This post was originally published on this site.
Erik Barmack writes every other Tue. for paid subscribers. He recently interviewed writer-director David S. Goyer on his new AI-powered startup; explained why 2025 will be the year of the full-length AI movie, and assessed how AI storyboarding is transforming pitch meetings.
“If anyone’s job should be replaced by AI, it should be the execs,” actor-writer-comedian Joel Kim Booster (Fire Island) said during the 2023 strikes. “In terms of the skillsets, I think it aligns a little bit more with what AI is able to do in terms of just sort of distilling algorithms and what can be bought and sold and things like that.”
Booster was right, but collectively we settled back into worrying about AI upending creative fields from voice work to making feature-length movies itself. (I still think we shouldn’t sleep on humanoid robots coming for the physical labor of production.)
Over the last few months, though, the emergence of a new set of products from the major large language model (LLM) companies — OpenAI, Google and Anthropic — started to excite people inside the AI world about the type of work AI could do. These “reasoning engines” are designed to think the way humans do by breaking down complex problems into steps and coming to logical conclusions.
The entire world had to pay attention to these reasoning engines when the Chinese company DeepSeek crashed the economy for one day late last month thanks to its R1 model being as good as OpenAI’s o1 while being more cost effective. These new models have been touted as being able to solve PhD-level science and math questions. They’re able to do the work of an expert computer programmer. They’re putting us on the path to so-called superintelligence and have techies worried about their careers.
So can they replace . . . me?
If Larry Ellison and Sam Altman say that these AI tools can solve age-old medical problems — and even lead to a cancer vaccine — then surely they can solve a problem like how to thrive in Hollywood as an independent producer. Or develop a U.S. distribution strategy for a Mexican movie. For all these AI companies to survive and live up to their lofty expectations (and valuations), they have to solve business problems so regular companies will buy their products. With the chatter around these new reasoning engines ascendent, I decided to put them to the test.
George Orwell wrote about all animals being equal, but some animals being more equal than others. I’m fine with the idea of AI upending Hollywood, so long as it doesn’t upend me first.
In this issue, you’ll learn:
-
How the promise of AI reasoning could inspire Hollywood’s own DOGE-like efficiency movement
-
How DeepSeek will accelerate studios’ AI adoption even if they never use DeepSeek itself
-
Why your next coworker may be an AI copilot
-
Where execs tasked with adopting AI will save time — and where they won’t
-
What the industry will lose by adopting AI for strategic analysis too early
-
Which AI engine — OpenAI, Google, Anthropic or DeepSeek — is best at the business of Hollywood
-
The challenge in asking AI for new business ideas
-
Where AI reasoning was accurate but still wrong
Generative AI may produce impressive (or at least intriguing) video and image results, but AI is fundamentally a tool for processing large swaths of data and recognizing patterns. In other words, in theory, it’s been trained just like an exec, whose job is to analyze trends, find market opportunities, develop distribution strategies based on past data, and of course, put all of this information into easily digestible bullet points. If you’ve ever used any AI chatbot, you know they can’t resist putting its findings into bullet points.
I know I complain about my job, but it’s certainly easier than curing cancer.