Open Twitter on any given morning and the timeline is wall-to-wall with the same panic: anyone can write anything now, build anything, generate anything, so creative work is finished.1 Koe starts from the same fact and walks off in the opposite direction. Yes, he says — when a machine can produce everything instantly, everything it produces becomes a commodity, worthless, not worth paying for.2
Here's the turn. When the output floods to zero value, the thing the output can't supply spikes in value.3 The flood doesn't drown the human. It makes the one spring that still tastes like a person the most expensive water in town. Commoditize the content, and you raise the premium on whoever's behind it.
The mechanism is a price seesaw. When the supply of something goes infinite and free, its price falls to nothing — that's the content side, and AI does it overnight.4 But that fall pushes the other end up: whatever is genuinely scarce, that a machine can't mass-produce, gets bid higher precisely because everything around it got cheap.5
What's on the expensive end? Koe is specific: the human perspective, the point of view, the curation, the taste, the story, the followed trajectory.6 Not the words — five billion beautiful words can be generated for free. The scarce, paid thing is which words a particular human chose to keep, and why anyone trusts that human's choosing.7 Strip the jargon and it's one sentence: the writing got cheap, so the writer got dear.
Picture two columns at a market stall. One column is "things you can get anywhere" — they're free, because scarcity set the price and there's no scarcity left. The other column is "things only this one person has." AI doesn't empty the second column; it empties the first, and by emptying it, makes the second column the only place value can live.
Koe runs this against the "dead internet" objection head-on. If AI floods the internet with content, doesn't content become worthless? Yes, he answers — and then asks what that does. It increases what can be paid for something human.8 The logic is almost mechanical: value is relative, so cheapening one side of the ledger enriches the other. The human point of view isn't valuable in a vacuum; it's valuable because the machine-made stuff around it collapsed to zero, throwing the human into relief.
In the Multiple Interests video he sharpens the scarcity claim. Most things will become commodities, but attention won't — there's only so much of it to go around.9 So in a world of infinite content, the bottleneck isn't making things; it's being one of the few things people actually know about and pay attention to.10 The premium doesn't just attach to the human — it attaches to the known human.
This page supplies the economic engine that several vault pages assume but don't fully justify. You Are the Last Defensible Moat claims you're the one thing that can't be copied — this page explains why the price of that uncopyable thing goes up rather than merely staying constant. Attention as the Last Moat is the sibling that carries the attention half of Koe's argument; this page carries the value-inversion half. And it gives Essay Craft as AI-Defense Moat its market logic — craft defends you not by being un-generatable, but by being the scarce thing whose premium rose. Together these pages stop treating "be human" as a slogan and start treating it as a pricing argument.
Koe's tightest illustration is a single rhetorical question, and it's worth reading slowly because it does all the work: AI can spit out five billion words that are amazing, super-beautiful prose — but which ones are you going to select to stick around?11
The five billion words are free. They exist, they're good, and they're worthless precisely because there are five billion of them. The valuable act isn't the producing; it's the selecting, and the selecting requires a human with a point of view who can be trusted to pick well.12 That's the premium made concrete. Notice the case study doesn't show a human out-writing the machine — Koe concedes the machine writes beautifully. It shows a human doing the one thing the flood can't do for itself: caring which of its outputs deserves to survive. The premium isn't paid for better prose. It's paid for trustworthy judgment over an ocean of equally-good prose.
It's Monday and you're scrolling, and every third post is plainly AI — fluent, competent, forgettable, and indistinguishable from the one above it.13 You feel the flood. The instinct is to compete on the flood's terms: post more, post faster, match the volume.
Don't. Watch instead for the rare post that could only have come from one specific person — a take you couldn't get from a prompt, a story attached to a real life, an opinion with a spine. That one stops your thumb. Now do the un-obvious thing: instead of adding to the flood, you write the post only you could write. It's slower. It's riskier. It exposes a point of view that a generic post would hide. That exposure is the entire move — you're not trying to out-produce the machine, you're trying to be the spring in a landscape of floodwater.14 You hit publish on the thing with your fingerprints on it, and you let the commodity posts scroll past.
You can feel yourself sliding into the worthless column. The signs:
Two Koe videos, both [PARAPHRASED] and [POPULAR SOURCE]. The argument is internally consistent but rests on an unproven premise: that the value freed up by cheap content reliably flows to humans rather than to, say, platform owners or the AI companies themselves. Koe asserts the seesaw; he doesn't demonstrate that the human end is where the value lands. 🚩 SINGLE-FRAMEWORK — the value-inversion claim is intuitive but unverified as economics.
A real tension with the "dead internet" crowd he's rebutting: they could be right that content floods and the premium fails to materialize for most humans — that a tiny number of already-known humans capture the entire premium while everyone else drowns. Koe's own attention argument cuts against his optimism here. If attention is scarce and only flows to the few who are already known, then the human premium isn't broadly available; it's a winner-take-most prize.19 He presents the premium as an opportunity for "you," but his attention economics suggest it's an opportunity for very few of you. He doesn't resolve this.
His claim that promoting a theory makes it more likely true (from the same video) should also make a reader discount his optimism — a creator selling a "become the scarce human" path has commercial interest in believing that path pays.20 🚩 MOTIVATED REASONING.
Koe and David Perell sit close here. Perell's Everybody Is a Media Company argues the individual can now own distribution and become a sovereign economic unit — and Koe's premium is what makes that sovereignty pay. Perell tells you that you can be a media company; Koe tells you why being the human behind one is the asset that appreciates as AI cheapens everything else. They converge on the sovereign individual but emphasize different halves: Perell the distribution, Koe the scarcity. Hold both and the picture completes — you need to own the channel and be the irreplaceable thing flowing through it, because owning a channel that pipes commodity content is just being a smaller flood.
There's productive friction with the Taste/Judgment/Labor framework. That framework treats taste and judgment as inputs you contribute to AI-assisted work — a division of labor between human and machine. Koe's premium argument adds a price tag the framework leaves implicit: those human inputs aren't just necessary, they're the only part anyone will pay for once the labor is automated. Read together, the framework describes the workflow and Koe describes the economics of it — and the economics are harsher than the framework's neutral tone suggests. If only taste and judgment command a premium, then the "labor" third of the framework isn't just delegated to AI, it's de-monetized entirely, which reshapes how the whole division pays out.
Plain version: cheap machine output is a market-pricing story, but what about a human can't be copied is a story about creativity and selfhood — and the two only make sense together.
Handshake one — to creative-practice, Essay Craft as AI-Defense Moat. That page argues the essay — thinking made visible, idiosyncratic, voiced — is hard for AI to replicate and so protects the writer. Koe's premium argument gives that defense its economic shape. The essay isn't valuable just because it's hard to copy; it's valuable because copying everything else got free, throwing the un-copyable essay into sharp relief and raising its price. Stack the pages and a non-obvious thing surfaces: the moat isn't really the difficulty of the craft — AI is getting frighteningly good at fluent prose. The moat is the attached human, the trajectory and point of view the essay carries, which is worth more in exact proportion to how cheap fluent prose became. So the craft page and the premium page together relocate the defense: it's not "AI can't write a good essay" (it can), it's "AI can't be the trusted person whose essays you follow, and that person's premium rises as their words' raw scarcity falls." Craft is the vehicle; the human premium is the fuel — and neither the craft framing nor the pricing framing reaches that conclusion alone.
Handshake two — to behavioral-mechanics, Everybody Is a Media Company / Sovereign Individual. Perell's claim is structural: technology lets one person own distribution and become an independent media entity. Koe's premium tells you what that entity's actual asset is. Owning distribution is necessary but it's just a pipe — what determines whether the pipe is worth anything is whether commodity water or premium water flows through it. The handshake reveals a dependency the sovereignty thesis tends to gloss: the sovereign individual only stays sovereign if they're the scarce human, not a one-person commodity factory. As AI lets everyone become a media company, simply being a media company stops being a moat — the moat collapses back to the same place, the irreplaceable human behind the channel. So Perell's structural revolution (anyone can own distribution) and Koe's value revolution (the human premium rises) are two faces of one event, and together they expose the trap: the sovereign individual who isn't also the scarce human has built a free, infinitely-replicable channel — which is to say, no moat at all. The pricing logic disciplines the sovereignty hype. There's a final twist the pair surfaces that neither page volunteers alone. Perell's revolution democratizes the channel — anyone can own distribution now, the barrier fell. Koe's premium concentrates the value — only the scarce, known human captures it. Those two movements run in opposite directions, and the friction between them is the real shape of the creator economy. Distribution got infinitely available exactly as the value flowing through distribution got winner-take-most. So the sovereign-individual dream and the human-premium dream are not the same good news told twice; they're a democratization and a concentration happening at once, and a person can win the first while losing the second — own a channel, broadcast freely, and still earn nothing, because the channel is now as common as the content. Reading Perell's structural optimism through Koe's pricing lens corrects a flattering assumption baked into a lot of creator advice: that owning your distribution is the hard part. It isn't. Owning distribution is the part that got easy. Being the human the premium flows to is the part that stayed brutally hard, and the two pages together are what make that visible — the access revolution and the value concentration are one event, and most people will experience only the first half.
The Sharpest Implication. If the premium is real but attention is scarce and winner-take-most, then "be the human" is not a broadly available strategy — it's a lottery with a human-shaped ticket. The sharpest reading of Koe is darker than his framing: AI doesn't raise the human premium evenly, it raises the premium for the few humans who are already known and zeroes out everyone else, because the same flood that makes the known human precious makes the unknown human invisible. The premium and the obscurity are the same event seen from two sides.
Generative Questions.