AI is the existential threat to screenwriters, artists, musicians, whilst giving you, the human, all the rights to the produced work! You needn’t credit any of the seemingly endless references. ChatGBT can likely write a better version of this blog post, write the tweet to promote it, build a community, find a stronger voice, and live my best life for me.
Yet… I’m a supporter because much like the advent of other giant technological leaps in the past, you either adapt or perish. I don’t want to perish into an oblivion darker or lighter than I currently feel like I’m sinking into. So I embrace. My first direct foray into AI came at the hands of an AI writing company called Sudowrite. At the top of the year, one of my closest friends introduced me to the CEO and cofounder of the company. I had a pleasant coffee with the CEO and we discussed screenwriting challenges I struggled with, as well as my experiences as an assistant to a showrunner. I had witnessed writing from many points of view and had a well rounded idea of what we bang our heads against. We hate research, claim we hate story structure and save the cat, wish there was a database where we could describe a scene and a query of movies pop up displaying all the results that we could reference and steal from. Perhaps Sudowrite could solve all of these needs. Our coffee lasted hours and we developed a quick kinship, even resulting in walking to his airbnb to meet his partner and their dog. We would stay in touch and eventually he would demo new products and invite my feedback. I couldn’t believe how powerful some of the AI tools were for generating new ideas. I could spit in a premise and the program would defecate so many options of where the story could go. The results were only as good as the premise, in my opinion at least. So they weren’t perfect, but as a brainstorming tool, I never thought AI was already that close to creating something usable. Up until that point, I had only laughed at memes where AI would attempt to write a Tarantino style script. It was nonsense. But to go from that script, a pastiche of ideas, the middle school version of me trying to act and sound like Lex Fridman, to what Sudowrite was ideating… I was impressed to say the least.
Eventually, Sudowrite hired a young, strapping product designer named Ryan, who led my demos thereafter. This man was clearly a talented and eager designer but as we got to know one another, I learned that he, as well the founders, were all fiction writers. So they genuinely were trying to create the tools to fuel their own idea generation. Our relationship developed, and we shared with one another our value systems and an affinity for tea. Some short time passed and then he hit me with a pitch: he would come down to Los Angeles and be my writers assistant for a week for research on the product. He asked if I was up for it, as if I had something to lose. I loved the idea, I could only see upside. So we arranged the dates, had a precall to discuss what I’d have him do, he flew down and we were off to the races. In hindsight, I underestimated the duties I had prepared for him and overestimated the speed at which we could operate. I wanted to be ambitious with the week we had, foolishly believing we could pivot from one project to the next, from outline to outline. What I should have realized, especially from past experiences, is going from one project to another can feel like going from one planet to another. I’ll dive into specific projects in a future paywalled post, but what Ryan was able to help me with was a lot of research, creating a virtual cork board on figma, ideation, and overall he shared a very different perspective on all of my ideas. Very quickly Ryan understood that if his product could generate a targeted research query, it would save writers and their assistants so much time. Sudowrite already had tools to distill your written work and spit out loglines, summarize beats, generate potential future ideas. When he himself found the work difficult to grasp, or if we were both stuck, he’d work on refining these tools so we could use them to generate new ideas, tailored to what I needed. When I was struggling to see the big picture, the tools would summarize my ideas and spit them at me in a different way with hope that it would inspire a look at the target from a different angle. I didn’t like any of the ideas but the fact that this machine could already do this never ceased to amaze.
Sudowrite does have a ton of competition in AI writing tools. And as I write this, ChatGBT has quickly surpassed them all. I felt a sense of purpose slipping during the middle of the week. I was never a naysayer, but I did feel that threat amplified because I was failing to grasp the tools handed to me. I felt like a three year old stubbing my thump with a hammer for the first time and no longer trusting the hammer. Meanwhile, I imagined the other kids in day care thriving with not only hammers but also screwdrivers :(. I saw a future where producers would use these tools to remove me, the writer, from the equation. But also an actor’s voice, their likeness, soundtracks, picture editing, the list and the roles go on. We could all be replaced. Even the director’s duties are sure to change. They would be the guardrails, vessels feeding data to the AI to generate their vision. Writing has often been compared to sculpting, in chiseling away from a block what does not belong to create beauty from what remains. A post-AI director will do just that, listing prompts to divert ideas away from certain arenas and herding digital sheep into the fences of their imagination until they have their David.
If you are a product of your environment, then we were two very different products. Many of my discussions with Ryan converged at discussing the tech vs film industries and about trauma with past romantic relationships. Before eyes are rolled, the specific project demanded some of these discussions and this is also not uncommon in a traditional writers room! Writers rooms are therapy, scripts are bleeding onto the page. Ryan was beyond helpful in forcing me to attack certain problems but there were major pitfalls that made me examine my own process. He provided so many frameworks for me to fill in, such as character details, themes, episode arcs, and it was up to me to stamp my view on them. What I said to him after a day of brainstorming and organizing ideas was that I would do the work once we separated and I had some time alone to think. I thought I would break the pilot plot this way, thought I’d come up with characters this way, but I’d come short, exhausted after a long day. Unfortunately, this part wasn’t unfamiliar to my process and reminded me of the days of working with my former writing partner. Ryan is smart and eventually he caught on and the second half of the week he called me out telling me he need to be more of a project manager for me. I did take solace in remembering this was a similar conversation that I once had with my former boss, when I was the assistant. How frustrating it would be when my boss would change their mind, erase what had been on the board, pivot in a direction that would undo the entire foundation that was agreed upon. And eventually I would have to retreat from the creative partner to the one who holds the whip. Realizing this could be a pattern in all assistant dynamics, I empathized with myself and better understood my bosses of the past, as well as questioned whether Ryan could ever program his AI to shift gears on a dime like he was able to because of his EQ. I saw on linkedin of all places my former employer describing AI as “augmented intuition” whatever that means. Perhaps clever marketing. Ryan’s ability to change his approach was absolutely his intuitive read on where I was struggling. But could he have seen it from the words on a page alone? Or were they all social cues, compounding discussions about our elusive lovers, and shared views on mental health stigmas that led him to fit my needs?
We spent our final day reflecting on it all. I wanted so badly to see the report he’d end up writing for his higher ups and I admitted that if we were to do this again, I’d better structure our days with clearer benchmarks and promise to focus on only one project. He asked what he could do better and what feedback I’d have for him. I told Ryan that beyond having someone take notes for our sometimes brilliant conversation, he was a great assistant. AI has a long ways to go before it can be a Ryan. He adapted better than I ever could and he delivered on everything that he ever promised me. Besides a look into the process of a writer like me, I could not say I delivered the same.