retail); how to package it. 4. One shareable launch moment (a demo video) + a weekly engine I can run solo. 5. Draft the reach-out that opens on THEIR desire (own the future first), not my spec sheet.
My product: [one sentence]. Wedge buyer: [who bleeds most]. 1. My 10 most likely FIRST buyers — segment, role, the trigger that makes them need this now. 2. The ONE channel they already live in, ranked by how fast I reach them with $0. 3. Pricing: 3 options + the number they'd actually pay; how to package it. 4. One shareable launch moment + a weekly engine I can run solo. 5. Draft the reach-out that opens on THEIR pain, not my product.
Run synthetic users first, find the #1 bug, fix it, re-run — then show one real person.
Simulate 20 on-call engineers trying my escaped-bug finder. For each: first 30s, first action, where they get confused, drop vs wow, would they pay $30/mo? Aggregate: top 3 onboarding bugs, top 3 positioning bugs, the SINGLE biggest fix. Then tell me to fix it and re-run, and show the diff. Then: send it to one real engineer and watch — don't help.
Simulate 20 chronically-single users trying my one-intro-a-week service. For each: first 30s, first action, where they hesitate, drop vs wow, would they pay $25/mo before meeting anyone? Aggregate: top 3 onboarding bugs, top 3 trust/positioning bugs, the SINGLE biggest fix. Then tell me to fix it and re-run, and show the diff. Then: get one real single person to try it.
Simulate 20 makers hitting my 1-foot-robot pre-order page (animated demo, no hardware yet). For each: first 30s, first action, where they hesitate, drop vs wow, would they pre-order at $199? Aggregate: top 3 page bugs, top 3 positioning bugs, the SINGLE biggest fix, and the one skill that makes it a must-have. Then tell me to fix it and re-run, and show the diff. Then: get one real maker to react.
Simulate 20 [your ICP] users trying my product. For each: first 30s, first action, where they get confused, drop vs wow, would they pay $[X]? Aggregate: top 3 onboarding bugs, top 3 positioning bugs, the SINGLE biggest fix. Then tell me to fix it and re-run, and show the diff. Then: send it to one real [ICP] and watch — don't help.