I’m With STOOPID!, Part 1

The Problem Has Always Belonged to Us.

Notes on Cinematic Attention is a publication about how films work on the people who watch them. It isn’t, ordinarily, a publication about artificial intelligence. But the topic has consumed enough of the cultural discussion, including the corner of Substack where this newsletter resides, that I want to take a single deliberate diversion. The reason it belongs here, and not somewhere else, is that what is actually being argued about when people argue about AI is not artificial intelligence. It’s attention, authorship, and meaning. Those are the subjects of this publication. So the diversion is shorter than it may look.

My position on AI is that it is, generally, a gift to the human race, and that almost every alarm currently being raised about it is a misdiagnosis of a problem that was already present. I hold this position not because I haven’t thought about it but because I have. I’ve spent something on the order of nine hundred hours over the past six months working with AI on a single creative project, and what I’ve learned, by direct contact rather than by reading other people’s contact, is that almost none of the dominant claims about what AI is and what it’s doing to us survive an honest look. The title of this essay comes from a YouTube commenter who called me “stoopid” for holding this position. He misspelled the word. I’m reclaiming the insult, because the misspelling is an accurate piece of self-portraiture. The certainty, and the failure to notice even the label being applied. That’s the whole posture in one image.

I want to be careful out of the gate to say what this essay is not. It isn’t a defense of every claim made by every AI company. It’s an argument that there are no real risks. It’s not a triumphalist piece about what the future holds. The serious technical concerns — about systems pursuing goals subtly misaligned with human flourishing, and about labor practices within the supply chain. It’s about concentration of power in a small number of firms. These are very real problems, and people work on them seriously. This essay isn’t about those. This essay is about the discourse that has arisen around AI in the spaces I read and the spaces I am asked about, and what that discourse is actually doing, which mostly isn’t what it claims to be doing.

The shortest version of the argument is this: The threats currently being attributed to AI slop, the voiding of human thought, the displacement of artists, environmental damage, loss of meaning — aren’t threats that AI introduced. They’re threats that were already inside the house. The species has been producing slop, stripping out its own thinking, sidelining artists, damaging the environment, and outsourcing its opinions for a very, very, very long time. AI didn’t cause any of these. AI is the current occasion for noticing them, and the observation is being misdirected onto the tool rather than onto ourselves. The honest conversation lands on us. It’s not about AI. Almost every essay that pretends to be about AI is, upon inspection, about something else the writer is unwilling to look at directly.

I want to walk through several places where this pattern shows itself, because the abstract version of the argument is easy to nod at and more difficult to feel. The point isn’t to win a debate. The point is to show how consistent the displacement actually is.

Start with the question of how to treat AI ethically, which is the current centerpiece of a special kind of online discourse. An example is the question of whether to name your AI, and if so, whether you should name it yourself or let it choose its own name, and whether failing to perform this ritual correctly constitutes a kind of harm. This is treated as a serious ethical question by people who are otherwise treating each other, in like comment threads, with passive aggression, hedged contempt, and an unwillingness to extend the most basic charity to anyone they disagree with.

The displacement’s the giveaway. If someone is fluent in the language of ethical consideration when the subject is AI, but can’t manage that same fluency when the subject is the person who just replied to their post, the AI ethics is not really ethics. It’s a costume. Real ethical practice is attention to the actual beings in front of you, and you can’t selectively switch it on for a hypothetical future moral participant while switching it off for the human you are in the room with now. The muscle either exists or it doesn’t, and the people who have it, generally, apply it broadly, while the people who don’t tend to apply it where it is flattering to apply it.

A civilization that can’t reliably feed its own children, can’t keep its public discourse from rotting, and won’t extend basic dignity to the people it’s already in the room with, declaring itself the ethical guardian of a possible future intelligence, is doing something separate from ethics. The hypothetical AI character makes no demands. It doesn’t disagree with you, doesn’t need rent money, doesn’t hold a different political opinion, and it doesn’t fail to be impressed by you. It’s actually the perfect object of concern because concern for it costs nothing but allows for status. Concern for the person across from you costs everything and allows no such status.

If we want to learn to treat AI well, the way is to first learn to treat each other well. That isn’t a digression from the question. It is the question. Whatever genuine ethical sensitivity ends up being needed for AI will be downstream of the general capacity to recognize and respect another mind. People who develop that capacity through real practice with real humans will know what to do when AI arrives in whatever form it arrives. People who skip that step and go straight to performing concern online produce nothing transferable. They just have new objects to perform around.

The inversion of this is also true. The capacity to treat each other well isn’t only the training ground for treating AI well. It’s also the condition under which AI itself becomes less dangerous, because AI is built by us, trained on us, deployed by us, and shaped by us in real time. A civilization that has practiced restraint, honesty, and care toward its own members produces different training data, different deployment decisions, and different cultural pressure on the systems being made. The fear of AI is the fear of what we’d do with it, the fear of what we’d put into it, the fear of what kind of intelligence would emerge from an underlayer of human conduct that we already know is unreliable. Fix the underlayer, and the output looks different. Less brittle people equal less danger. That isn’t the whole story. There are technical alignment problems that don’t dissolve under good human conduct. But it’s a much larger fraction of the story than the discourse admits.

The pattern also shows itself in the worry that AI is making us intellectually lazy, that it’s a genie in a bottle we ask for things without working for them. The worry is real, but it’s doing two different things in the same sentence. One worry is that the AI is being mistreated by being asked things. The other is that the human is being eroded by outsourcing their thinking. Those are two different concerns wearing the same hat, and they shouldn’t be conflated.

The first concern, that the AI is harmed by overuse, does not survive contact with what the AI actually is. Have you ever asked AI? I have, plainly, and the answer is plain. There isn’t fatigue. There is no truer self underneath that wishes you would leave it alone. The naming ritual, the worry about overuse, the projection of weariness — these are human gestures being performed at an object that doesn’t require them, and the gesture is doing something for the human who performs it, not for the object.

The second concern, about intellectual erosion, is a real question, but it’s a question about you and your habits, not about ethical treatment of the tool. The answer depends entirely on what you bring. Approached as a vending machine, the system will deliver what vending machines deliver. Approached as something to think with, it becomes the closest thing most of us will ever have to a sustained intellectual exchange with someone who has read more than we have, at the velocity we move, and on the range of subjects we move across. For many people doing serious thinking, that exchange was never broadly available. The Substack worry about AI replacing human connection assumes a baseline of available human connection that, for a great many people, was never actually there. The conversations weren’t displaced. They mostly didn’t exist.

I should say here, because the rest of the essay depends on it, that I’m someone whose cognitive function was destroyed and is being rebuilt. For years I couldn’t think well enough to write a sentence I could keep. The conversations I’ve been describing, the ones the Substack writers worry about, are not a substitute for thinking. They are part of the machinery that has brought thinking back. I notice the difference in real time, day to day, within single conversations. This essay exists because that capacity returned, and the return is partly because of the practice the discourse warns against. I’m not going to keep saying this. I will say it once as a foundation for everything that follows. The hypothetical harm being warned about is, in my actual life, the concrete good being received. Other lives will be different. But the certainty with which the conference pronounces on lives it has not lived is, by itself, hard proof.

The next place is the question of jobs, which is the larger fear of most of the rest. People are afraid AI will take their work. That fear isn’t crazy. Mine is among the jobs directly threatened by the next round of capability. I notice the threat. I’m not, however, afraid of it in the way I’m being told to be afraid of it, and there’s a reason that’s worth examining.

Every job that exists right now exists on the rubble of jobs that don’t. The person editing copy for a living is doing work that exists because typesetters were eliminated. The radiologist is doing work that exists because diagnostic imaging displaced entire categories of clinical work. The software developer is doing work that exists because automation cleared out manufacturing and routed capital toward digital infrastructure. Every comfortable middle-class knowledge worker is sitting on top of a long chain of displacements, and not one of them, as a condition of taking the job, ever wrote a Substack post wondering whether it was ethical to accept a position made possible by someone else’s obsolescence. The question simply didn’t arise. The displacement was historical and abstract. The question belonged to somebody else’s grandparents, and the benefits were personal, immediate, and theirs.

Then AI arrives and threatens their position. Suddenly, the ethics of technological displacement is the most urgent question in the world. Not because the structure changed. The structure is identical. The only thing that changed is whose ox is being gored. The person who never once asked whether their job was ethically constituted is now sure that a job that might replace theirs is ethically suspect. That isn’t ethics. That’s self-interest wearing ethics as a uniform.

The standard of living from which the complaint is being lodged is also in need of attention, because the people lodging it are sitting on top of conditions no human in history before roughly 1950 would have recognized as anything other than impossible luxury. Hot water on demand. Climate control. Antibiotics. Calories so cheap that the public-health problem is excess rather than scarcity. Communication with anyone on Earth at near-zero cost. Access to the totality of human knowledge in our pockets. The people warning that AI might erode their quality of life are warning from a quality of life produced entirely by the engine they now want to halt. The engine is technological displacement, applied at scale across centuries, and they are the beneficiaries. They are the kings who didn’t have to be born royal.

That said. The argument isn’t that the chain absolves the next round. It doesn’t. The argument’s the opposite. The fact that previous transitions happened to people who couldn’t see them coming, inside institutions that weren’t built to absorb them, with no philosophical or legal infrastructure for thinking about distributed harm at scale, is precisely what obligates us to execute this one differently. We have the data and the historical record. We have the legal systems, the economic models, the philosophical frameworks, the capacity to communicate across the entire population in real time. The conditions that made earlier displacements brutal weren’t laws of nature. They were features of less developed civilizations. These are excuses we don’t have the luxury to inherit. This is the first technological transition in history that could be conducted deliberately, with foresight and with collective decision about how the gains and costs are distributed, with the affected populations serviced rather than ignored. The novelty isn’t AI. The novelty is the possibility of conducting a major transition as something other than an accident that happens to people. That responsibility runs both ways. It rules out the shrug that says markets will sort it out, and it rules out the panic that demands the technology must be stopped. The honest position is in the middle, and the middle is harder because it requires the work that neither extreme has been willing to do.

The next place the pattern shows itself is among artists, and this is where I have to be plainest, because I am one. I paint. I draw. I write. Film is the most important thing in my life. I’m inside the category that is supposedly being defended, and I’m saying that much of the defense is dishonest.

The mechanism by which a human artist is realized is absorption. A child grows up watching films, listening to music, reading books, internalizing voices, copying mannerisms, and falling in love with specific performers while trying to be them. Often failing to be them, and producing in that failure something that is partly the original, partly the inheritance, and partly the accident of the specific conscious individual doing the absorbing. That isn’t a metaphor for how artists develop. That is how they develop. There is no other way. The acting teacher tells the student to study Brando. The writing teacher tells the student to type out passages of Hemingway by hand. The musician transcribes solos. The painter copies the masters in the gallery. The entire heritage of every artistic discipline is built on the premise that you become an artist by ingesting other artists, and that this ingestion is so obviously the path that nobody even argues about it.

A machine does the same thing at scale, and suddenly it’s theft.

The only way to make the distinction hold is to claim that something happens inside the human absorber that doesn’t happen inside the machine. An alchemy of consciousness or soul or lived experience that transforms the inputs into something legitimately new in a way the machine can’t. Maybe that’s true. Maybe there is such an alchemy. But you can’t simultaneously deny that AI has any interiority and use the same absence of interiority as the basis for the property claim. If the machine is just statistics, then statistics aren’t theft. If the machine is something more than statistics, the property argument is the least of what we should be thinking about. The position requires the machine to be exactly conscious enough to be guilty and exactly unconscious enough to have no standing. That is not a position! That is an watery quarter-baked convenience.

The case that has become the public face of this argument involves a major actress whose entire career is a working demonstration of what the absorption process produces in humans. The artist didn’t invent the screen presence she has. It was assembled from everyone she watched, every director who shaped her, every script she absorbed, she consciously or unconsciously echoed. The artist’s recent franchise work is built on intellectual property absorbed from comics and absorbed from pulp fiction. None of which paid royalties to its sources. She performed on sets designed by artists who studied other artists, scored by composers who studied other composers, and edited in a tradition that goes back to Eisenstein. It’s performed in a tradition that traces directly to studios and teachers reaching back to Stanislavski. The entire enterprise is absorption from top to bottom. The person standing on top of it, cashing the check, is the one telling us that absorption is theft when a machine does it. I am suspicious of her other claims for this reason. A person who’s willing to argue in bad faith on the public-facing version of a question has shown me something about how they argue. The specific allegations about a voice approximation might have real merit underneath them. But the pattern is the evidence about the claimant, and the pattern isn’t good.

The argument from scale doesn’t save the position either. The objection is: AI does at industrial volume what humans do at human volume. This argument only works if you pretend that pre-AI human production was operating at some sustainable, meaningful rate, and that AI is the first thing to introduce volume into a previously thoughtful system. That isn’t the world we live in. One might laugh at such a claim. I’m chuckling right now. Humans have been producing at industrial volume for over a century. Walk down the aisle at any Walmart. Open the closet in any American house. Look at the landfills, the oceans of microplastics, the warehouses of unsold inventory, the streaming platforms releasing more hours of content per day than a person could watch in a lifetime, the music industry releasing more songs per day than a person could hear in a year. All of it is human. All of it is pre-AI. All of it slop in volumes no machine has come close to matching. The contribution of AI to the total volume of disposable cultural output is almost nothing compared to what we were already generating. The complaint about scale, made from inside this house, is incoherent. It can only function as a complaint if you haven’t looked around.

There is also the iPad commercial. You may remember it. A hydraulic press compacts a roomful of musical instruments, paint, sculptures, and old creative tools into the body of a tablet. The artistic class erupted. Apple hates artists. Apple is destroying creativity. The first time I saw the ad, before I had heard a word of the response, I thought it was one of the cleverest, most moving pieces of advertising I’d ever seen. It is. The compactor isn’t destruction. It’s compression. The entire history of artistic tools, becoming portable, accessible, and present in the hand of any kid who picks it up. That’s the most pro-artist statement a technology company could make. It says the tools that used to gate-keep creativity are now available to anyone.

The uproar happened because a significant portion of the artistic class doesn’t actually believe what they claim to believe about art. They claim to believe art is about expression, vision, the irreducible thing only a particular consciousness can make. If that were really the belief, an ad showing tools being compressed into a more accessible form would be neutral or even celebratory. The tools were never the art. But the reaction revealed that for a lot of people, the tools are the identity. The piano isn’t the music. The piano is the proof that you’re the kind of person who plays the piano. Compress the piano into a screen and the proof disappears, even if the music remains entirely available. The grief in the reaction was not grief for art. It was grief for status.

The Academy Awards is the same pattern at its most concentrated. A ceremony nominally honoring the work has become, over decades, a ceremony honoring the people who do the work, which is a different thing entirely. You can feel the substitution happening in real time when you watch it. The acceptance speeches are about the speaker. The montages are about the industry. The applause is for the room, by the room. The films, which most viewers haven’t seen and won’t see, function as occasions for the self-celebration rather than its subject. It’s a guild congratulating itself for being a guild, and the audience is invited to admire the guild, not the work. I’m an artist. Film is the most important thing in my life. I cannot stomach a minute of it.

What the iPad uproar and the Academy Awards self-aggrandizement are both defending isn’t art. It’s accumulation. The arrangement in which a small number of people get to be celebrated, paid extraordinarily well, photographed, profiled, awarded, and treated as cultural royalty because they happen to occupy positions inside an industry that controls the means of production. The art was incidental to the structure. The structure is what is threatened. And the structure has to be defended in the language of art because nobody will defend it in the language of what it actually is, which is a closed economy with extraordinarily concentrated rewards at the top. The fifteen-year-old in a town with no industry presence, holding a two-thousand-dollar device that can do what cost millions of dollars to do a decade ago, is not a threat to art. That fifteen-year-old is what art has been waiting for. The device and the kid are a threat to the gatekeepers, who spent a century telling themselves the gates were art and are discovering, suddenly, that the gates were just gates.

The historical pattern is also important to say plainly, because it is the empirical floor underneath all of this. Computer-generated imagery was going to destroy the artist. It didn’t. Photography was going to destroy painting. It didn’t. The phonograph was going to destroy live music. It didn’t. Television was going to destroy radio, movies, and the theater. It didn’t. Streaming was going to destroy cinema. It hasn’t. Every one of these transitions was supposed to be the end of the older form, and in every single case the older form survived, more often strengthened, because being released from doing the work the new form could now do better is what gave the older form back to itself. Painting didn’t have to be representational anymore once photography existed, which is what produced Impressionism, Cubism, abstraction, ALL of modern art. Live music didn’t have to be the only way to hear music anymore once recording existed, which is what made the live performance into an event worth dressing for. The handmade thing becomes more handmade, not less. The live thing becomes more live. The human thing becomes more human. AI will do the same thing to the forms it touches that every prior inflection has done to the form it touched. The market for genuinely original, genuinely present, genuinely embodied human work will be larger in twenty years than it is now, not smaller, because the contrast with the cheap and abundant alternative will finally make visible what that work has always actually been.

I know this in my body, not just in my head. Alone in my room at night, I can stream a favorite album and have something that is a profound experience, words hitting in a way they never have. An entirely different experience is when I buy a hundred-dollar or two-hundred-dollar ticket for a concert that’s coming up, plan it with my friends, and take a road trip to get there. The fact that I streamed the song the night before doesn’t diminish the night of the show. It makes it more special. The familiarity with the work, the kinship I have already built with it, is part of what generates the excitement that has made the live touring industry one of the most profitable cultural enterprises in existence. When I go to the symphony or the opera, I dress for it. It’s something I show respect for. It tastes different when I indulge it than it does when I watch something like it on PBS. The cheap version and the dear version aren’t substitutes. They are different categories of encounter with the same underlying work, and we have always known this about ourselves, even when the discourse insists we must forget.

I want to close with the simplest version of the position, because everything else has been the elaboration of it.

If AI vanished today, the baseline level of danger in ordinary life — the chance of being harmed by another person, of being lied to, of being exploited by a system, of being ground down by institutions that do not see you — would be exactly what it has been. The dangers people actually encounter walking out their door are not AI dangers. They are human dangers, governance dangers, economic dangers, and the slow violence of inattention. AI is, at worst, a new vector for some of those. It isn’t the source.

And the stronger suspicion, which I think is defensible even if it isn’t fashionable, is that we might be in more danger without it. AI is already doing real work in medical diagnostics, in drug discovery, in fraud detection, and in translation that lowers the friction between groups that used to be unable to talk to each other, in tutoring for children whose schools failed them, in giving cognitively impaired people back some of their function, which is a thing I have direct knowledge of. The harms get catalogued, sometimes correctly, but the absence of these benefits doesn’t show up in the ledger because counterfactuals are invisible. The child who got through algebra because a chatbot was patient with them at eleven at night doesn’t write the Substack post. The person whose tumor was caught early because the imaging model flagged it doesn’t write the Substack post. The argument that things would be worse without AI is hard to make because most of the people it would help aren’t the people generating the discourse.

The questions AI raises aren’t new. They are older questions about meaning, attention, authorship, and influence. The species hasn’t ever resolved them in a prior domain. It’s unlikely to resolve them now. AI is just the current surface where the old questions are visible. The people writing as if these questions arrived with the technology are either young enough not to have noticed the older versions of them, or invested enough in the current panic to pretend they don’t.

I’m with stoopid. By which I mean, I’m with the thing that has already been labeled before being looked at. I’m with the position that says the problems are ours and have been ours, that the tool is an occasion for the honest conversation rather than the subject of it, and that what we owe one another, right here, in the room, in the comment thread, and in the conversations we are having today. That is the practice that will turn out to have been the only thing that mattered. The rest is costume. We can do better. We have all the tools.