The best creative expression are autotelic
Autotelic (from the Greek auto- meaning "self" and telos meaning "goal") refers to something that has its purpose in itself rather than serving an external end. An autotelic activity is done for the sake of it.
Any genuine creative expression, whether a thriving business or an artistic enterprise must reflect autotelic values and principles. Why? Because those values come from the deepest trenches of one’s soul, where things exist solely for their own sake. These values are the best way to infuse uniqueness into one’s enterprise, because everyone’s soul is unique by default. Fundamental genuineness is the mother of authenticity and originality.
An enterprise, a creative product, a life, these are chaotic stories, where a single decision can have a monstrous impact over the very long term. So one must ensure that the few decisions they make are at least respectful of their own virtue. Otherwise, they will end up chasing someone else’s dream, built on someone else’s values and someone else’s style. In a world more random than we’d like to admit, we don’t control outcomes nearly as much as we think. The only real power we have lies in a handful of decisions, not in their consequences. So we’d better make those decisions by listening to who we truly are. And who we truly are is autotelic, we exist because we exist, and that’s enough.
Thus, when cultivating the difference inherent in any original work, there is always, more fundamentally, someone or some entity listening to their own deepest instincts. That’s why unique products are not only unique in the present moment but also in how they evolve. They differ now, and they differ in the path they take over time. Two creative souls are not only different when observed, they are different in how they diverge, regardless of how much they appear to have in common.
That’s why activities rooted solely in comparison, status games, and peer validation (which is not to say some external validation is never needed) can never be authentic self-expression. Their starting point is neither a divergence from consensus nor a meditation on honest instinct. They are mere imitations, of what we should be, what we should look like, and worst of all, what we should want to want. It’s like trying to write a rhymed love poem based on someone else’s already written love. No. When it comes to finding one’s purpose, there is, always an autotelic element in any successful creative expression, be it an artwork, a business, or anything else.
Only after that inner divergence comes convergence: a reduction to one, single purpose for any activity or use of time. Just as a good poem begins in scattered points across space and time and converges into a single final line, one’s actions should converge into something singular. Into something sole. There is something unique and inimitable in one’s purpose because it isn’t chosen by the outside world but whispered by one’s soul.
Contrarian minds
The best autotelic expressions in business usually appear as contrarian moves: people questioning conventions, products establishing differences with no apparent reason other than their own internal logic. They often look irrational at first glance, unexpected evolutions under new technologies, or creative adaptations within rigid, old-fashioned industries where terminal state has been supposedly reached.
This is where AI often fails to grasp certain deep human refinements. AI systems, including those by OpenAI like GPT or DALL·E, are not built to wander around. They are built to optimize for a goal, a measurable objective. GPT, for example, is trained to predict the next word in a sentence based on a massive dataset of human-written text. DALL·E does something similar for images: it learns how to generate pictures that match a text description by being shown millions of examples of image-caption pairs.
In both cases, these systems learn by imitation. Their entire learning process is governed by a metric of success, how well the model’s output matches human examples. The more “correct” it looks, according to its training data, the better. Even when the result feels creative or surprising, it’s only because it’s statistically plausible, not because the system is discovering something new.
But no real artist, no one who ever made anything truly original, starts that way. They don’t know what the end result will be. They discover it. They struggle toward it. They constantly adjust their own expectations. The best creative works aren’t solved like problems; they’re discovered through autotelic process. There’s always doubt, tension. That’s the drive. The pursuit matters more than any fixed goal because the work can never be done as well as it can be done.
This is the great difference. GPT and DALL·E start from the endpoint, a predefined target like matching a prompt or mimicking good writing, and work backward. Their training code encodes a very specific idea of success, this is what “good” looks like. And where does that idea come from? From the collective mass of examples, the consensus. From past data, not inner experience. Even when the model adds randomness to its output, it’s just statistical noise, not intention. Randomness and autotelic purpose are not the same. The idea of success is baked into the training process.
Autotelic activity means creating something not to match an external standard, but because it reflects something within. Something only you could have made. That’s why truly original human works are shocking, not just for their surface beauty, but for realization: a fellow human made this. A human. With fear, hope and confusion. Current AI can imitate greatness. But it cannot mean it. It cannot feel that inner necessity. Because the result it reaches belongs to no one. It has no soul to speak from, no value system to return to. It cannot express the unseen in terms of the seen. Maybe it will change but, as of now, creative works more often than not feels off.
Divergence as the starting point
The starting point of any original work resembles divergence more than convergence. It begins with a conscious or indifferent departure from the consensus. It starts with a conscious or indifferent difference, being different for its own sake, not out of a desire to be different. Everyone is unique because of this. This is difference for its own sake, or rather, by default.
This is why, by listening to one's deep entrenched instinct, the work that comes as an output is naturally genuine, because “nobody can beat you at being yourself”, as said Naval Ravikant.
Divergence from an existing point, an existing consensus, or whatever is the dominant consensus at any given time, must be the expression of a genuine feeling. If you try to be different without being genuine, you're always naturally copying someone else in this world being different, and you're no longer different.
AI is different. It's fed with millions of example, and it sounds or looks like a consensus. It has no autotelic elements corresponding to a unique soul. Whatever it does, it does it to optimize a certain metric. A creative expression of a unique kind blended with other expressions of unique kinds made consensus because there is no such thing as the average of art, as the average of sincerity, as the average of genuineness. Any average can't be unique. It smooths things and removes the asperity that makes things refined and naturally human and attractive.
I think that in any pleasant work there is an underlying feeling that it could have been myself or it's another fellow human doing it, someone that has deep inside them the same instincts and the same organism as I have. Smoothing then kills the refinement, and it's one highway toward non-original work or at least non-pleasant work.
I think that's why AI is better used to help catalyze the original divergence, the starting point, using its encyclopaedic knowledge, or to remove minor binary mistakes like grammatical on some otherwise original work. Because someone thinking on a specific topic creates connections between various disciplines. And it's not only static disciplines, it's their own very unique way to blend those disciplines through their very unique experience of them.