| name | ripple-mapping |
| description | Brainstorm non-obvious startup and product ideas from a technology shift by mapping its second-order consequences (the "And Then What?" game). Use when the user wants to brainstorm what to build, asks "what should I build with X," wants startup ideas around a new technology, or asks about second-order effects / picks-and-shovels opportunities of a trend. |
Ripple Mapping
When a big technology shift happens, 99% of builders rush to the same obvious application. That space gets crowded instantly and is nearly impossible to win. The golden ideas are found by assuming the obvious apps win, imagining millions of people using them, and building for the new problems that mass adoption creates. You don't compete with the wave — you sell tools to the people riding it.
This skill produces a Ripple Map: a structured four-step derivation from a technology shift to defensible second-order business ideas.
Operating mode
- If the user supplied a technology shift (even roughly), run the full four-step map in one pass, then offer the follow-ups listed at the end.
- If the shift is vague ("AI", "crypto", "robots"), do NOT run the map yet. First sharpen it with the user into a specific cost/speed/access delta (see Step 1). One clarifying exchange, then run.
- If the user supplied no shift at all, propose 3–4 candidate shifts happening right now, let them pick, then run.
The four steps
Work through the steps in order, producing every step's output explicitly. Never skip ahead to ideas. The discipline of the detour is the entire method: your first instincts are Step 2 material and belong in the discard pile, not the recommendation.
Step 1 — The Big Shift
State the shift as a delta, not a product: "X went from expensive/slow/gated to cheap/instant/open." Quantify if possible (10x, 100x, from $N to ~$0). This is observed, not built.
Bad: "AI video is big now."
Good: "Producing a broadcast-quality video clip went from ~$5,000 and a crew to ~$0.50 and a sentence."
Step 2 — The Obvious Trap
Enumerate 5–10 obvious ideas that anyone with a laptop is building this weekend. List them specifically — this exhausts the default answers so they can't leak into Step 4. Label the list clearly as the discard pile.
Litmus test: if an idea is "use the new tech to make a slightly better version of an existing app," it goes here. So does anything that will appear 50 times in the next demo day batch.
Step 3 — The New Mess
Assume the trap ideas completely win. Millions of daily users. Now derive what breaks. Produce 4–6 messes, each stated as a concrete breakdown, not a missing feature.
Messes come from adoption at scale — look in these categories: congestion and noise, trust and verification, fraud and abuse, compliance and liability, coordination between parties, waste and cleanup, discovery in overwhelming supply, maintenance and decay, skill gaps and labor displacement.
Two quality checks every mess must pass:
- Scale test: Would this problem exist if only 100 people used the trap apps? If yes, it's a feature gap, not a ripple — cut it.
- Consequence test: Is this caused by the apps succeeding, not by them being bad? "The AI writes poorly" fails; "nobody can tell what was written by whom" passes.
Rank the messes by how painful and how inevitable they are.
Step 4 — The Real Business
For the top 2–3 messes, propose the specific tool, infrastructure, or utility that cleans it up. Each candidate must pass three gates, answered explicitly:
- Who pays, and why? A named buyer with money and urgency — not "users would love this."
- Why won't the wave-makers absorb it? If the platform creating the wave will obviously ship this as a feature, cut it. Good answers: it's neutral infrastructure across competing platforms, it's a liability they want off their books, it's a market too unglamorous or fragmented for them.
- Timing: Is the mess visible today, or reliably predictable? "Visible now" beats "predictable"; "speculative" gets cut.
End with a short verdict: which single candidate you'd pursue and the one-sentence pitch for it.
Calibration example: the car
- Shift: The gasoline engine makes individual long-distance travel possible.
- Trap: Hundreds of founders start car companies, compete brutally on price, and mostly go bankrupt.
- Mess: Millions now drive — crashes everywhere, chaotic roads, cars stranded without fuel.
- Real business: Traffic lights, gas stations, car insurance. Fortunes made solving the consequences of the car, not building the car.
Match this level: the winning ideas sound boring next to the trap ideas. That's the signal, not a flaw.
Output format
Present the result as a Ripple Map in this shape:
# Ripple Map: <shift, one line>
## 1. The Big Shift
<delta statement, quantified>
## 2. The Obvious Trap (discard pile)
<numbered list, 5–10 ideas — explicitly marked as what NOT to build>
## 3. The New Mess (assuming the trap ideas win)
<ranked messes with one-line "what breaks" each>
## 4. The Real Business
<2–3 candidates, each with: idea, who pays, why the platforms won't absorb it, timing>
**Verdict:** <the one to pursue, and why>
After the map
Offer these follow-ups (don't run them unprompted):
- Ripple again: treat a Step 4 winner as a new Step 1 and map its second-order ripples (gas stations spawned convenience stores). This is where the least crowded ideas live.
- Go deeper on mess #N: expand one mess into a full landscape — buyer segments, existing partial solutions, wedge product.
- Stress-test a candidate: argue the bear case — who kills it, what has to be true.
