A Scribbler’s Take on AI Creativity and Productivity
Throughout my career as a TV writer and producer, I've always thrived in the collaborative environment of a writer's room. The banter, the clash of ideas, and the energy in those rooms are irreplaceable. But as those daily interactions became less frequent post-Covid, I found myself searching for a new way to brainstorm, riff ideas, and get takes on my stories without the immediate social pool of fellow scribblers. Enter AI.
My LLM Awakening
After spending my first couple of hours using Chat GPT, I had an epiphany that much of my career had involved operating similarly to a large language model (LLM) for the showrunner. They would come into the room with a general idea, and the writers would build out the specifics and generate drafts, which the showrunner would then rework. Recognizing this helped me get over my trepidation. I realized AI wasn't just an inevitable threat; it was a potential tool that could amplify my creative process.
A Tool in the Box
Over the past year and a half, I’ve been experimenting with ChatGPT, Claude, MidJourney, Gemini, Eleven Labs, and others. They’ve become an essential part of my workflow. These tools don’t just mimic collaboration—they bring their own kind of haecceity, offering a fresh perspective on my writing.
After I input my scribbles, LLMs help me brainstorm ideas, identify themes, and refine my content. While AI hasn’t yet fully executed a final product that I love, I see a future where it might. I look forward to the day I can ask an LLM to generate a sequel to Shane Black’s The Nice Guys, solely for my own entertainment.
My Research Intern
On the scribbling tip, LLMs have also become invaluable research assistants. Pulling together the specific info I need for my writing could take hours, sometimes days, to track down. AI streamlines that process. While the information it provides isn’t always correct, I run it through a multi-step verification process to ensure accuracy. With a bit of work, I usually get it far enough that I can add the data to my imagination.
Adventures in Screenplay-Adaptation
I tried using Claude to turn an unsold screenplay into prose that I could potentially publish as a novella. It wasn’t a one-click wonder. The process was a grind of back and forth, and calling the final result less than ideal would be kind. But when I ran the content through Eleven Labs text-to-speech, the few chapters generated in audiobook form sounded kind of cool.
The experiment taught me a lot about the differences between writing for the screen and for the airport turnstile. So while the final product wasn’t something I’d rush to publish, the process was an accelerant for my learning curve.
The ADHD Whisperer
LLMs have this incredible ability to take my ADHD-influenced prose and make it more grokkable to those who aren’t living inside my head. It turns my headcanon into readable text for humans.
This post was forged from a series of disjointed rambles into the mobile version of ChatGPT while I waited to pick my son up from his job at the ice cream shop. AI platforms can take my garbled notes, organize and analyze them, and, with obsequious deference, make suggestions for clarity and adjustment.
Rapid-Fire Idea Testing
My off-the-charts ADHD keeps my brain spinning at high velocity, churning out more ideas than I know what to do with. LLMs help me whip up versions of these ideas, enabling me to see what’s good and what sucks. It gets me to the fail stage of development quickly. Then I can decide whether to iterate or discard.
AI-Powered Game Design
One area where AI has been particularly helpful for me is exploring tabletop game design. As a lover of skirmish games like Warhammer 40k, I’ve always wanted to create one of my own but was intimidated by the complexity of balancing the systems. My neurodiverse-induced inability to grasp numbers had kept me from diving in. But AI was able to alleviate that anxiety enough that I could experiment with the narrative elements that really interest me. Whether or not any of it works, I have no idea. But it’s been a fun and engaging hobby that brings me joy.
New Tools and Techniques
I’ve always seen parallels between video game design and TV writing. After experiencing the challenge of crafting serial stories for Lost and Alias, I adapted elements of agile software development for scribbling TV. On Heroes, we used a system that, while not exactly Scrum, involved similar principles. On American Gods season two, my writing staff was on an insanely tight schedule. Teamwork was essential, and agile techniques got the job done.
The Next-Gen AI Filmmakers
My favorite book, Neuromancer by William Gibson, has a quote that I’ve taken to heart: "The street finds its own uses for things." I see that spirit alive in my film school-bound son, who is using rudimentary AI tools to bring his scripts to life. He’s rotoscoping himself into drawings amidst photos of miniatures he’s kitbashed and art from his notebook. He’s also using tools from Runway to add insane digital elements to his iPhone shorts via MoCap. It’s a long way from my Super-8 filmmaker days, blowing up Playdough monsters with firecrackers, but the creativity and innovation are the same.
Embracing the Inevitable
If you’re curious about how you might incorporate LLMs into your process, I recommend the recent episode of the Think Like A Game Designer podcast. Host Justin Gary discusses ethical AI use in game design with Ethan Mollick, a professor at Wharton. Their experiences using AI tools to augment their design process and boost productivity mirror my own.
I never want to be the guy who tells the next generation to "get off my lawn." I’m always trying to activate my “beginner's brain,” seeking out new tools and ways of working to ensure my creativity stays fresh and relevant as I navigate the ever-challenging and competitive creative landscape.
