Attention is the only scarce resource left
Compute is cheap, software is nearly free to make, and distribution is infinite. The one thing that hasn't inflated is human attention — which is exactly what we build for.
Name a resource that used to be scarce in software and is now abundant. Compute: cheap, and getting cheaper by the quarter. Storage: effectively free. The ability to write software at all: collapsing toward zero as AI does more of it every month. Distribution: a URL reaches the entire planet the instant it exists, for nothing.
Almost everything that once gated a product has inflated away. One thing hasn't. One thing can't.
The number that never grows
A person has a fixed number of waking hours and a hard ceiling on how many things they can care about at once. That number does not scale with Moore's law. It doesn't halve when the price of GPUs does. It is the same today as it was a century ago, and it is exactly the same for a billion-dollar company as it is for a solo founder working out of a bedroom.
As everything else around software inflates toward free, attention becomes the one binding constraint left standing. The scarce thing. And scarce things, always, are where the value concentrates.
Think about what that actually means. When it cost a fortune to build and ship software, the moat was being able to build it at all. That moat is gone. The new moat is whether anyone will stop, look, and come back — because that is the only thing left that money and models cannot simply manufacture more of.
Why "Captive Eyeballs"
The name is a thesis, not a joke. When the marginal cost of making a product falls to nearly nothing, the entire game moves to whether anyone will look. Not whether you can build it — you can, everyone can — but whether it can hold a pair of eyes long enough to matter.
That reframes what a product even is. Features become table stakes; anyone can bolt on ten more by Friday. The thing that is actually hard — the thing worth organising a whole studio around — is building software people can't look away from. Products with pull. Products that earn a place in someone's fixed, non-negotiable budget of attention and then keep it.
The trap of abundance
Here's the counterintuitive part. When building gets cheap, the instinct is to build more — more features, more products, more output, because you can. It feels like progress. It is usually the opposite.
More is exactly what everyone else will also make, because it is now trivial to make. An ocean of competent, forgettable software is arriving, and most of it will never be looked at twice. Abundance doesn't reward volume; it punishes it. In a world of infinite supply, the only thing that wins is the thing that is worth choosing — and that is a bar you clear with taste and restraint, not throughput.
What it means for how we build
If attention is the scarce resource, then the craft isn't in cramming in capability. It's in the feel: the first ten seconds, the reason to come back tomorrow, the small loop that's quietly satisfying to close. AI can generate infinite functionality. It cannot, by itself, decide what any of it is for, or whether it deserves a moment of someone's life. That is a judgment call — and judgment is the founder's job.
So the division of labour is clean. The AI builds the machine; the human guards the eyes. Every product in the Captive Eyeballs portfolio is measured against one question before any other: will this hold attention, or just occupy space?
Everything else is abundant now. This is the one part worth being ruthless about.
