No connection? No problem. This is a full session — the exact prompts and the responses you'd get back. Read each prompt, read the response, do the ✎ paper step, move on. When wifi returns, paste any prompt and keep going live. Same flow as the slides: 6 stages, 6 artifacts.
STAGE 01Ideation — generate, then attack
YOU → AI
You are a serial founder who has killed more ideas than you've shipped.
I want to build a consumer humanoid robot. My unfair advantage: I can ship
hardware cheaply and I know the maker community. Hair-on-fire problem: people
want a robot companion, but full-size androids are $20k and terrifying.
Generate 10 angles, score 1–10 on wedge, buyer urgency, defensibility,
distribution. Kill 8. Defend 2. Write each survivor's obituary.
AI
01 Full-size home android 2/10 — $20k, years, scary
02 Robot vacuum, but cuter 3/10 — iRobot owns it
03 1-foot-tall desk humanoid 9/10
04 Telepresence robot 4/10 — covid fad faded
05 Kids' coding robot kit 5/10 — Sphero land
06 Robot pet (no humanoid) 7/10
07 Warehouse pick-arm 3/10 — capital-heavy B2B
08 AI smart speaker with a face 4/10 — Alexa already lost
09 Animatronic collectible 4/10 — novelty, no repeat
10 Open-source robot brain SDK 5/10 — devs don't pay
SURVIVORS
#03 + #06 are the same little body. The wedge is
tiny + humanoid + safe on a desk — companionship at 1/20th the price and zero intimidation.
OBITUARY
"1-Foot Robots died at month 10: a cute gadget people used twice. It survived only where the robot had a
reason to exist on your desk daily — a personality + a job, not just a wave."
YOU → AI (the attack — in the room you'd attack a pod-mate's idea)
Now destroy it. Attack as a skeptical seed investor, then the buyer who'd churn.
AI
INVESTOR · WEDGE 6/10
"Consumer hardware is a graveyard — Anki, Jibo, Kuri all died with funding." Fair. They died on
burn, not demand.
The moat is the personality/software that makes a $200 toy feel alive — and a margin that survives without VC oxygen.
BUYER CHURN · WEEK 3
· "Cute for a day, then it lives in a drawer."
· "It just does what my phone does, but slower."
·
"A wave" is a demo, not a reason to keep it charged.
THE VERSION THAT SURVIVES
A
desk companion with a daily job — reacts to you, does one genuinely useful thing (pomodoro coach / standup buddy / mood check-in), and has a personality that grows. The software is the product; the body is the smile.
✎ ON PAPER: circle the survivor. Write one obituary line for your idea in the workbook.
STAGE 02Product — animate the wow in HTML, then iterate, iterate, iterate
YOU → AI · round 1
Build a SINGLE HTML page that ANIMATES my wow moment: a 1-foot robot sits on a
desk, looks up when you "arrive," waves, and a speech bubble says "Ready to
focus? 25-min sprint?" with a Start button. One file, no auth.
AI
INDEX.HTML — ONE FILE, OPENS IN ANY BROWSER
<div class="desk">
<div class="bot">🤖</div> <!-- head tilts, eyes blink -->
<div class="bubble pulse">Ready to focus? 25-min sprint?</div>
<button>Start sprint →</button>
</div>
What you see: the bot wakes, eyes track to you, a little wave → "Ready to focus?" pulses → Start. Five seconds, and you get why it's on your desk.
YOU → AI · round 2 (react to what you saw)
The wave is the magic but it feels robotic and dead-eyed. Give it a personality
— a name, a mood that shifts, and make the focus pitch sound like a friend
nudging me, not a productivity app.
AI
Named it Pip. Added an idle loop (looks around, fidgets) and a mood meter that brightens when you finish sprints. Rewrote the line. Before: "Ready to focus? 25-min sprint?" After: "You've been staring at that tab for a while 👀 — wanna knock out 25 minutes with me?"
YOU → AI · round 3 (out of ideas? hand it the wheel)
I'm out of ideas. You take it — make it more convincing as a daily companion.
AI
Added: a streak counter Pip celebrates with a tiny dance · a midday "how's it going?" check-in · Pip slumps if you ignore it for hours (gentle guilt, not nagging). Pro tip: when you stall, asking the AI to improve it is a real strategy — then react to what it gives you.
YOU → AI · round 4 — logo ideation
Now the logo / mascot mark. Give me 5 fun concepts for "1-Foot Robots" —
playful and friendly, but it should still read as a real product, not a kids'
toy. One line each.
AI
1 A tiny robot standing on a ruler "1 ft" mark — the size IS the brand
2 Robot head winking, antenna shaped like a heart — friendly companion
3 A robot inside a lowercase "ft" ligature — name baked into the mark
4 Two-tone rounded bot silhouette, big single eye — Pip's face as logo
5 A robot waving, footprint as a power-button circle — alive + on
Strongest: #1 (the one-foot size is the whole pitch) and #4 (works tiny as an app icon).
YOU → AI · round 5 — iterate the logo like the product
Love #1. Iterate: warmer, rounder bot, make the "1 ft" unmistakable at a glance.
Give me 3 variants as TRANSPARENT-BACKGROUND PNGs — 1024px for the site,
512px square avatar, 32px favicon. Flat vector style.
AI
3 VARIANTS — ALL TRANSPARENT-BACKGROUND PNG, 1024 / 512 / 32
A) Rounded bot on a clean ruler, bold "1 ft" tick — crisp and literal.
B) Same + a soft glow ring — reads "alive / powered on."
C) Just Pip's two-tone head, "1ft" tucked in the chin — sweeter, app-icon first.
CHECK IT LIKE A TESTER
At 32px, A's ruler ticks turn to mud —
C wins the favicon, A wins the homepage header. Asking for transparent PNGs at three sizes up front is the trick: you see instantly where a logo breaks.
YOU → AI · round 6 — close the loop
Drop variant A into the page header and put Pip's face (variant C) as the
favicon. Reload me the page.
AI
Done — header carries the ruler mark, favicon is Pip's face. The page went from demo to brand in six rounds. That's the loop: build → react → iterate → hand it the wheel → brand it.
✎ ON PAPER: sketch your page in 4 boxes (input → animation → pulse → button), then doodle TWO logo concepts and circle the one that survives at favicon size.
STAGE 03Positioning — 20 lines, keep one
YOU → AI
Generate 20 positioning lines for 1-Foot Robots (a tiny desk humanoid companion
with a personality and a daily job). ICP: remote workers who want company at
their desk. Alternatives: a $20k android, a smart speaker, nothing. Score
specificity, urgency, memorability. Kill 18. Show top 2. Pick the winner.
AI
"The future of robotics" 1/10
"AI companion, reimagined" 1/10
"Your personal robot assistant" 2/10
"Big robot energy, small package" 5/10
…14 more, killed.
TOP 2
"A one-foot robot that actually lives on your desk — and gives a damn about your day." — 9/10
"All the company of a robot. None of the $20k or the uncanny valley." — 8/10
WINNER
The first — it names the size, the place, and the emotional job, instead of the category.
✎ ON PAPER: write your own winner. Read it aloud to your pod — would a buyer repeat it?
STAGE 04Pitching — simple Khosla seed deck
YOU → AI
You are a Khosla Ventures seed partner. Zero patience for buzzwords.
Positioning: "A one-foot robot that actually lives on your desk — and gives a
damn about your day."
Traction: 220 pre-orders at $199, 70% from one maker-community drop.
Build the simple seed deck: what we do · problem · solution + magic · why now ·
market (bottoms-up) · why we win · business model · go-to-market · traction ·
team + ask. For each: a headline and the ONE RISK it must de-risk.
Then a 60-second narrative in MY voice. Then a skeptic pass — 3 weakest slides.
BONUS: which investors fit? Re-tailor to the top firm, draft the intro email.
AI
1 WHAT WE DO "A $199 desk robot you actually keep on." risk: toy, not company
2 PROBLEM "Robot companions are $20k and scary." risk: is it a need?
3 SOLUTION 1-foot humanoid + personality + a daily job risk: retention
4 WHY NOW Cheap actuators + LLMs make it feel alive risk: fad?
5 MARKET Remote workers × desk gadgets, bottoms-up risk: TAM theater
6 WHY WE WIN Personality software + cheap hardware stack risk: copyable?
7 MODEL $199 device + $5/mo personality plus risk: hardware margin
8 GTM Maker drops → creator unboxings risk: scale past fans
9 TRACTION 220 pre-orders, 1 community drop risk: tiny
10 TEAM + ASK $1.5M → tooling + first production run risk: hardware is hard
60-SECOND NARRATIVE (your voice)
"Everyone wants a little robot buddy, but the real ones cost $20k and stare into your soul. We make a one-foot humanoid for $199 that lives on your desk, knows your name, and actually nudges you through your day. Two hundred people pre-ordered from one community drop. The hardware is cheap on purpose — the personality is the product."
SKEPTIC PASS — WEAKEST 3
Retention (slide 3), hardware margin (7), why-you (10). Fix retention first — show the streak/check-in loop that keeps Pip charged.
BONUS — INVESTOR FIT
Hardware-friendly seed funds and angels who backed consumer robotics / playful hardware (the Anki/Sphero/Nothing lineage, eyes-open about the graveyard). Re-tailored hook: "We learned why Jibo died — and built the margin and the loop it didn't have." Intro email drafted — 60 words, one 15-min ask.
✎ ON PAPER: write YOUR hook in your own voice. Say it to your pod in under 10 seconds.
STAGE 05Go-to-Market — first buyers, reach, price
YOU → AI
Who are my 10 most likely FIRST buyers, with the trigger that makes them need
this now? The ONE channel to win first with $0? Pricing — 3 options and the
number they'd actually pay? Then the reach-out that opens on THEIR pain.
AI
FIRST BUYERS
Remote-working makers, gadget early-adopters, and desk-setup obsessives. Trigger: they just posted a "battlestation" photo or backed a hardware crowdfund this quarter. (Also: parents wanting a non-screen desk buddy for a kid's homework hour.)
THE ONE CHANNEL
Maker / desk-setup communities + creator unboxings. One good 30-second "Pip waves and roasts me into focusing" clip beats any ad. Win the makers; they film it for you.
PRICING
$149 early-bird ·
$199 standard + $5/mo personality plus (the product) · $299 founder's edition. Anchor against the $20k android, not other gadgets.
REACH-OUT (cold DM to a desk-setup creator)
"Your battlestation is unreal but it's missing one thing — something that looks back. Pip's a $199 one-foot robot that lives on your desk and nudges you to focus. Want one to film before we ship?"
LAUNCH + ENGINE
Launch: a community drop with a numbered founder's run. Engine: every owner clip of Pip doing something charming becomes the ad — the personality markets itself.
✎ ON PAPER: name your first buyer, your one channel, and a price. Three blanks, fill them now.
STAGE 06Validation — synthetic users first
YOU → AI
Simulate 20 remote workers hitting the 1-Foot Robots page. For each: first
30 sec, where they get confused, drop or wow, what they'd tell a friend, would
they pre-order at $199? Aggregate: top 3 bugs + the single biggest fix.
Then re-run.
AI
20 PERSONAS RUN
13/20 hit the wow ("it looked up at me and I smiled — I want one on my desk").
BUG · 9/20 hesitated: "cute, but will I still use it in a month?" — the daily-job loop isn't visible on the page.
11/20 would pre-order at $199. 6 said no — "feels like it'll end up in a drawer." 3 wanted a cheaper version.
TOP 3 FIXES
1) Show the streak / daily check-in loop right on the page, not just the wave. 2) A "30 days with Pip" timeline so buyers see longevity. 3) Real owner clip for proof.
RE-RUN AFTER FIX #1
17/20 wow (was 13) · pre-order intent 15/20 (was 11). Ship it — then put a real Pip on one person's desk. ✎ ON PAPER: predict YOUR #1 buyer hesitation, and the one fix you'd ship. Tell your pod.