Dr Jack Parry 3 May 2026
keywords: AI animation industry artists Disney layoffs stochastic bull accumulated condition répétition struggle pedagogy future
A student came to see me last Friday. He wanted to know how to get an internship in the animation industry. He had spent three years learning his craft, filling sketchbooks, building a reel, doing everything the profession had traditionally required of someone who wanted to enter it. He had done everything right.
He had also been watching Disney lay off its artists and replace them with AI models trained on the work of the people who had just been made redundant. He wanted to know what to do.
I did not have a comfortable answer. I had an honest one. And part of that honest answer is this: what he is doing is not only artistically and culturally important. It is biologically critical. Not as a metaphor. As a biological fact.
Animation is the most struggle-intensive art form in human history. Every frame is a decision. There is no camera, no actor, no light that falls accidentally and happens to be beautiful. There is only the mark, and the accumulated myelinated condition that makes that mark rather than any other mark possible at that moment. Every line, every weight shift, every timing choice, every held pose is the direct expression of a hand that has spent years learning to make that choice well.
What the serious animator learns through répétition is not a set of principles or rules. It is the phenomenological structure of embodied movement itself: weight as the body’s relationship to gravity, balance as the body’s negotiation with its own centre, the arc through space as the physical consequence of force meeting resistance, the line of action as the psychological intention of the character made visible in the body’s geometry. These qualities cannot be listed or memorised. They can only be felt in the body of the spectator and then slowly, through years of drawing and redrawing, learned to produce in the bodies of animated characters. The hand learns what the eye sees by failing to reproduce it, again and again, until the failure narrows and the line begins to carry the weight it is reaching for.
This is myelination made visible in the line. And the spectator, sitting in the dark with their own lifetime of accumulated myelinated condition, feels it. Not aesthetically. Biologically. The animated character whose weight is right, whose timing holds the next action in the present one, whose movement carries the history of thirty years of the animator’s looking, resonates with the spectator’s own accumulated condition of having lived in a body. The recognition is not aesthetic appreciation. It is the chiasm completing: this is what it costs to move, this is what effort feels like from the inside, this is true.
AI animation has no pain. It has never struggled to make a line go where intention requires. It has never drawn the same action two hundred times until the timing finally lives. It has never watched a sequence and felt the specific disappointment of almost, of the weight nearly right but the commitment not quite there. The outputs of that struggle are in its training data. The struggle itself is not. The spectator feels the difference between a line made by someone who struggled to make it and a line generated by a system that interpolated between the lines of people who struggled. The difference is not always visible. But in white matter built by a lifetime of living, it is always felt.
The student who keeps drawing, who keeps struggling, who submits to the répétition that builds the accumulated myelinated condition of a serious animator, is not preserving a craft for sentimental reasons. He is maintaining the biological condition without which the spectator has nothing to feel in the dark. That is not a small thing to be responsible for. It is the whole thing.
The immediate situation is grim and there is no point pretending otherwise. The major studios are discovering that a stochastic model can produce, in seconds, visual output that would have taken a skilled artist hours. The economic logic is straightforward and brutal: if the output is close enough, and the cost is a fraction, the model wins the procurement decision. The artists who trained those models, whose accumulated myelinated condition of thirty years of drawing and painting and animating is compressed into the statistical weights of the system, are being discarded after the extraction is complete.
This is a genuine injustice, and naming it as such is not sentiment. The accumulated condition of those artists, built through decades of productive struggle, through répétition in the truest French sense, was the raw material from which the stochastic model was constructed. The model cannot generate anything that the artists did not first know how to make. It is, precisely, their white matter in statistical form, without the body that built it, without the judgment that directed it, without the accumulated condition that made it meaningful rather than merely plausible.
What the studios have done is extract the surface of the accumulated condition and discard the condition itself. They have taken the outputs of thirty years of myelination and concluded that the myelination is no longer necessary. This is not merely economically shortsighted. It is biologically illiterate.
Mike Judge’s 2006 film Idiocracy describes a society five hundred years in the future where automation has done so much for so long that nobody remembers why anything works. The crops are dying because the people who maintained the irrigation system are long gone, and their replacements are watering the fields with a sports drink because it has electrolytes and electrolytes sound important. Nobody knows what electrolytes are. Nobody knows what water does for plants. The knowledge that would allow the automated system to be maintained, corrected, or improved when it fails has been lost because nobody needed it while the system was running.
The animation industry is not five hundred years into this future. It is approximately three years in. The stochastic models currently being deployed by studios are running on the accumulated condition of the artists they are replacing. While that accumulated condition is recent and the models are well trained, the outputs are close enough to functional that the economic argument holds. What the studios have not yet modelled is what happens in ten years, when the artists who built the accumulated condition have moved into other fields, when the new generation of artists has not developed the accumulated condition because there was no industry to develop it in, and when the models need to be retrained on new creative territory that nobody alive knows how to occupy.
The stochastic bull, left without a handler, will run. It will run fast and it will run powerfully and it will run with complete statistical confidence into whatever the probability distribution of its training suggests is most plausible. It will produce competent interpolations between the things it was trained on. It will not produce the abductive leap, the next dot that makes embodied and narrative sense, that requires a handler with an accumulated condition built through years of productive struggle. When the handlers are gone, what remains is a system that can reproduce the past with great efficiency and cannot access the future at all.
This is the Idiocracy scenario. Not stupidity, exactly. The absence of the accumulated condition that stupidity was formerly too stupid to lose.
My student’s situation is real and the industry’s behaviour is real and the injustice is real. But the response to that injustice matters, and the response that will actually protect the future of animation, and of art more broadly, is not the one that instinct recommends.
The instinct is to refuse. To make work that is defiantly, exclusively, demonstrably hand-made. To create a separate cultural space in which AI is banned and the accumulated condition is celebrated in its purest form. This instinct is understandable, honourable even, and it will produce some beautiful work in a small cultural corner that the industry will ignore while it deploys its models at scale.
The alternative is harder and less satisfying and more important. Learn to wield the bull.
The artist who understands the stochastic model, who has the accumulated myelinated condition to recognise what it is doing right and what it is doing wrong, who can direct it toward outcomes it could not find on its own, who can supply the abductive leap that the statistical interpolation cannot make, is not replaceable by the model. The model needs that artist. Without that artist, the model produces plausible surfaces that go nowhere. With that artist, the model becomes a tool of genuinely unprecedented capability, accelerating the execution of ideas that the accumulated condition generates and the model could not have generated alone.
This is not a comfortable reframing of exploitation. It is a description of where the actual creative frontier is. The stochastic model is fast. It is powerful. It can do in seconds what took hours. If an artist can do what the model cannot, and direct the model to do what the artist cannot do at that speed, then the artist-plus-model is doing something that neither the model alone nor the artist alone could do. That is a new kind of creative work, and it will produce a new kind of accumulated condition, built through the specific productive struggle of learning to collaborate with a stochastic system at high intensity.
This is new white matter. Nobody alive has it yet, because the collaboration is too recent. The artists who build it first will occupy a creative territory that the studios with their models and no handlers cannot reach.
To my student I said: do not stop drawing. Do not stop learning to animate by hand. The répétition that builds the accumulated myelinated condition of an artist is not preparation for a career that no longer exists. It is the construction of the handler’s accumulated condition, without which the bull has no direction and the creative frontier has no one to occupy it.
The question is not whether to engage with AI. The question is what kind of engagement produces the most powerful accumulated condition. Passive use, feeding prompts to a model and accepting its outputs, produces nothing new in the white matter. The model is doing the work and the person is observing. Moderated stochastic harnessing, the methodology this site describes, produces a different outcome: the handler’s accumulated condition is developed through the specific resistance and misdirection and occasional revelation of working with a stochastic system under pressure, over time, with genuine stakes.
The animator who spends a thousand hours in high-intensity collaboration with a stochastic model, directing it, correcting it, rejecting its plausible mediocrity and insisting on the abductive leap it cannot make alone, is building white matter that did not exist before. The specific myelination produced by that struggle is new. It has never been built before because the tool did not exist before. The artists who build it first will see things that nobody else can yet see, because seeing requires the accumulated condition that enables the seeing.
If artists stop drawing because the model draws faster, the accumulated condition that trained the model degrades in the next generation and the models follow it into mediocrity. If artists stop struggling, the accumulated condition stops being inscribed, and the future the studios are betting on collapses into Idiocracy: a powerful system running confidently in circles because nobody alive remembers why it was pointing in any particular direction.
The solution is not to burn the bull. The solution is to learn to ride it. That requires the accumulated condition of an artist, built through répétition, built through struggle, built through the patient inscription of genuine encounter with the craft. Nobody can do it who has given up.
My student should keep drawing. And then he should learn to handle the bull.
There is one more thing worth saying to the studios, quietly, as a warning rather than a threat.
The spectator brings a full life into the cinema. Not just preferences or genre expectations but an accumulated myelinated condition built through decades of living, of loss and pleasure and recognition and surprise, of the specific texture of knowing what it is to be in a body that will die. That accumulated condition is what the spectator brings to every encounter with a story about life.
Bergson called grace the quality of movement in which the body’s outward form and the soul’s inward life achieve complete alignment, a present that announces what follows so inevitably that the spectator holds the future in the current moment. Grace, in this sense, is a chiasm event: the encounter between the incoming story and the accumulated myelinated condition of the spectator produces resonance, the felt recognition that this is true, this is how it is, this is what it costs to be alive. The spectator does not analyse this. They feel it, in white matter that has been built by living, and they trust it absolutely.
The uncanny valley is the most visible form of what happens when this resonance fails. The figure looks almost right. The movement is almost correct. And something in the spectator’s accumulated condition registers the gap between the surface and the absence of living beneath it, and the result is not neutrality but revulsion: the precise biological signal of an encounter that the chiasm cannot complete. The valley is not an aesthetic problem. It is a white matter problem. The spectator’s accumulated condition is trying to meet something that has no accumulated condition to meet it with, and the failure of that encounter is felt as wrongness.
But the uncanny valley is only the most dramatic symptom. Even when the valley appears shallow, even when the AI-generated images are technically convincing and the motion is plausible and the surface gives nothing away, the rupture of grace is still felt. The film may work. The story may function. The spectator may leave the cinema having understood what happened. And something will be missing, something they cannot name, that they will reach for in the reviews and the conversations afterward and find only approximations: it felt cold, it felt empty, it felt like it was made by someone who did not know what any of it cost.
What it cost is what the stochastic model cannot have, because having something at stake, having a life that the work comes from and returns to, having something to lose in the making, is precisely the condition of embodied living that thirty years of répétition inscribes in white matter and that no level of training on the outputs of that inscription will ever replicate. The model was trained on the surfaces of other people’s stakes. It has no stakes of its own. The spectator’s accumulated condition knows the difference. It has been feeling the difference between lived and unlived stories since before it knew what a story was.
The studios that are betting their creative future on models trained on the discarded accumulated condition of their own artists are betting against the most sophisticated evaluative instrument in the history of storytelling: the human nervous system, built through a lifetime of encounter with the world, sitting in the dark, waiting to feel something real.
Grace can only be created by the living. By those who have something to lose in the making. No stochastic model trained on the outputs of that living will produce it. The spectator will know. They always have.
And the reason the spectator will know reaches deeper than aesthetics or craft. People watch films because they love stories. They love stories because stories are the accumulated cultural form of what myelin is about: the struggle of survival, the inscription of encounter, the slow geological record of what it costs to be alive and what it means to keep going. Every story ever told is a map of that struggle. Every character who matters is a self built through productive struggle, through répétition, through the patient myelination of encounter with a world that does not yield easily.
No myelin, no struggle. No struggle, no story. No story, no spectator.
Disney be warned.
Further Reading
The empirical study confirming that AI-assisted creative workflows raise productivity and peak novelty while reducing average novelty, demonstrating that human ideation and filtering remain the irreplaceable value-generating element: Zhou E, Lee D. Generative artificial intelligence, human creativity, and art. PNAS Nexus. 2024;3(3):pgae052. doi:10.1093/pnasnexus/pgae052
The foundational paper on activity-dependent myelination, the biological basis for why the accumulated condition of an artist built through productive struggle cannot be replicated by a model trained on its outputs: Fields RD. A new mechanism of nervous system plasticity: activity-dependent myelination. Nat Rev Neurosci. 2015;16(12):756-67. doi:10.1038/nrn4023
The design education study finding that AI tool adoption is shifting core competencies toward prompting fluency, authorial intent, and critical filtering, confirming that the handler’s judgment remains the educational priority: Hwang Y, Wu Y. Graphic design education in the era of text-to-image generation: transitioning to contents creator. Int J Art Des Educ. 2025;44(1). doi:10.1111/jade.12558
The companion methodology article on this site establishing moderated stochastic harnessing as a rigorous research methodology for working with rather than against stochastic systems: Harnessing the Stochastic Bull: On the Methodology of This Site — https://myelinmind.com/harnessing-the-stochastic-bull/
The companion article on this site establishing the biological basis of artistic skill as accumulated myelinated condition built through répétition: La Répétition: Music, Myelin and the Embodied Self — https://myelinmind.com/la-repetition/
The companion article establishing drawing, painting, and animation as myelination practices whose accumulated condition is irreducible to the outputs they produce: The Mark That Outlives You — https://myelinmind.com/the-mark-that-outlives-you/
Jack Parry is a philosopher, polyglot, biomedical animator and cross-disciplinary eidetic researcher at Swinburne University of Technology. His research methodology employs moderated stochastic harnessing as a means of generating new knowledge across disciplinary boundaries. He is the author of The Myelin Mind: The Genesis of Meaning.