Crossed Wires: Hollywood — dying voices on the wind

On 6 August a feature movie called The Cully Hill Boys was released. I watched about 10 minutes of it online. It is not really my genre (London underworld/violent/rap/drug story), but there are plenty of these sorts of films out there. Some are very good, including Netflix’s 2019 hit Top Boy which was revived from an earlier UK Channel 5 release. Others are less so, but even so there seems to be an insatiable audience for them – mainly teenage boys and young men.

The Cully Hill Boys is one of these. It was produced by a film and video company called Higgsfield in the San Francisco area. The screenplay was written by Timothy Planagan, who had worked in independent film and had been shopping this script for years. The two lead directors are Aitore Zholdaskali and Adilet Abish, both filmmakers from Kazakhstan.

I am pretty sure that no one who reads this will have seen this movie. It is not on any film circuit and can only be found online.

But it is turning out to be one of the biggest movie stories of the year. The news is not that it was made with AI assistance (that is already old hat) but that this is an entirely different kettle of fish from the other big 2026 AI-film story, which saw the Iranian-themed film Dreams of Violets admitted into the Tribeca Film Festival, of which I have written previously.

First, some facts, which set up a story of the coming upheaval of a legacy industry.

The 110-minute film was made in four weeks (as compared to an industry average of one to two years). It was made for $2-million, almost all of it in compute costs (as compared to an industry average of many tens of millions of dollars for this sort of genre). It had a crew complement of 28 (as compared to an industry average of hundreds).

But it is not the startling production economics that are the big story.

It is where and how the human talent has been redistributed.

About half of the crew had real film experience, in addition to the writer and directors. The Higgsfield filmmaking platform (built on top of the leading Chinese text-to-video generator Seedance 2.5) has been designed specifically for the top creative team – the director, cinematographer, art director.

Filmmakers do not think in prompts – they think in cameras, lenses, lighting, movement and storytelling. Higgsfield shifts the user away from writing adjectives such as “cinematic” or “filmic” into prompts and towards making the kinds of creative decisions a director or cinematographer would make. Users select virtual camera bodies such as an Arri Alexa, choose lenses, focal lengths and camera movements and film-style lighting setups, and construct scenes as though they were working on a physical set. The “language” of the production process is immediately familiar to traditional creative filmmaking teams, and the same creative decisions remain in the hands of the director and cinematographer, with no new language required.

Key breakthrough

Many companies can now generate video, but Higgsfield attempts to build the environment in which an entire movie is planned and produced. Instead of asking AI to create a clip, the filmmaker creates reusable assets – characters, locations and props – which persist throughout the production. Once these assets have been established, they can be reused repeatedly, allowing visual consistency from one shot to the next. This is presented as the key breakthrough that transforms AI video from a novelty into a filmmaking tool.

This is different from other emerging AI visual workflows. Traditional text-to-video systems require every scene to be regenerated independently, making continuity almost impossible. Higgsfield instead treats filmmaking as a sequence of connected shots built from a common asset library. Characters retain their appearance, locations remain recognisable, and camera angles become deliberate creative choices rather than random outcomes.

And so, the creative instincts of the director and his team are facilitated as seamlessly (if not more) than if they were on a live set.

I have left the real shocker to last. The actors.

The cast is, in fact, a collection of simulacra of real people, including minor celebrities. They include Israel Adesanya (former UFC middleweight champion), Quinton “Rampage” Jackson (MMA veteran and former UFC champion), N3on (Mikyle Rafiq, popular Kick streamer) and Matt “MK” Kiatipis (streetball creator and athlete).

These people (well-known to their young masculine audience) signed over IP rights to their “likeness” and voices. They apparently never even visited the production studios (except to be scanned). The AI actors are startlingly similar to their real-life counterparts, presumably close enough for their fans who see this film not to notice or care.

There is a great deal of money to be made here – the economics of licensing “likeness” must certainly be seductive for any online influencer or someone who was once well-known IRL and is now seeing their star fade, or just about anyone recognisable seeking an additional income stream.

We should pause here. An actor is a vehicle to carry a writer’s dialogue and dramatic intent, and the director’s interpretation. If this can be executed authentically via some other approach than flesh and blood, why would an audience care (and how would they even know)? Is it not their prerogative to judge its impact on themselves, provenance notwithstanding?

Completely believable

My short viewing left me a little wobbly. Had I not known, there was no way that I could have told it was AI-generated (and I have considerable expertise in digital special effects) – the real-life fidelity of the digital images was completely believable. It was a glimpse of an imminent and somewhat disturbing future. Those of us born and bred and nurtured alongside the great art of filmmaking and the larger-than-life characters that burnished it now face the prospect of an onslaught of something else entirely. A new kind of film, acted by lookalikes and imaginaries, written and helmed by real and perhaps even exceptional human talent.

Is the movie any good, at least the piece that I watched? No, it is a bit wooden and a little silly, as many subgenre B-movies are. But a 15-year-old boy might disagree with me, and he too deserves his say. I must also admit that a better script or different story might well have landed differently with me.

There is a final matter here. We watch visual content in theatres or on monitors or smartphone screens. There is an entire economy that carries the content from concept to our eyes and ears, and the costs and risks are well understood by the businesses that finance them. When the economics change this radically (an order of magnitude faster and cheaper than traditional methods within the space of a year or so), there is no way to really predict the industry deformations that will appear in its wake. But it will not be trivial. 

If you were a large movie studio dependent on a few big hits for profit, but then were offered many small low-risk, low-cost mini-hits for the same cumulative profit, what would you choose? How would you re-evaluate the business risk? Would you ever fork out $250-million for a big-budget sci-fi movie again?

The bigger part of me screams – no, I won’t watch these, ever. It is an insult verging on apostasy. But the truth is that soon no one will care what I think. Or my filmmaker friends (of whom I have many, both artists and artisans), schooled in passion and craft. Or what the unions want or demand. Or what Leonardo DiCaprio or Martin Scorsese or Aaron Sorkin think. There is a new generation of eager storytellers and audiences untroubled by legacies and history, and they will colonise the land.

They will make and watch these new films, and they will construct the yardsticks, and they will decide on their artistic merit, and our outrage will be as dying voices on the wind.

(If you are old-school and troubled by the prospect of AI-films diluting a once-proud art form, do not visit the Higgsfield website. It will ruin your day.)  DM

Steven Boykey Sidley is a professor of practice (ex-JBS, University of Johannesburg), a partner at Bridge Capital and a columnist-at-large at Daily Maverick, where he writes the weekly Crossed Wires column. His new book, It’s Mine: How the Crypto Industry is Redefining Ownership, is published by Maverick451 in South Africa and Legend Times Group in the UK/EU, available now.