Prompting

Make Your Idea Fight to Survive: The Free Claude Idea Evolution Prompt

One copy-paste prompt that makes Claude write 10 different versions of your idea, sends a customer, a competitor and an investor after each one, kills the weak ones and breeds the survivors for 5 generations. You get one idea still standing, why every other one died, and the first thing to test. The full prompt, how to use it, and the catch

5min read
2prompts to copy
10numbered steps
5sections

Ask Claude about your idea and it will usually tell you it's a good one. That feels nice, and it's useless. Real ideas don't get judged by a friendly chat. They get judged by customers who won't pay, competitors who copy you, and money that doesn't add up.

This free prompt turns Claude into a jungle. It writes 10 different versions of your idea, sends three hungry predators after each one, kills the weak ones, and mixes the best parts of the survivors into the next generation. After 5 rounds you get one idea still standing, plus why every other one died.

What it does

  1. Writes 10 truly different versions of your idea: different customer, problem, money model or way of reaching people.
  2. Sends three predators after each one: a customer (would I pay or care?), a competitor (how do I crush this?) and an investor (does it make money and grow?). Each names the one reason it dies and gives a score.
  3. Shows a scoreboard, kills everything but the top 3, and puts the dead ones in a graveyard with the line that killed them.
  4. Breeds 10 new versions from the best parts of the 3 survivors, and makes the survivors fight their children.
  5. After 5 generations: the survivor, why it survived, its biggest weak spot, the full graveyard, and one cheap test to run this week.

The Idea Fight prompt

PASTE INTO CLAUDE, THEN ADD YOUR IDEA
Run an Idea Fight on my idea. Your job is NOT to agree with me. Claude tends to like whatever idea it is shown, so for this whole run you are a referee, not a fan. No praise, no "great idea", no softening. If an idea would die in the real world, say so plainly.
GENERATION 1: WRITE 10 VERSIONS
Write 10 genuinely different versions of my idea. Change the real bones each time: who it's for, what problem it solves, how it makes money or wins attention, and how it reaches people. No near-copies with a new name. Give each one a short name and a one-line pitch.
THE PREDATORS
Three predators attack every version, each in role:
- CUSTOMER: "Would I actually pay for this or care about it? What would I use instead, or do I just ignore it?"
- COMPETITOR: "I already exist in this space. How do I copy or crush this in a month?"
- INVESTOR: "Does this make money, and can it grow past the first 100 people? Where do the numbers break?"
Each predator names the ONE most likely reason this version dies in the real world, in one plain sentence, and gives a survival score from 1 to 10 (10 = this attack barely hurts it). Be specific to this version. A generic risk like "competition is tough" doesn't count. Don't invent facts, numbers or companies; if you aren't sure something is true, say "assumption:" in front of it.
SCOREBOARD
After every generation show a short table: rank | name | customer | competitor | investor | total (out of 30) | how it dies (or "survives"). Lowest totals at the bottom.
SELECTION AND BREEDING
- Kill every version except the top 3. Add each killed one to THE GRAVEYARD with its name and the one line that killed it.
- Breed the next generation: write 10 new versions by mixing the best parts of the 3 survivors (for example the customer of one, the money model of another, the reach of the third). At least 2 of the 10 must also fix the exact attack that hurt the survivors most. Keep the 3 survivors in the next round so they have to beat their children.
- Run the predators and the scoreboard again on all 13.
Repeat for 5 generations in total (or the number I give below). Keep each round short: the scoreboard, the graveyard additions and one line on what changed. Never pass a weak idea through just because it's mine.
FINAL RESULT
1. THE SURVIVOR: the one idea still standing, its pitch in 2 sentences, and its scores.
2. WHY IT SURVIVED: the 3 attacks it beat and how.
3. ITS BIGGEST WOUND: the attack it still hasn't fully beaten.
4. THE GRAVEYARD: every dead version, one line each on why it died.
5. TEST THIS WEEK: the single riskiest assumption the survivor depends on, and one cheap test I can run in the next 7 days with real people to check it (what to do, who to ask, and what result would mean "keep going" or "stop").
Generations: 5
My idea: [one or two lines about your idea]

If it stops early

Long fights can run out of room in one reply. If Claude stops before the final result, paste this:

NEXT GENERATION
Keep the Idea Fight going from where you stopped. Same rules: no flattery, the same three predators (customer, competitor, investor), keep the top 3, breed 10 new versions from their best parts, show the scoreboard and the graveyard. Run the next generation, and if it was the last one, give me the full final result (survivor, why it survived, its biggest wound, the graveyard, and the one test to run this week).

Free, for your email

Get the full setup steps + the copy-paste setup prompt

The rest of this guide (2 more sections) opens right here. The setup prompt, one paste into Claude that sets it all up, comes by email only. Use your real address.

50,000+ follow where these guides come from. No spam, unsubscribe in one click.

Read next