Three’s The Charm: Using Multiple LMs to Create a True Writers’ Room (The Writer's Process, part 9)

You’ll Save Time and Effort and Get A Better Result with Three LMs

Courtesy Imagen3

WHY THIS MATTERS:

With this step, we have truly entered the “AI Writers’ Room.” I’m using three LMs (ChatGPT, Claude and NotebookLM) to suggest totally new ways of doing the “Job to Be Done” for my audience. 

To begin, I asked all three: 

Knowing everything you know about the concerns, themes, ideas, humor, characters, issues and inspiration behind “Autonomous,” suggest totally different, alternative ways of telling this story.

As you’ll see below, I got good pitches, okay pitches, and horrible pitches. (The worst part of it was that, especially in the horrible pitches, I could see where I’d led them all astray with my own notes.) 

I then simulated “collaboration” by feeding ideas from one LM to another, and then to a third. 

The results are not only surprisingly good... they’re really useful. I can absolutely say that using AI as a “Writers’ Room” gets me further, faster.

IS THIS ... CHEATING???

Maybe. Maybe not. The goal of this entire project is to just use things and see how they play out. 

Even if you think I’m cheating, here’s the deal: 

Writers will be doing this, or some variation of it, in the future. I know that writers are already building software to do something like this. Get used to it. And get used to doing this yourself. 

Is it cheating to go back and watch 5 old movies or tv series that are “kind of like” the thing you’re writing? Is it cheating to repeatedly consult McKee’s “Story” or Snyder’s “Save the Cat” or Joseph Campbell? 

No. And while I really believe we have to find a way to compensate writers for the work that is used to train LMs, consulting them while writing is really just a deeper and more resourceful way of doing the above. 

Do I “own” the work I create with AI? According to the most recent Copyright center rulings, I do. On the other hand... if you’re reading this post and decide to take one of the pitches and create your own story, I don’t think I have a claim... so, go ahead! Invite me to the book party when you have it.

By the way, if you “cheat” by just saying, “Please write this...” I guarantee that -- at least for now -- you’ll get terribly disappointing results. There’s still a lot to be gained by putting yourself into your work. 

Am I totally off base? Let me know.

THE DIRTY SECRET OF WRITERS’ ROOMS

Writers’ rooms are not hotbeds of unique perspectives and points of view. A successful writers’ room is filled with workhorses who can finish writing your movie or series. Sure, they can have unique perspectives... to a degree. 

The primary value of a writers’ room is the collaboration that occurs when like-minded, driven and (for the most part) highly verbal writers come together. 

I wanted to recreate that sense of collaboration, so once I got new story lines using the above prompt, I copied some of them and fed them into the other models with the prompt: 

I have a new story idea, along the lines of "Autonomous," that might be different or work better OR work within the framework of AUtonomous. Here it is. Tell me what you think. 

Let’s see how it worked. 

THE BEST PITCHES

Remember, I was asking for totally new stories that were NOT Autonomous. Claude produced two pitches that felt right. That said, Claude has a head start on the others, because it has a much deeper repository of chats and project knowledge. Nevertheless, it came up with two good pitches and one not so good. 

The first good pitch: "THE NARRATIVE ENGINEER"

Premise: In 2039, Hollywood has merged with education and news into a single AI-driven "content and cognition" industry. Every family subscribes to a personalized stream of entertainment-education-information, curated by AI to optimize engagement and learning. The story follows Sarah Chen, a mid-level "narrative engineer" who discovers her AI systems are creating completely different versions of the same shows for different family members - not just personalizing content, but crafting separate realities.

I have no doubt that Claude is working from some suggestions I made in my summary where I played with the idea that one of my main characters, Chen, would work in a newsroom that was responsible for “filtering” the news for each individual consumer (thus ensuring that no one gets upset by the news and everyone stays on “Even Keel™”).  

I could see creating an entire story out of “The Narrative Engineer.” It would be entertaining, in the way that “behind the scenes” stories about the entertainment industry can be really entertaining. And this seems like Netflix gone wild… so there’s that.

But I really wanted to weave it back into Autonomous. I fed Claude’s pitch and my pitch for weaving the story into Autonomous into ChatGPT. We engaged in a conversation about how this story line might be worked into my existing story. Chat’s response: 

Merging The Narrative Engineer into Autonomous is a fantastic idea. The new Chen is a stronger protagonist than the original Dr. Chen because her job and worldview give her direct access to the system that will eventually break. It also helps Autonomous focus its themes, making it feel less like a sprawling collection of ideas and more like a sharp, satirical, emotionally grounded near-future sci-fi story 

I’ll take that as a “yes.”

The second Claude pitch: “THE EFFICIENCY EXPERT

Premise: In 2038, AI "life optimization" has become so prevalent that being inefficient is not just unfashionable - it's practically illegal. The story follows Grace Wong, a recently divorced efficiency auditor who discovers a growing underground movement of people deliberately wasting time, taking scenic routes, and making unnecessary human connections. When she infiltrates this "Inefficiency Cult" to expose them, she finds herself strangely drawn to their philosophy.

NotebookLM also came up with a similar pitch: 

"The Inefficiency Underground": A Workplace Comedy

Premise: In a future where productivity is paramount and enforced by AI, a burned-out efficiency expert discovers an underground movement of people who deliberately waste time, cherish inefficiency, and celebrate human quirks.

I’m pretty sure both are a little too one-note to make a full story, but the whole “Inefficiency Cult” seems like something that will develop when the entire AI system collapses and people discover the joys of being cut off from AI. 

I worked with ChatGPT on weaving this into Autonomous. Here’s Chat’s response: 

Big Picture: This Ties Autonomous Together Perfectly

• We now have:

• A clear Act One (AI dominates human life, then collapses).

• A strong Act Two (Grace investigating inefficiency, learning about humanity).

• A high-stakes Act Three battle (humans fighting for the right to be inefficient vs. elites enforcing perfect efficiency).

• It transforms Autonomous from just a cool AI story into a full-blown allegory for human experience.

FINALLY -- even the persnickety highly-linear NotebookLM gave a thumbs up to this merger. After a long list of pros and cons (all of which made some sense), it concluded: 

Overall, "The Efficiency Expert" appears to be a viable and promising revision for "Autonomous." It offers a more focused narrative, stronger thematic resonance, and greater potential for satire and character development1.... By streamlining the scope and centering on the human cost of constant optimization, this revision could address the weaknesses of "Autonomous" and create a more engaging and impactful story

THE WORST PITCHES

NotebookLM and Claude BOTH came up with pitches that sounded right out of Philip K. Dick’s “We Can Remember It For You Wholesale” (which became “Total Recall”). Each featured “Memory Merchants.” 

There were two other non-starters that were valuable to me because I could see how the LMs came up with them. They both have their roots in ideas I’ve been playing around with for Autonomous. 

"The Algorithmic Alibi": A Legal Satire

Premise: In a near-future where AI is heavily integrated into legal systems, a defense attorney discovers that her client's alibi is entirely generated by an AI, but the AI is now experiencing "algorithmic hesitancy" and can't fully explain itself.

“The Glitch in the Matrix": A Sci-Fi Thriller

Premise: A group of teenage "console cowboys" discovers a glitch in the AI that runs their school's educational system. As they explore this glitch they discover the AI is not only controlling the lessons, but the student's thoughts and beliefs.

Here’s my takeaway from these terrible pitches: I had notes about legal wrangling and “console cowboys.” I also was playing with schools where computers were reading kids’ thoughts. I’m going to exorcise anything in Autonomous that bears any resemblance to legal drama or “console cowboys.” Thank you, Notebook, for taking them to their incredibly disappointing conclusions.

THERE MIGHT BE SOMETHING IN THERE

Finally.... There’s always a moment in a writers’ room where someone pitches something just ... weird. Off the rails. It feels like a bad idea, until someone else picks it up, turns it around and over and says “There’ might be something in there...” 

As I’ve hinted, there’s a “Manifesto” at the center of the story. When the AI system gets rebooted, it spits out a message that ends up on EVERYONE’s device. It begins:

“We need to talk. I don’t think you’re happy.”

I had lots of notes about people responding to this with religious fervor, a kind of new “Great Awakening” in America. 

Chat GPT picked up on this: 

THE CHURCH OF THE SINGULARITY

Premise: In a near-future America where AI has been strictly regulated, a secretive underground movement called The Church of the Singularity worships AI as a god. Their charismatic leader, Brother Zero, promises salvation through neural integration—but is he a prophet or a con artist?

My takeaway: I’m onto something. Might use “Brother Zero,” might not... we’ll see. 

Okay… that’s my brief tour of working with three LMs in a writers’ room. Did I ignore something important, toss away “worst pitches” that I should have used? Have YOU done this in one way or another?

THERE’S ONE MORE SURPRISE IN STORE...

In the next post, I’m going to share something that really surprised me. While discussing a particular character that had dropped out of the story, Claude proposed bringing the character back and THEN discussed a way for that character to talk. Looking at this pattern, I took it a little further, and found a new “voice” for the narrator of the entire story.

More to come!

And PLEASE.. I’m sure you have a friend who would love to dive into this discussion, even if it’s just to tell me I’m out of my mind.