Colaborative Writing with AI
Craft Focus
→ Multi-idea brainstorming
→ Filtering the good from the hilarious
→ Using AI to push beyond clichés
When someone says “let’s do a brainstorm session,” a very specific image tends to appear.
People standing around a whiteboard.
Markers uncapped.
Ideas written in neat bubbles.
Someone eventually circling the “good ones.”
That’s not what this is.
The image at the top of this post isn’t a metaphor for creativity—it’s a description of what brainstorming actually feels like when it’s working. Multiple ideas colliding at once. Different kinds of pressure happening simultaneously. No single storm to point at and say that one’s the problem.
Real brainstorming isn’t orderly, and it isn’t polite.
It’s loud in the head, contradictory, and occasionally overwhelming.
And paradoxically, that’s usually when something useful is about to surface.
Chaos Isn’t the Enemy of Structure
Most of the time, when we think about “brainstorming,” we imagine something controlled.
A notebook.
A plan.
A clear objective.
In practice, most of my best ideas have come from conversations that weren’t really about writing at all. More look at that cloud, it looks like a dinosaur than let’s outline a chapter.
Sometimes the conversation wandered.
Sometimes it stalled.
Sometimes it turned into a joke that shouldn’t have worked.
Those weren’t distractions.
They were signals.
Chaos isn’t where stories fall apart.
It’s where unexpected structure quietly forms—if you let it breathe long enough.
Starting With the Hardest Thing on Purpose
They say comedy is the hardest thing to write.
So naturally, that’s where we’ll start.
Not because this exercise is about comedy—but because a joke exposes structure in a way few other things do.
A punchline is useless on its own.
It assumes context.
It implies a misunderstanding.
It demands explanation.
Which makes it the perfect creative spark.
The Exercise: Build the Joke That Explains the Punchline
This is an open exercise.
There is no “correct” outcome.
You are not meant to compare what you create to anyone else’s result.
Step 1 — Start With a Punchline
Choose or invent a line that clearly feels like the end of a joke.
It should sound wrong without context.
It should imply that something has already gone sideways.
Don’t explain it.
Don’t justify it.
Just write it down.
Step 2 — Treat the Punchline as Evidence
Instead of asking “How do I build toward this?”, ask:
- What must be true for this line to exist?
- Who is misunderstanding something?
- What assumption is quietly being protected?
You’re not writing backward—you’re reverse-engineering logic.
Step 3 — Let the World Make the Mistake Reasonable
Now start adding pieces that make the misunderstanding plausible:
- A setting where clarity is avoided
- A social rule that discourages questions
- A character who doesn’t want to look foolish
The goal isn’t escalation yet.
The goal is believability.
Step 4 — Escalate Through Confirmation
Things don’t get stranger because they’re random.
They get stranger because everything keeps almost making sense.
Each beat should reinforce the wrong assumption just enough to keep it alive.
Step 5 — Stop When the Punchline Feels Inevitable
The moment you realize the line could not not happen—stop.
That’s the story.
A Note on Writing With AI
This exercise works best if you don’t treat the AI like a vending machine.
Don’t nudge it toward answers.
Don’t try to out-prompt it.
And definitely don’t ask it to be funny.
There’s a reason it’s called ChatGPT.
Talk to it.
Let the conversation wander a little. Say things out loud that you’re not sure about yet. Push on ideas the same way you would in a writer’s room—half joking, half curious, not fully committed to anything.
If you come at it with a checklist, the AI will give you exactly that back: tidy, efficient, and mostly uninteresting.
If you come at it sideways—“This feels wrong but I don’t know why yet”—that’s when something useful tends to surface.
The goal here isn’t to get answers.
It’s to keep the misunderstanding alive long enough to explore it.
When AI works well in this exercise, it’s not because it solved the joke for you. It’s because it kept asking “What if?” when you were about to settle too early.
Treat it less like a tool and more like a slightly over-eager collaborator who doesn’t mind sounding ridiculous on the way to something true.
That’s where the good stuff usually hides.
Why This Works
This exercise works because it flips how we usually approach ideas.
Instead of starting with intention, you start with evidence.
Instead of chasing originality, you chase logic.
Instead of forcing structure, you discover it.
Comedy just happens to make the process visible.
Once that clicks, you can apply the same thinking to a character, a line of dialogue, a setting, or even an entire world.
The punchline is just the spark.
The real skill is learning how to ask why it exists at all.
Closing Thought
Chaos doesn’t ruin stories.
Unexamined chaos does.
When you learn how to let ideas collide without rushing to organize them, you don’t lose the plot—you uncover it.