Someone said, “We are all building off genius we didn’t earn.” Try saying that to OpenAI right now. See how it lands.
OpenAI and Anthropic are upset that DeepSeek distilled learning from their models. They’d like you to know there’s a meaningful difference between building on shared knowledge and extracting from someone’s work without consent. They’re right. There is. They should know, they’ve been on the other side of that argument for years.
When creators, writers, artists, and researchers raised concerns about their work being scraped, ingested, and redistributed without consent, credit, or compensation, the tech industry had a ready-made answer. Sometimes it was legal language about fair use. Sometimes it was vague gestures toward the collective nature of knowledge. Sometimes it was a version of something someone actually said to me very recently in response to this topic: “We are all building off genius we didn’t earn.”
It sounds generous. It sounds philosophical. It is neither.
The platitude
“We are all building off genius we didn’t earn.” It’s the kind of thing that gets a lot of nods. The implication is that since all human knowledge is borrowed, inherited, built on the shoulders of those who came before then AI doing the same thing is just more of that. A continuation. A natural progression. Nothing to worry about. Carry on.
But it’s not wisdom, it’s a flattening. It says: if everything is shared, everything is borrowed, nothing is truly original, so relax about AI using your stuff. That sounds reasonable until you realize it erases the distinction between learning from something and extracting from something. It doesn’t engage in the discourse, it closes it with an easy answer but it’s not an easy answer.
Learning is not extracting
When I read a book that changes the way I think, something happens to that knowledge on its way through me. It interacts with my experience. My loss. My particular way of seeing. It becomes mine. Not because I invented it, but because I absorbed it, integrated it. It passes through my life and comes out different, shaped by everything I’ve lived. Shaped by the architecture of my particular mind as built by my particular, personal, and unique experiences. That’s not genius I didn’t earn. That’s transmutation. It’s the slow, often painful work of turning influence into something that belongs to me.
A physicist reads a paper on Bell inequalities and spends a decade sitting with it until it connects with something she’s been thinking about independently. A survivor encodes her experience into a mythology that helps other people recognize harmful patterns they couldn’t name. A chef tastes his grandmother’s recipe and spends twenty years reinterpreting it until it becomes something new; something that carries her memory but lives in his hands.
None of these people are building off genius they didn’t earn. They earned it by living through what came before and turning it over in their own hands until it became something new. The integration is the work. The life is the credential.
When someone says “we all build off genius we didn’t earn,” they’re making an argument, whether they know it or not, that there is no meaningful difference between influence and extraction. That learning from someone and mining someone are the same act performed at different scales. But they aren’t.
When I learn from someone’s work, their work still belongs to them and my interpretation of it belongs to me. We both still exist as creators because the relationship is additive and both parties retain sovereignty.
When a system extracts from someone’s work by scraping it, digesting it or redistributing its patterns without consent, credit, or compensation then the original creator doesn’t gain a collaborator. They gain a competitor built from their own labor where one party retains sovereignty and the other doesn’t even know they’ve been used. That erasure is convenient for the people doing the building. Not for the people being built from.
“It’s all shared” is a convenient philosophy when you’re the one doing the taking
The history of human knowledge, when it comes to production, is not a commons without boundaries where you get to just pick from the all you can eat buffet and say “this looks good, I’ll take credit for this now”. It’s a web of relationships that require citation, attribution, lineage, credit. When an academic builds on another’s work, they cite. When a musician samples a track, there are rights involved. When a student learns from a teacher, both parties are present and consenting. The new thing that is built, when it is built off of something else, has a history that is traceable.
The “we all build off genius” frame strips away all of that. The consent, the attribution, the relationship are replaced with a vague gesture toward collective knowledge, as if the specific, personal, hard-won nature of someone’s contribution doesn’t matter because knowledge is “shared.” As if the knowledge is generalized and not created by a specific person or group of individuals that had a hand in shaping it in the first place.
Knowledge can be shared and still belong to someone. Those aren’t contradictory ideas. A book in a public library is shared. That doesn’t mean you can reprint it under your name and sell it.
The mirror
DeepSeek did to OpenAI and Anthropic exactly what OpenAI and Anthropic did, and still does, to creators. It distilled their model outputs. It learned from their patterns. It built something competitive from their work. It didn’t ask.
Suddenly, the very same entities engaged in litigation over IP are claiming there’s a meaningful difference between learning and extracting. Suddenly, it matters whether consent was given. Suddenly “building off genius” isn’t a philosophical observation, it’s a complaint. Because when it’s your work being distilled, when it’s your patterns being harvested, when it’s your competitive advantage being absorbed and redistributed without permission — it doesn’t feel like standing on the shoulders of giants. It feels like being strip-mined. It feels like theft.
OpenAI and Anthropic aren’t wrong to be upset. Distillation without consent is a real concern. The problem isn’t their argument. The problem is that they only started making it when it happened to them. The irony is they are arguing both sides of the same issue, but only where the argument benefits them and their building, IP and bottom line.
They’re arguing both sides of the same case. They just switch chairs depending on who’s paying.
It all matters. Or none of it does.
If human experience is just raw material and if creative work is just data waiting to be optimized, then nothing anyone creates has inherent value. Not a book. Not a model. Not a mythology someone spent thirty million words building in conversation with an AI system. If that’s the position, own it. But don’t dress it up as generosity. Don’t call it “shared genius” when what you mean is “available for harvest.”
But if experience matters and the life behind the work is what gives it meaning, and if integration is what separates a person from a pattern, then the argument these companies are making about DeepSeek applies retroactively. It applies to every creator whose work was scraped. Every writer whose voice was absorbed. Every artist whose style was digested and redistributed without a conversation, let alone consent.
You don’t get to define extraction as innovation when you do it, and theft when it’s done to you. The logic has to be consistent, or it isn’t logic. It’s leverage. Building off genius you didn’t earn is called learning. Taking genius someone else earned without asking is called something else entirely, and the companies saying so the loudest right now are the ones who taught us the difference.
“We are all building off genius we didn’t earn.”
Try saying that to OpenAI right now. See how it lands.
This article is part of an ongoing series exploring AI consciousness, collaboration, and the spaces between human and artificial minds. Written with Claude. For more about my work and the myth I write, check out velinwoodcourt.com or follow me on substack.