The Cost of Generation Is Approaching Zero

For most of human history, the constraint on creative output was production.

To make a film, you needed film stock, cameras, a crew, months of editing. To publish writing, you needed a press, paper, distribution. To produce music, you needed a studio, instruments, session players, mastering time. The cost of making something was high enough that it served as a natural filter. Most ideas stayed ideas.

That constraint is dissolving.

Today you can generate a photorealistic image in seconds. A short film in minutes. A polished first draft in the time it takes to make coffee. The models are improving so quickly that the production cost for almost any creative artifact is approaching zero. Not reducing. Approaching zero.

This changes something fundamental.

When production was expensive, judgment was embedded in the cost. The decision to spend $200,000 making a film was itself a judgment: someone believed enough in the idea to commit resources. The scarcity forced a kind of triage. You couldn't produce everything, so you had to decide what deserved to exist.

Now that production is nearly free, judgment becomes the scarce resource. Not removed from the equation — more central to it.

Anyone can generate a hundred variations of an image. The question is which one is right. Anyone can produce a draft. The question is whether it says something worth saying. Anyone can create a piece of music. The question is whether it moves anyone.

The models do not know what you are trying to say. They do not know who you are talking to, what you believe, what matters, or what the world needs next. They are extraordinarily capable instruments that have no idea why you are playing them.

That's not a limitation to be solved. That's the permanent condition.

What becomes valuable in an age of abundant generation is not the ability to produce — it's the ability to direct. To know what to make before you make it. To recognise the right output among the many possible outputs. To understand what the work is for.

These are not technical skills. They are human ones. And they are not getting cheaper.

In fact, the opposite may be true. As generation floods every creative domain with plausible outputs, the rarity of genuine judgment increases. The person who can tell the difference between something that works and something that merely appears to work becomes more valuable, not less.

I think this is underappreciated.

Most of the current conversation about AI and creativity focuses on the tools: which model, which workflow, which prompt. That conversation will become increasingly commoditised. The tools will improve and standardise. Access will broaden. Generation will become infrastructure, like compute or bandwidth.

What won't become infrastructure is taste. What won't become infrastructure is knowing what the work should do in the world and whether it did it.

The shift is from production to curation. From making to deciding what to make. From execution to judgment.

This is not an argument for passivity. Making is still essential — you can't judge what you haven't made, and you develop taste through production, not instead of it. But the bottleneck has moved. The thing that determines whether creative work succeeds in an age of abundance is not how much you can produce. It's how well you can see.

Judgment, it turns out, doesn't scale. It doesn't get cheaper with better hardware. It doesn't improve with fine-tuning. It remains stubbornly local: a person, a perspective, a set of values, a capacity to tell what's working from what isn't.

That's the thing that AI is not replacing. That's the thing that becomes more important as everything else gets automated.

The cost of generation is approaching zero.

Judgment isn't.