← Collected essays

On AI · · 3 min read

AI Is Not a Tool. It Is a Limb.

On the quiet stratification of human capacity, and the music most people don’t even know is playing.

By Chris Wyatt · Collected edition, September 2026

Download AI Is Not a Tool. It Is a Limb. (PDF)
Contents

Two people walk through the same grocery store. One hears the chord progression of the muzak playing overhead and can name the key. The other doesn’t even register that there’s music. Both have working ears. The difference isn’t in the sense. It is in what the sense has been wired to.

That’s roughly where we are with AI, and it is the thing nobody is naming correctly.

The current discourse treats AI as a tool. A better calculator, a faster search, a clever writing assistant. Companies measure “AI adoption” the way they used to measure spreadsheet adoption, as if the question is whether you click the button. This framing is going to age badly. It is wrong in the same way that calling literacy a “writing tool” is wrong. Yes, technically. But it misses what is actually happening to the species.

Knowledge has been getting closer to the human animal for as long as we’ve been an animal. Oral tradition kept it inside people. Scrolls and libraries made it external and addressable. Printing made it replicable. The internet made it queryable. AI makes it generative and conversational. The next step, whatever it is, will probably make it ambient and predictive. Each transition didn’t just shrink the distance to knowledge. It changed the kind of mind that could hold it.

We are in the middle of one of those transitions right now. And the people inside it are already starting to look different from the people outside it.


“Tool” doesn’t capture this. “Sense” doesn’t either, because a sense is roughly equal across the species. The cleanest frame comes from evolutionary biology: extended phenotype. The beaver’s dam is part of the beaver, even though it is external. The spider’s web is part of the spider. They are not accessories. They are constitutive of the animal.

When you talk to someone running a real agent stack, you aren’t talking to them. You’re talking to them-plus-agents. The boundary of the person has moved.

This produces a stratification that nobody is mapping well, because everyone is mapping the wrong axis.

The popular axis is access. Will the rich have better AI than the poor? Probably, at the margins. But frontier models are collapsing in price. The bottleneck isn’t the model. It is integration, and integration is a habit, not a purchase. Three tiers are emerging.

TIER ONE

Treats AI as a fancier Google. Ask, answer, close tab. Productive but not transformational.

TIER TWO

Treats AI as a collaborator. Push back, iterate, use it to sharpen reasoning. Most thoughtful professionals have reached this tier.

TIER THREE

Treats AI as infrastructure. Agents running in the background, persistent workflows, the model woven into the bones of how the person operates.

The gap between tier two and tier three is the one that will matter. Tier three isn’t a productivity story. It is an identity story. People at this level don’t “use” AI any more than you “use” your visual cortex. It is how they see, how they decide, how they show up. They are not faster than tier two. They are a different kind of entity.

You will know them by their output. It looks like the work of someone with a team of six. There is no team. There are agents. The dam is doing work.


The fair counterargument is that this is transient. “Internet expert” was a category in 1998 and a punchline by 2010. The tools democratize, the gap closes, and the early movers stop looking special. That is possible. But I don’t think it is right this time, because the limb itself compounds. The person who designs an extended cognition for fifty agents today is in a better position to do it for five hundred next year. The skill is the moat, and that doesn’t plateau the way “knows how to Google” did. It accelerates.

Back to the grocery store.

The interesting thing about tier three isn’t that they have more information. It is that they perceive more. They hear the chord progression. They notice the structure underneath the noise.

The species is growing a new limb. It is unevenly distributed, and it is going to stay that way for a while.

The interesting question isn’t who hears the music. It is that most people don’t even know it is playing.