Author's Note #1: I misspelled Don Allen Stevenson III’s last name as “Stephenson” in Mondays’s essay in a handful of instances. Those errors have been corrected.
Author's Note #2: A reminder that I am running a survey about the seven entrepreneurs or "builders" I interviewed this year, all of whom sit at the intersection of Artificial Intelligence (AI) and media.
What else would you want to learn about them in a new offering? And in which other formats?
There have been two conflicting themes in my essays over the past five years: The exciting possibilities of technology and the “hard knocks” reality of constraints (market, legal, technological).
When I launched this subscription newsletter back in 2019, legacy media companies were boldly promising investors that their streaming services could reach hundreds of millions of subscribers worldwide.
I was skeptical.
I had learned as an executive within Viacom (now Paramount Global and soon to be “TBD”) that the older generations of TV and movie executives did not understand the emerging technology or retail business models of direct-to-consumer. The extraordinary profits from the cable business model—which effectively subsidized their media businesses—incentivized everyone to invest in their digital futures suboptimally.
Over the past five years, these executives have learned the “hard knocks” consequences of suboptimal investments in technology and direct-to-consumer business models. High churn rates suggest the demand for their content libraries is weaker than their business models require. Most are struggling to scale and/or to be profitable.
Even Netflix faces a tougher march to 500 million subscribers. All are losing consumer attention and engagement to YouTube on TVs across the U.S..
Today, the seeds of a new five-year dance between possibility and constraints lie in emerging generative artificial intelligence (AI) tools. Over the past year, I have written about the possibilities of generative AI through the eyes of entrepreneurial “builders” like Jordi van den Bussche aka "Kwebbelkop", Fable Simulation’s Edward Saatchi and Don Allen Stevenson III aka "Don Allen III".
Generative AI technology is in a primitive phase where its use cases are being tested and proven out. So, there seem to be exponential possibilities and very few constraints. The question is how to think about what the constraints might be. That is what makes Marshall McLuhan’s work arguably more valuable than ever.
Key Takeaway
"Builders” like Jordi van den Bussche (Kwebbelkop), Fable Simulation’s Edward Saatchi and Don Allen Stevenson III are pushing past the traditional constraints of traditional media to identify new dynamics of storytelling within the medium of generative AI.
Total words: 1,400
Total time reading: 6 minutes
Three Problems With McLuhan
There are three fundamental problems with recommending Marshall McLuhan in 2024.
He was a philosopher;
Who was writing for an audience of academics (in “Understanding Media” he frequently cites the unreadable modern novel “Finnegan’s Wake” by James Joyce); and,
Unlike Clayton Christensen, he never utilized business case studies to support his arguments.
In short, his writing and argumentation can be dense and opaque. Anyone citing him beyond his adage “the medium is the message” may not necessarily understand what he wrote (and that was the gist of the joke behind McLuhan’s famous cameo in Woody Allen’s “Annie Hall”).
A more accessible version of his writing can be found in “The Medium is the Massage: An Inventory of Effects”, a 160-page book published in 1966. It is composed in an experimental, collage style by graphic designer Quentin Fiore (and with a recently redesigned cover by the artist Shepard Fairey).
That book offers a simpler, more digestible outline of his core argument: Mediums—or processes—matter more than the core content delivered within them. It argues that the key to understanding the possibility of generative AI business models is to identify the “invisible ‘groundrules [sic], pervasive structures and overall patterns” that work in the background of mediums.
Whereas, a focus on the various, emerging content formats with generative AI distracts us from identifying those constraints. McLuhan writes “the ‘content’ of a medium is like the juicy piece of meat carried by the burglar to distract the watchdog of the mind.”
In 1967, television was “a totally new technology which demand[ed] different sensory responses” from audiences. In that framing, generative AI in 2024 seems to be television’s logical replacement in that sentence.
However, unlike television—constrained by radio waves (broadcast), fiber optic cable installation, (cable) and television manufacturer capabilities—the constraints and processing of generative AI are less familiar to us in large part because they are still emerging.
Possibility
This is why business models like Kwebbelkop AI, Showrunner or Stevenson’s consulting model seem more interesting than business models that leverage AI tools to make Hollywood production models faster and cheaper. They actively seek those constraints and then iterate accordingly.
Whereas van den Bussche, Saatchi and Stevenson seem focused on the possibilities for storytelling created by generative AI beyond movies and TV series. None are betting on using an arsenal of AI tools to produce the volume and the quality of content that previously was only expected of much larger studios.”movies or TV series as the final output to emerge from generative AI.
Promise Advanced Imagination, an AI startup that emerged last week, is focused on the more familiar outputs of movies and TV series. It is focused on “using an arsenal of AI tools to produce the volume and the quality of content that previously was only expected of much larger studios.". There are possibilities in that model, but they are more technical, tactical and narrowly focused on building new workflows for studios “from the ground up”.
Van den Bussche is hacking his way into building a suite of AI tools that generate every aspect of a YouTube video with his “Kwebbelkop’ persona: “audio, video, person, voice, concept, everything” as he told me in a recent conversation. He started down this path after suffering what he called “burnout” from posting videos almost every day for 10 years.
The interactive storytelling on Showrunner starts with the idea of the consumer—via text-to-video prompts—and then leverages generative AI to create an episode of a show. However, its founder Saatcxhi is not focused only on TV episodes as the final output. His company Fable Simulation recently shared research on SIM-1, “a framework for AI wargames that powered an AI simulation in Sim Francisco”. Fable recently released a simulation in Sim Francisco of the 5 days from when Sam Altman left OpenAI to when he returned. It also released a research paper on Showrunner Agents in Multi-Agent Simulations.
As for Stevenson, he told me "In the past, I would always kind of tailor my ideas off of how much time do I think I have to execute on that idea? If I started to get too creative with the idea, I would just pull it back and say, ‘You know what? Let's not go down that route because we won't have time if that's the wrong route.’”.
Now, with generative AI tools, he can explore “super complicated” concepts without punishment:
“And if it's the wrong direction an hour from now, we'll change directions. And that creative workflow, that creative pipeline is now one that's far more forgiving to the individual. You can explore without punishment. In the past, if you went to explore, there's a major risk. If you're wrong on your exploration, you've just wasted so many resources in trying to get to that idea that was wrong.”
He produces short Instagram videos, and as TV shows or movies, as the final outputs, and monetizes them with other business models: Instagram subscriptions, exclusive content for “luxury audiences”, and consulting for technology companies). He is using traditional media to bypass the traditional economics of content production.
After TV, Movies, YouTube, Instagram…
All three entrepreneurs are pushing past the traditional constraints of traditional media. The question is how to identify the obstacles to all this possibility. Only van den Bussche and Stevenson are actively monetizing content produced with generative AI (van den Bussche via ads and sponsorships).
Through the lens of McLuhan, the constraint that seems easiest to identify is consumer demand. McLuhan wrote that in 1967, what television’s audiences understood and television’s critics failed to understand is that “a totally new technology [...] demands different sensorial experiences. Van den Bussche’s fans rejected his previous attempt at an AI Kwebbelkop last Fall, and he suspended the earlier version of the project last December. He told me he will be announcing his next experiment in AI shortly.
Stevenson is more interesting because he has found audience demand but the majority are not paying to watch his content. Instead, they are paying to watch “luxury” exclusive content or for videos to inspire developers, as customers like Apple, Adobe, Snap and Nvidia do.
To paraphrase a point from McLuhan in “Understanding Media”, generative AI technology is “within the gates”, and “we are numb, deaf, blind, and mute” to its encounters with both the demand side and supply side of the media business. The sense is that by exploring the possibilities of storytelling, these entrepreneurs are discovering new dynamics of storytelling made possible by the medium of generative AI.
The danger with possibility is that it implies there are no constraints. It is in following “builders” more closely that we get early, valuable signals for what those new constraints around storytelling are and will be in the age of generative AI.