A moat is the water around a castle that the attackers can't cross. In a world where AI can copy your writing, clone your app, and flood the internet with infinite content, almost every wall around your business is about to get scaled. Dan Koe's claim is that one moat still holds, and it holds for a reason no amount of compute can fix: there's only so much human attention to go around, and a machine can't manufacture more of it.1
The counterintuitive core has two halves. First, attention is a moat because it's scarce — and scarcity is the one thing AGI can't dissolve. It can make survival cheap and content infinite, but it cannot add hours to the day or eyes to the species.2 Second, the thing everyone calls "saturation" is a myth in reverse. The internet isn't overflowing with too much good content fighting for too little attention. It's the opposite: a flood of content most of which you skip, which means there's an attention deficit — a shortage of stuff actually worth paying attention to.3
Strip away the AI drama and the claim is an economics one. A moat is anything durable that protects your business from competitors. Most moats are made of things AI is about to commoditize — code, copy, design, information. Attention is different because its supply is fixed by biology, not technology.4
Koe's chain is tight. To make money from what you do, you need people to buy it. For people to buy it, they need to know it exists. Knowing-it-exists is attention. And attention "is one of the last moats because it's very scarce."5 The mechanism: when everything else becomes a commodity — when "anyone can write anything or build anything or create anything" — the only thing left that can't be infinitely supplied is the finite pool of human notice.6 Whoever holds a share of that pool holds the scarce resource everyone else is fighting over.
The Future of Work transcript sharpens it into a single line: "attention is the ultimate leverage."7 Not a moat among moats — the leverage that everything else routes through. You can have infinite production capacity and it buys you nothing if no one is looking.
The strongest objection to all of this is "but everything's saturated — there's too much content already." Koe dismantles it by asking you to look at your own behavior.
Count your day. You skim past hundreds of posts. You skip most of the YouTube videos in your feed. You read maybe one long-form piece. Of everything you consumed today, "there's maybe 1 or 2 things, if that, that actually change your life."8 That pattern — mountains of content, almost none of it landing — isn't saturation. It's a deficit. There's a "massive deficit of valuable and beneficial and useful content," because most of what's produced "just sucks."9 Saturation would mean the good stuff is crowded out; the deficit means the good stuff barely exists.
The Future of Work transcript runs the AI version of the same logic. When AI floods the space with mediocre content, "it doesn't matter because there's not enough attention to pay attention to all of that. And if it's mediocre, then... people don't want it and they're going to shift their attention to what matters."10 More volume doesn't crowd you out — it raises the value of being the thing worth not-skipping. The flood makes the signal scarcer, and scarce signal is exactly the moat.
And it's not winner-take-all, which is the part that keeps the moat from being a lottery. "This isn't a winner takes all market. Almost nothing is."11 Attention shifts; anyone can go viral; people who build slowly and charge premium prices "make more than people with millions of followers."12 The moat isn't one castle that wins — it's that holding any genuine share of finite attention is defensible, because that share can't be cloned the way a product can.
This page is the reason the other distribution pages matter. They tell you to build attention; this one proves the attention is worth building because it can't be taken from you the way other advantages can.
It gives Own Distribution vs Manual / Bot / Borrowed its economic justification — owned reach is a moat precisely because attention is the scarce thing, and you own a slice of it. It completes Every Business Is a Media Business by explaining why the attention that media generates is defensible and not just nice to have. And it directly extends Esoteric Interests as an Irreplaceability Moat, which locates the moat in what you're uniquely interested in; this page locates it in the attention that interest earns. The two moats stack: a weird, specific point of view earns a defensible share of finite attention.
The Future of Work transcript names two living proofs, and they sit at opposite ends of the same mechanism. Koe says: "just look at Elon Musk or MrBeast."13
Musk holds so large a share of attention that "his tweets quite literally shift the markets" — he attracts capital, talent, and resources through the attention itself.14 The attention isn't marketing for some other asset; the attention is the asset, and it converts directly into the things money usually buys. That's the moat at maximum scale: a pool of notice so large it bends material reality.
MrBeast shows the compounding loop. He "would reinvest what he gets from the attention back into the production, and he is slowly becoming a very powerful individual."15 Attention earns resources, resources buy better production, better production earns more attention. The moat doesn't just sit there — it can be reinvested like capital, widening every cycle.
But Koe immediately pulls the case study away from the lottery reading. He doesn't want to be Musk or MrBeast — too risky, too stressful, and not what he finds meaningful.16 The point of the two examples isn't "go be the biggest." It's that attention behaves like capital at every scale, including the small-and-premium one where someone with a modest audience charging high prices out-earns someone with millions of followers selling cheap.17 The mechanism is the same; only the dial setting differs. The moat is real whether your pool of attention is a continent or a village — because even a village is a slice of something genuinely finite.
It's 11:40 at night and you're doing the thing you swore you'd stop doing — thumbing your feed in the dark. Skip. Skip. Skip. A video about a productivity app, gone in half a second. Three reels you won't remember. A thread you scroll past without reading the second line.
Then something stops your thumb. You actually read it. You think about it after you put the phone down.
That stop is the whole business, and most nights you don't notice it. Tonight you do. You ask the question Koe's deficit argument hands you: out of two hundred things tonight, what made my thumb stop — and why? It wasn't the most produced. It wasn't the loudest. It was the one with a real point of view, the one that said something the other hundred-and-ninety-nine didn't.
That's your spec. Not "make more content" — the feed is already drowning in more. Make the thing that stops a thumb. Tomorrow morning, before you write anything, you run the skip test on your own draft: would I stop for this, at 11:40, exhausted, skipping everything? If the honest answer is no, you don't post it. The moat isn't built by adding to the flood. It's built by being the one drop in it that someone refuses to skip.
You can recognize a business that misreads the attention moat:
The two sources converge cleanly, which strengthens the claim: Multiple Interests gives the scarcity-and-deficit argument, Future of Work gives the leverage-and-examples version, and neither contradicts the other.21 When two separate Koe videos make the same structural argument independently, the core idea ("attention is the scarce, defensible resource") earns more confidence than a single-source claim would.
The deepest open question is whether attention is really non-substitutable or just currently scarce. Koe asserts "AGI can't make attention a commodity."22 [POPULAR SOURCE] But this is an assertion, not a proof. AI agents that consume, filter, and act on content for us could change what "attention" even means — if my agent reads everything and surfaces the three things worth my time, the scarce resource shifts from human attention to agent attention, and the moat may move. Koe doesn't address this; the claim that human attention stays the bottleneck is the load-bearing assumption the whole moat rests on, and it's contestable. [CONTESTED — needs corroboration]
A 🚩 motivated-reasoning flag: "attention is the ultimate leverage, become a creator" is, again, a conclusion that grows the audience for Koe's creator products. The argument is sound on its own terms, but the alignment between the advice and the advisor's interest is worth naming.23 There's also a survivorship-bias risk in the Musk/MrBeast examples — pointing at the two biggest winners to prove the mechanism quietly ignores the millions for whom the same strategy went nowhere, even though Koe partly inoculates against this with the "not winner-take-all / small premium audiences win too" caveat.
Koe and Esoteric Interests as an Irreplaceability Moat are building the same wall from two sides. Jack Moses' page locates defensibility in what you uniquely care about — the weird, specific intersection of interests no one can copy. Koe locates it in the attention that earns. The convergence is that both reject product features and skills as durable moats and both reach for something harder to commoditize. The tension is instructive: Moses' moat is about supply (the rare thing you make), Koe's is about demand (the finite notice you capture). Held together, they reveal that an esoteric interest with no attention is a treasure nobody's looking at, and attention with no esoteric core is a crowd that'll wander off the moment something shinier scrolls by. Neither moat holds alone; the irreplaceable interest earns the defensible attention.
Against Perell's Paradox of Specificity, there's a sharp convergence Koe doesn't name but clearly assumes. Perell's paradox: the more specific and narrow you go, the broader your appeal becomes, because specificity is what makes a thumb stop. Koe's "skip test" is the paradox in action — the content that survives the deficit is the specific, opinionated, polarizing piece, not the broadly inoffensive one. The tension worth holding: Koe sometimes talks about attention as if scale (Musk's millions) is the goal, while Perell's paradox says the route to scale runs through narrowness. The reconciliation is that you win the finite-attention game by going more specific, not more general — the moat is crossed by the few who care intensely, not the many who care mildly.
Plain version: this idea — that attention is the one moat AI can't dissolve — connects to three others: the irreplaceable-interest moat, the paradox that being more specific wins more attention, and reputation as your core asset. Each connection shows what the attention moat needs to actually hold.
To Esoteric Interests as an Irreplaceability Moat (business). That page argues your defensibility comes from the strange, specific combination of things you're into that nobody else has assembled — a moat made of who you are. This page argues the moat is the finite attention you can capture. The structural parallel is that both are searching for the one asset AI can't replicate, and both have given up on the obvious candidates (product, code, information). Stack the two and something becomes visible that you miss reading either alone: an esoteric interest is the supply side of an attention moat, and attention is the demand side of an irreplaceability moat — they're two faces of one coin. A person with a genuinely rare point of view but no captured attention has built a moat around an empty castle; a person with captured attention but a generic point of view has a castle anyone can rebuild next door and lure the crowd away. The durable position is the intersection: a weird, specific obsession that earns a defensible slice of finite notice because it's weird and specific. Scarcity on the supply side (rare interest) plus scarcity on the demand side (finite attention) compound into a moat that's twice as hard to cross — which is why Koe and Moses, working from opposite ends, keep arriving at the same fortified spot. The practical upshot: don't pick between cultivating a rare voice and building reach. The first without the second is a hobby; the second without the first is a borrowed crowd that leaves. You're building one moat with two walls, and a castle needs both.
To Perell's Paradox of Specificity (behavioral-mechanics). Perell's paradox holds that narrowing your focus widens your reach — the more specific you are, the more people feel "this is exactly for me," and the more it spreads. Koe's attention-deficit argument is the demand-side proof of why the paradox works. In a feed where everything gets skipped, the only thing that stops a thumb is specificity — a real opinion, a polarizing angle, a point of view sharp enough to register. Generic content drowns in the deficit; specific content is what's scarce, and scarce is what gets attention. The parallel: both say the path to broad attention runs through narrow specificity, not broad appeal. Stack them and the real lesson lands: "saturation" is the fear that makes people generalize to seem appealing to everyone, and generalizing is the exact move that guarantees the skip. The deficit and the paradox together flip the intuition completely — the crowded feed should push you narrower, not broader, because narrowness is the only thing the deficit leaves room for. The moat isn't crossed by appealing to more people; it's held by being unmistakably for some people, intensely. And there's a second-order point hiding here: the people who feel "this is exactly for me" are also the people who defend what they found, who share it, who pay premium for it. Specificity doesn't just win the skip test once — it converts a passive viewer into an active carrier, which is how a small slice of finite attention compounds into the slow-built premium audience Koe says out-earns the millions-of-followers crowd. The narrow moat isn't smaller; it's deeper.
The Sharpest Implication. If "saturation" is actually a deficit, then the entire emotional logic of "I shouldn't start, it's all been done" is built on a misread of your own feed. The market isn't full of good content fighting for scraps of attention — it's full of skippable content, and the scarce, defensible position is being the thing worth not-skipping. The barrier to entry isn't crowding; it's quality of point of view.
Generative Questions.