Made, Not Generated

There is a particular kind of pleasure in holding something and knowing that another person made it. Not assembled it, not specified it into a CAD file, not prompted it - but made it. You can feel it before you even look closely: a slight warmth in the weight of it, an asymmetry so subtle you couldn’t measure it, a surface that tells you something about the hour of the day and the mood of the person who finished it.

These are not imperfections. They are evidence.

I have been thinking about this a great deal lately - about what it means to make something by hand, and what is lost when we stop valuing it. Not in the abstract, nostalgic sense. In a very precise one. Because the conversation about artificial intelligence and creativity is moving very quickly toward a question I think is the wrong one: whether AI can make art. The more interesting question - the one I keep returning to - is what we actually value when we value art.

If the answer is purely the surface, the finished image, the visual result - then yes, AI can produce something that resembles it. It can generate a painting in the manner of any artist who ever lived, a textile pattern of considerable sophistication, a photograph of a place that has never existed. It can do this faster and cheaper than any human.

But if what we value is the process - the decision-making, the struggle, the particular human consciousness that encountered a material and decided to do this with it rather than that - then we are talking about something that cannot be replicated. Not because the technology isn’t good enough yet. Because it has no hands. No afternoons. No memory of a grandmother’s tablecloth that makes you want to make something that lasts.

When I choose what to stock, I am not curating images. I am curating stories, and specifically the kind of stories that only happen when a person decides to make something by hand - which means deciding, failing, adjusting, and deciding again. The marks left by that process are not flaws. They are the record of something alive being involved.

Blue Light, Iron Salts, and One Particular Afternoon in France

Nothing I carry illustrates this better than the work of Alice Fraudeau.

Alice founded Joume Atelier in 2024. She works from France. She makes scarves - hand-made, one at a time - using the cyanotype process: a photographic technique that produces images in that deep, singular Prussian blue that you do not forget once you have seen it. Each piece takes days. No two pieces are the same. Not because the design changes, but because the process depends, in part, on something Alice cannot control at all.

The cyanotype process works by coating a surface with a solution of iron salts, laying a photographic negative or botanical element over it, and exposing it to ultraviolet light. Wash it in water and the light-exposed areas turn blue; the unexposed areas remain white. The resulting image is precise in one sense - the chemistry is exact, repeatable - and unrepeatable in another. The depth of the blue, the richness of it, depends entirely on the quality of the light on that particular day, at that particular hour, in that particular place.

Alice can control the formula. She can control the design, the materials, the process. She can mix the iron solution to the same ratio every time. She cannot control the sun.

Every scarf Joume makes carries, fixed into its shade of blue, the record of one afternoon in France - whether it was bright or overcast, whether a cloud passed during the exposure, how long the light lasted before it faded. The object holds that.

A Technology That Seemed Like Alchemy

The cyanotype process is not new. Sir John Herschel invented it in 1842 - the same year, as it happens, that Ada Lovelace was writing the first computer algorithm. The following year, a botanist named Anna Atkins used it to create Photographs of British Algae: Cyanotype Impressions: the first book ever illustrated with photographic images. She pressed specimens of seaweed and algae directly onto her prepared paper and let the sun do its work. The results were extraordinary - scientific documents of exact accuracy that were also, somehow, objects of ethereal beauty.

How did the world receive this? With considerable indifference. Hand-drawn engravings remained the standard for botanical illustration. Photography was still strange, still carrying the faint scent of alchemy. It was not considered a serious medium. Atkins was largely forgotten, and stayed forgotten for more than a century - until the 1980s, when art historian Larry Schaaf identified her as quite possibly the world’s first female photographer, and her prints began to be understood for what they are.

What I find instructive in this is not the injustice of her obscurity - though it was an injustice. It is the fact that the process itself, the chemistry, the iron and the light, were available to anyone. Herschel had published his method. The sun was no one’s. But the art was Atkins’. It was Atkins’ hands, and her particular way of seeing, that made the thing worth making.

What AI Can and Cannot Do

I am genuinely interested in artificial intelligence. I use it, I find parts of it useful, and I do not think the answer to any of this is fear or refusal.

But I think we are making a category error when we suggest that AI can replace the human creative act in the making of beautiful objects. And I think it matters that we get this right, because the error has consequences - for how we value the things we surround ourselves with, and for the makers who spend their lives in the effort of making them.

The objects I admire - and that I bring into The Leopard - are valuable precisely because they are the record of a human encounter with materials. The slight unevenness in a hand-thrown pot. The variation in a hand-marbled paper. The deep blue of a Joume scarf that contains, fixed in iron and light, the memory of one specific French afternoon. These things cannot be generated. They can only be made. And making requires time, and attention, and the particular vulnerability of a person who has put their hands into something and is waiting to see what comes back.

What AI can do - and here is where I think the more interesting conversation actually lives - is carry the weight that isn’t making. The administrative burden of running a small atelier. The search for historical references. The translation of a maker’s story into a language a new audience can find. It can, in other words, take friction away from the things that aren’t the craft, so that the maker can spend more time on the craft.

Not AI as a replacement for the creative act. AI as the infrastructure that allows more people to dedicate themselves to it. That is the version I find genuinely hopeful.

Why I Started This Business

When I think about why I started The Leopard, I come back to something simple: that the objects we live with shape us, and that we deserve to be shaped by things that carry a story. Not perfect things - beauty, as I understand it, has never required perfection. It requires presence. The presence of someone who cared enough to spend real time making something, who measured the chemicals and designed the print and stood in the morning light waiting to see what the sun would do.

AI cannot offer that presence. It has no afternoons. No mornings in the light. No memory of anything that moved it.

The cyanotype process is now a hundred and eighty years old. It still requires iron salts mixed by a person, a design chosen by a person, a day spent in the light by a person. And it still produces objects of extraordinary singularity. That is not a sentimental observation about the past. It is a precise observation about what makes something worth keeping.

One hundred and eighty years later, Alice Fraudeau is still standing in the light, waiting to see what the sun will do.

The Leopard