ISB AI Founders · portal
1
Week / Day 1 · Wed Aug 5 / Brainstorming problems
Studio · taught

Brainstorming problems

Leave with ONE one-sentence problem thesis that survives a public gallery walk.

Open this session's deck →
10:45
start
90m
duration
5
exercises
3
things you leave with
Welcome

The most important session of your week, and it runs on paper and sticky notes before software. You will probably arrive with a vague ambition, something like an app for studying. You will leave with a sharp sentence about a real person and their real problem. Everything you build this week inherits its quality from that sentence.

You walk out with
✓One problem thesis in the format
✓A market scan with real competitor names
✓A wall of problems the whole cohort can see
Why this session exists

Bad code can be fixed in an afternoon. Building something nobody wants costs you the whole week. Offline exercises make the whole room's problems visible before anyone falls in love with an idea.

The arc, minute by minute
10:45
Problems, not ideas + the thesis format and its falsifiability bar
10:55
BLOCK 1 · Fill the wall: inventory, day replay, sticky storm, one 30-min run, self-paced
11:25
Whole room: cluster, dot-vote, claim a problem
11:33
The ridiculous ten: 10+ wild solutions on stickies, 5 min, no judging
11:40
BLOCK 2 · Pressure-test: Claude attack, market scan x2 models, ten-yuan filter, thesis, kill it yourself
12:00
Gallery gauntlet, then PLANT YOUR FLAG: everyone commits to their problem, starred on the wall
Exercises
01
Block 1 · Fill the wall
One continuous 30-minute paper block, three stages on one mission slide: 10 annoyances minimum, day replay with 5+ frictions circled, 6+ notes to the wall. Self-paced; staff circulate.
02
Whole room · The wall decides
Cluster into territories, dot-vote the hottest, claim a problem worth your week.
03
The ridiculous ten
Five minutes, sticky notes, no judging: ten or more solutions to YOUR claimed problem, ridiculous encouraged. Wild ideas hide real wedges; the AI re-runs this storm with you in the afternoon studio.
04
Block 2 · Pressure-test
One continuous laptop block: Claude attacks your problem, market scan in two models (name the strongest competitor and the gap), ten-yuan behavior filter, write the thesis, then try to kill it yourself. Exit checklist on screen.
05
The gallery gauntlet
Read your thesis aloud. The room hunts three holes: vague person, vague moment, fake workaround. Survive it and it goes up.
prompt · give this to the room verbatim
Here is my claimed problem: [X]. Evaluate it like a tough startup mentor. 1) Attack the weakest part. 2) Who already solves this, which one is strongest, and why has it not killed the problem? 3) Role-play the person who actually has this problem for three exchanges. 4) Then give me a sharper version of the problem in one sentence.
Tools in play
Paper firstClaude or ChatGPT
Materials to prep
Sticky notes, hundredsSharpiesDot stickersWall space + tape
The checkpoint · nothing counts unless it is live
Gallery walk: you read your thesis aloud. Vague ones die here, kindly. Make yours sharp enough to survive.
The handoff · what the next session assumes
Leave withONE falsifiable thesis that survived the gauntlet, on the wall
DecidedYour flag is planted: the problem you are spending your week on, starred on the wall and said out loud
Bring nextThe starred problem is what you pitch in speed-dating; the star clusters are where teams form at 13:15
← previousAI Foundationsnext →Co-founder dating