Bernard Marr's Future of Business & Technology Podcast

There are a handful of genuine process insights - the iterative, non-linear nature of AI filmmaking, the voice-casting workflow, and VO3 as a June 2024 inflection point - but the majority of the runtime is consumed by generic AI-displacement commentary, storytelling-is-human platitudes, and speculative future-gazing that any podcast listener will have heard repeatedly.

“we found somebody online who was, uh, a drone pilot. We, um, we talked to them, we took what they said, and we used ChatGPT to kind of help us structure it. So, uh, that saved us about three or four days.”

“VO came out. So this is a, you know, widely accessible tool made by Google VO3. And when it did, it was clear that this was a leap forward in terms of capability and in terms of fidelity and things like, you know, character consistency.”

The argument that the 2-hour film runtime is a technical accident of reel lengths - and therefore 'filmmaking' will cease to be a coherent category - is a genuinely interesting contrarian take; however, most other framing (AI-as-fire, electricity analogy, Napster analogy, Kurzweil singularity, fight-flight-freeze) is well-worn tech-discourse recycling.

“the idea of filmmaking comes from film, which the reason film was a, was films were two hours. It's not just some arbitrary number. It has to do with the fact that they have reels in cinemas that were that length”

“it's more like something like fire, which is more like a platform that enables other things to happen. So a tool is like a hammer.”

Two working directors with a combined 40 years across documentary, commercials, and music videos who have actually built and shipped AI-generated films using current tools - genuine practitioners rather than pure thought leaders - but they are not major industry figures and their track record cited in the transcript is thin.

“we're still coming to the table with collective 40 years of filmmaking experience, working with actors, experiencing, uh, different types of shoots”

“We worked with a member of Arcade Fire who, you know, you know, she wrote a song 10 years ago that I really like.”

The episode names real tools (VO3, MidJourney, ChatGPT), a specific project (the Ukrainian drone mission film), a specific collaborator (Arcade Fire member), and a rough time saving ('three or four days'), but large claims like '$200 million Hollywood film for 1/100th of the price' and 'record sales are 10% of the music economy' float completely unsupported, and most future predictions are pure speculation.

“we found somebody online who was, uh, a drone pilot. We, um, we talked to them, we took what they said, and we used ChatGPT to kind of help us structure it. So, uh, that saved us about three or four days.”

“we can make, you know, a uh, $200 million Hollywood level feature film for, you know, 1 1/100 of the price”

The host asks basic biographical and scene-setting questions, lets lengthy monologues run without substantive follow-up, and never challenges a single claim - including the '$200M film for 1/100th the price' assertion or the James Cameron/Stable Diffusion claim; the 'give me three words' closing and explicit sycophancy confirm this as a PR-style chat rather than a rigorous interview.

“I literally could carry on this conversation forever.”

“Give me three words. What will the future of filmmaking look like?”