I keep seeing smart founders stare at a blank doc, type one prompt into ChatGPT, and get back ten ideas that all sound like cousins. Then they blame themselves for not being creative enough. I used to nod along with that story too, until I read a post from an AI founder that flipped my thinking completely.
The original poster runs an AI startup, and their whole argument is simple: this isn’t a creativity problem, it’s a prompting problem. As the author puts it, most people still brainstorm like it’s 2015. One prompt, one angle, and whatever idea the room agrees on fastest. Half the time that’s just the easiest option, not the right one.
What grabbed me is how this creator brainstorms now. No whiteboard, no loudest-voice-wins. They use structure instead of vibes: constraints, reversals, first principles, multiple perspectives. Each move pushes the model somewhere your own brain wouldn’t wander on its own. The expert says that single shift changed how they decide what to build next.
To make it usable, this industry pro pulled together 20 of their sharpest ChatGPT brainstorming prompts into one infographic. They shared a taste of six, and I want to break those down because each one solves a specific kind of stuck.
The 6 prompt frameworks worth stealing
- Rapid Idea Generator. This one asks ChatGPT for 30 ideas in a single shot, deliberately mixing the practical with the wildly ambitious. The point isn’t to love all 30. It’s to blow past your first three safe answers and force volume, because the interesting stuff usually hides around idea number 22.
- Reverse Brainstorming. Instead of asking how to solve the problem, the author flips it: how could you make the problem worse? List every way to sabotage yourself, then invert each one into a fix. It sounds silly, but naming the failure modes out loud surfaces solutions you’d never reach head-on.
- First-Principles Brainstorming. This framework strips a problem down to what’s actually true, no assumptions carried over from how everyone else does it. You break the thing into its rawest parts and rebuild from there. It’s the difference between a faster horse and a car.
- Constraint-Based Brainstorming. Here the creator hands the model tight limits on purpose, like no budget, one week, or a team of two. Constraints do the filtering for you. A wide-open prompt gives mush, but a boxed-in prompt forces sharp, resourceful answers.
- Competitor-Inspired Brainstorming. This one studies what already works for others and borrows the underlying mechanic without copying the surface. You’re not cloning a rival, you’re reverse-engineering why their move lands, then applying that logic to your own situation.
- Brainstorm, Score, and Rank. The finisher. You generate wide, then have ChatGPT score and rank every idea by impact versus effort. It turns a messy pile into a prioritized shortlist, so you leave with a decision instead of just a longer list.
Those are only six of the twenty. The through-line across all of them is the same idea the post’s author keeps hammering: a generic prompt pulls generic thinking out of you. A structured prompt pulls out the ideas that were buried under your obvious ones.
Why it matters: the model is only as good as the angle you hand it. Change the angle, and the same tool that gave you ten samey ideas suddenly gives you the one you’d have missed.
How to actually use this
I think the smartest tip from this contributor is the least flashy one. They suggest keeping the prompt collection somewhere you’ll see it, right next to your notes app, so you can pull it out the exact moment you’re stuck staring at a blank page. Good frameworks only help if they’re in reach when the panic hits.
Here’s how I’d put these to work this week:
- Start wide, then narrow. Run Rapid Idea Generator first for raw volume, then feed the output straight into Brainstorm, Score, and Rank to cut it down.
- Reach for Reverse Brainstorming when you’re blocked. If nothing’s coming, don’t push harder on the solution. Ask how to make it worse, then flip your list.
- Use First-Principles when you’re copying the industry by default. Whenever you catch yourself saying “that’s just how it’s done,” that’s your cue to strip it back to what’s true.
- Add Constraints when the answers feel generic. Tell the model the budget is zero and the deadline is Friday. Watch the quality jump.
There’s a bigger trend sitting under all of this. As more people use the same AI tools, the edge stops being access to the model and becomes the quality of the questions you ask it. The mind behind this post basically built a repeatable questioning system, and that’s a skill worth developing whether you’re a founder, a marketer, or anyone who has to produce ideas on demand.
If you build things for a living, this is the kind of thing worth having on hand. Head over to the full LinkedIn post to see the complete breakdown and the rest of the twenty prompts, then tell me which one you’re trying first.