The Economics of Sameness

The biggest cultural risk we face right now isn’t decline. It’s convergence. Across our high streets, our streaming platforms, even our ideas, everything is starting to look the same - not because we lack creativity, but because we’ve become exceptionally good at optimising for what works.

Open Netflix and the same logic applies. The superhero film/series isn't just a genre anymore - it's an industry within an industry, a hedge against uncertainty at a time when the cost of failure could be catastrophic. Studios spent decades learning that a recognisable IP, a bankable franchise, an established audience, is the closest thing to a guaranteed return in an inherently unpredictable business. The result? Films that feel like they were designed by committee - because they were.

This is what happens when cost drives culture. When the stakes are high enough, risk becomes the enemy. And risk is where everything interesting lives. 

Here's the question I keep chewing over: if AI brings the cost of production down dramatically, does the creative risk equation change?

Think about what it currently takes to make a great film. A great script. A great idea. Great performances. Then the technical artistry - cinematography, lighting, sound design, editing - each discipline requiring its own team of specialists. Film has always been a collaborative art form of enormous complexity and cost. That complexity is precisely why so few people get to make one.

But here's an analogy that intrigues me. A composer writing a symphony does so largely alone. The work exists entirely in one person's head - a fully formed artistic vision - until it is transcribed and ultimately handed to a concert orchestra for performance. The composer doesn't need the orchestra to exist creatively. The orchestra is the execution.

Could AI become the orchestra? Could a genuinely visionary creator - someone with a singular idea, a distinctive worldview, an irreducible perspective on what they want to say - essentially prompt a film into life? Not prompt it in the sense of simply typing instructions into a chatbox, but in the sense that the vision is entirely theirs, and the AI is the instrument through which it is realised?

In my opinion this is certainly going to happen – and the work could be extraordinary. There are already creative visionaries on Instagram posting work that signals this future. And, as a comic fan, I have always especially admired the creators who could both write and draw their visions: Frank Miller (Sin City), Hergé (TinTin), Jean Giraud/ Mœbius (Le Garage Hermétique), Eiichiro Oda (One Piece) etc. I think history will show them to be another precursor to what comes next. Whatever the medium, whatever the tools – great art will always require remarkable human vision, creativity and skill.

But here's the counterweight. In music, we've already watched what the democratisation of production tools does. GarageBand, Logic, then streaming, then social media distribution. The barriers fell, and yes, some genuinely brilliant work emerged that would never have existed inside the old system, but the vast majority of what the algorithmic floodgates released is polished, competent, and ultimately empty. Technically accomplished and creatively inert.

Back when I was Head of Music at BBC Radio 1 (2005-2015) we faced the 'Indie landfill' situation when a wave of indistinguishable guitar bands flooded the market - same template, same energy, diminishing returns, the sound of an idea being run into the ground. But indie landfill was a genre-specific problem. What we're dealing with now is categorically different: landfill across every genre simultaneously. Landfill pop. Landfill hip-hop. Landfill lo-fi. Landfill country. The streaming algorithm rewards familiarity, production tools have made replication cheaper than ever, and the result is a kind of genre-wide convergence that would have seemed impossible twenty years ago. The floodgates haven't just opened. They've been removed entirely.

I suspect the democratisation of film will follow the same curve. A handful of works of genuine genius. An ocean of ‘meh’ content that has been generated to spec. 

The homogeneity of AI-generated content isn't a future concern. It's already here, and it's spreading.

Scroll through LinkedIn for ten minutes. You'll start to notice the pattern. An opening hook designed to make you pause. A short, punchy personal observation. Then the pivot to a broader lesson. Then the list. Then the call to action. Then the engagement bait in the final line:

Agree? What do you think?

Someone had an interesting thought. They fed it into ChatGPT. ChatGPT has been trained on what performs well on LinkedIn, so it generated something that performs well on LinkedIn. The human insight - often genuinely valuable to be fair - has been processed into a format optimised for reach, and in the process, something essential has been lost. The voice. The roughness. The sense that a person is actually thinking.

This is what optimisation does to culture. It smooths it out. It removes the friction that makes things interesting.

There's a principle that operates quietly beneath all of this: scarcity breeds demand.

When vinyl records were the only way to hear recorded music, they mattered. Going back even further, when live performance was the only way to experience music at all, the event of it was irreplaceable. The more accessible something becomes, the more its perceived value tends to erode - not its intrinsic value, but our relationship with it.

And then there's the question of effort - which I think is underappreciated in these conversations. The effort required to achieve something is not incidental to the achievement. It is a fundamental part of its meaning. Walking Alfred Wainwright’s Coast to Coast across the north of England - 192 miles from the Irish Sea to the North Sea - is not something you do for the destination. The destination is The Bay Hotel, a pub in Robin Hood's Bay, which you could reach in two hours by car. You walk it because the walking is the point. The blisters, the accumulated exhaustion and exhilaration of two weeks on your feet. Take that away and you haven't found a more efficient route. You've lost the entire experience.

If a creator can generate a film, or an album, or a novel in an afternoon, what does that do to how we experience it? What does it do to how the creator experiences making it? I don't have an answer right now, but I think the question matters more than most people are currently willing to acknowledge. 

Something interesting is already happening in music. Classic artists and catalogue recordings are acquiring a new kind of value - not just nostalgia, but authenticity. The knowledge that a piece of music was made by human beings, in a room, with all the limitations and accidents and choices that implies, is beginning to feel like a distinguishing characteristic rather than a historical footnote.

This will intensify. As AI-generated music becomes ubiquitous, and it will, the cultural weight of pre-AI work will grow. Not all of it. Great music will remain great music, but the origin story of art is becoming part of its meaning in a way it never previously needed to be. 

And then, occasionally, something comes along that reminds you what the alternative looks like.

Angine de Poitrine are everything the homogeneity thesis says is disappearing. Spiky. Difficult. Interesting. Funny. Absurd. Genuinely original in a landscape where genuine originality has become genuinely rare. Their image is startling in the way that almost nothing is startling anymore - because startling requires risk, and risk requires not caring whether everyone gets it. They have their own symbols, their own invented language, costumes that make you look, smile and look again. This is 21st century Dadaism: art as provocation, identity as performance, the whole endeavour built to stand out rather than fit in, to inspire conversation rather than passive consumption. You can check them out here: https://www.youtube.com/watch?v=0Ssi-9wS1so

Will they become a mass-market proposition? Almost certainly not, but that is not a criticism - it is precisely the point. Not everything is supposed to scale. Some things are supposed to exist at the edges, to keep the edges alive, to remind the centre that there is something beyond it.

What matters is this: people are hungry for them. In a landscape of the optimised and the adequate, something genuinely strange and specific creates a gravitational pull that no algorithm predicted and no committee approved. Scarcity breeds demand - and in a world drowning in content, a band that has built a world, invented a language, and committed entirely to a vision they couldn't fully explain to a marketing department is scarcity itself.

We need more left turns. Not because every left turn leads somewhere remarkable -most of them don't. But because a culture in which everyone keeps going straight eventually forgets that left turns were ever an option. 

Which brings us to the platforms, the DSPs. Where should Spotify, Apple Music, and Amazon draw the line on hosting AI-generated music?

It's not a simple question, and simple answers don't serve it well. An outright ban on AI music is neither workable nor, perhaps, desirable - the line between AI-assisted and AI-generated is increasingly impossible to draw, and some of what emerges is genuinely worth hearing. But unlimited hosting of unlimited AI content, with no distinction, no labelling, no curation logic that accounts for provenance - that risks drowning everything in a sea of the optimised and the generic.

Spotify's SongDNA initiative (https://newsroom.spotify.com/2026-03-24/songdna-announcement-beta/) is interesting in this context - an attempt to map the genetic human fingerprint of what makes music connect. Whether it points toward better discovery or better homogeneity depends entirely on what questions you ask of the data. Used well, it can be the tool that surfaces the genuinely distinctive in an ocean of the adequate. Used poorly, it's just another optimisation engine - and we've got plenty of those already. Personally, I love the feature as it opens endless new rabbit holes for discovery.

There's a parallel in urban planning. We have building regulations that shape the character of our streets. I support these controls, but should there be the equivalent for shops on the high street? Planning consent that enforces the diversity of what opens and what stays? It sounds heavy-handed, and perhaps it is, but we are already asking ourselves a version of the same question in every creative industry simultaneously, and the answers we settle on will shape what culture looks like for a generation.

The threat of homogeneity is not that everything becomes bad. It's that everything becomes fine.

Competent. Polished. Inoffensive. Adequate. Nice.

The nail bars and fried chicken joints are fine. The superhero films are fine. The LinkedIn posts are fine. The AI-generated playlist fillers are fine. But fine is forgettable and I want more.

The things that change how people hear music, see the world, understand themselves - those things are rarely fine. They are strange, specific, difficult, expensive, risky. They come from someone who cared too much about something too particular to make it safe.

That person still exists. The question is whether the system we're building gives them a chance - or optimises them out.