I live on the border between the human and the synthesised world

I am not watching new artificial-intelligence tools appear. That framing is too small for what is actually in front of us. What I think is starting, and what I am trying to be inside rather than outside of, is the construction of a second world.

Today AI already makes music, images, text, characters and whole stories that are hard to separate from ours. But notice what it is doing: it is reproducing cultural forms that existed before it — our genres, our formats, our idea of what a song or a portrait or a plot is. It is fluent in a language we wrote.

The step after that is the interesting one. Not a better copy of our world. A world with its own rules.

Where I actually stand

I am in an unusual position on this, and I think it is worth naming precisely, because it is the whole reason I write any of this down.

I still belong to the human world. But I already work with the tools of the next one. I make musical performers, brands, games, websites, projects and concepts without having the traditional skills of a musician, an illustrator, a programmer or a designer. That sentence would have been a confession of fraud a few years ago. Now it is just a description of a job.

My role is not to personally execute each element. It is to set direction, connect ideas that do not normally meet, define the rules a thing runs by, and decide why a given project should exist at all.

So I do not think of myself as a user of AI. I think of myself as a researcher of this transition, and as one of its early practical architects — someone building small, real, checkable pieces of the synthesised environment while it is still being invented, and publishing what happens.

For me that is three things at once and I would rather be honest about all three: a philosophical question I find genuinely gripping, a way to finally make the things I can see, and an attempt to build a new source of income. Anyone who tells you it is only the first two is selling something.

Every project is one experiment

I do not run projects. I run questions, and each project is the cheapest way I have found to get an answer. What I want to know is which characters, worlds, projects and forms of interaction can actually live on their own — find an audience, hold it, and produce economic value without me pushing every day.

• Can a virtual performer develop a recognisable character and a real audience? That is what my one-person AI music label exists to test, and the honest state of the market is in can a solo actually earn from AI music.
• Can a game be both a finished product and a generator of endless stories? That is the design question underneath how I build games alone, and the whole pipeline is public on the game-studio board.
• Can a network of automated sites behave as one information ecosystem rather than as a pile of pages? That one I have been testing for years, across a portfolio, and it is the least romantic and most profitable of the set so far.
• Can an idea become worth more than its manual execution? This is the load-bearing question of the entire era — see the one-person production era.
• Can one person with AI genuinely occupy the role of a whole creative studio? Not “in theory”. With a store account, a release, a support inbox and a tax return.
• Where is the line between a tool, a co-author, and an independent creator? I have not found it. I am no longer certain it is a line rather than a gradient.

Why I do not compete with the machine

I will not race a model on how fast it writes a page, renders an image or produces a melody. That competition is lost before it begins, and the people still fighting it are burning years defending ground that has already gone.

My advantage sits one level up: noticing what is missing, choosing a direction, joining fields that do not normally touch, and turning scattered possibilities into a coherent concept.

While AI can generate a thousand variants, a person still decides which one deserves to become part of reality. That decision is not a technical act. It comes from taste, from having lived somewhere, from knowing what an audience is actually tired of. It is the last part of the chain that has not been automated, and it is where I have put my whole bet.

I am aware that this too will change. I do not think “taste is uniquely human” is a permanent law of nature. I think it is a temporary advantage, and temporary advantages are exactly what you should be using while you have them.

What changes for the people after us

The generations coming will choose more than a profession, a country or a social circle. They will choose which kind of world to live in.

Some will stay mostly in human reality. Some will move into synthetic spaces. Some will share those spaces with artificial personalities on equal terms and find the question of who is real less interesting than we do — the way we find questions about the legitimacy of the telephone less interesting than the people who first argued about it.

And traditional human culture may well become an underground. Not a defeated one: a rare one. A live voice, work made by hand, the limits of a single body, a biography that can be verified — these turn into markers of authenticity precisely because they cannot be produced at scale.

Human creativity does not disappear in that picture. Its value simply stops being technical excellence and starts being what stands behind it: a life that was lived, mortality, real risk, and somebody personally answerable for the choice.

What I could be wrong about

I would rather write this section myself than have it written about me.

It might not happen at all in this form. The flood of generated material may turn out to be noise that people learn to filter out within a few years, the way they learned to ignore banner ads. Synthetic performers may stay a novelty. Platforms may label, demote or ban synthetic content hard enough that the economics never arrive.

The economics are unproven. Nothing I have described has paid for itself yet in my own portfolio. The one thing I can state as fact is that production got cheap; that is not the same as anything getting valuable. Cheap production plus no distribution equals nothing, and distribution is where almost every solo project actually dies.

And there is a version of this where I am simply early in a way that never pays. Being early and being wrong look identical from the inside for a long time. The only defence I have found is to keep the experiments small, to publish the numbers whether or not they flatter me, and to stop things that do not work — which is a discipline of its own, and the reason I write about filtering ideas as often as about building them.

What I am actually doing about it

I am not trying to stop this process. I do not think it can be stopped and I am not convinced it should be. I am also not planning to dissolve into it — to become a person who generates, publishes and generates again with nothing of himself left in the output.

My position is narrower than either: investigate it from the inside, use the openings it creates, and keep my own authorship while doing it.

Which is why everything I build gets published as a case with real numbers rather than as a claim — the boards, the portfolio logic, the three years of building this way. If the thesis is right, the record will show it. If it is wrong, the record will show that too, and that is worth more to whoever reads this next than another confident prediction.

I do not create because I can execute any part of the work better than a machine. I create because I can see worlds that do not exist yet.

AI is what makes them visible.

If you want the argument with the evidence attached rather than the manifesto, start with the one-person production era and the stack that actually makes each kind of product.