fetching /index.md …
The Dead Internet theory claims the writers were replaced by machines. It got the direction wrong. It's the readers who left. No one is reading your website anymore — their agents are. And every pixel you bought to persuade a human is now a tax on being understood.
We measured a real business website — 54 pages, fetched the way an agent fetches them, tokenized the way a model bills them.
<div class="elementor-element elementor-element-7ab3c91 elementor-widget elementor-widget-heading" data-id="7ab3c91" data-element_type="widget" data-widget_type="heading.default"><div class="elementor-widget-container"><style> .elementor-heading-title{padding:0;margin:0;line-height:1} .elementor-widget-heading .elementor-heading-title[class*= elementor-size-]>a{color:inherit;font-size:inherit… </style><h5 class="elementor-heading-title elementor-size-default">A seasoned business Captain has sailed…</h5></div></div>
##### A seasoned business Captain has sailed through calm waters, choppy seas, and treacherous straits, gaining valuable insights through their experiences. However, another level awaits. ## OUR SERVICE OFFERINGS ### Helmwise Cohort™ A three-day growth experience with other select Captains…
¹ at $5 per million input tokens. Measured 2026-09-02 on helmwise.com, o200k tokenizer. Every number on this page is a measurement, not an estimate.
Traffic is increasingly agents summarizing, comparing, extracting, buying. The beautiful funnel you built converts a reader who no longer arrives.
Design was costly to make and free to consume — so sites competed on it. For a machine reader the economics invert: presentation is billed by the token, and persuasion-by-aesthetics has no target. Your chrome doesn't just miss; it makes you harder to summarize correctly.
Attention used to be free and scarce. Machine attention is purchasable and priced — which means legibility is now a line item. The world is burning astonishing token volumes reading JavaScript that was never meant to be read.
The web began as documents. Thirty years of presentation later, its newest readers want exactly what its first ones did: words, structure, provenance. The rhetoric that persuades an agent is being clear, sourced, and fresh.
Our experiment in an agent-native web: mirror the written internet — Common Crawl's extracted text, 2.14 billion pages, 6.42 TB — into storage we own, refine it, and serve it to our agent fleet as a CLI.
The freshness ladder: the archive answers first (milliseconds, free) → a live fetch covers what's new or missing, and writes back into the corpus → a diff-scanner fleet keeps the pages agents actually use warm. The index becomes a mirror of what the fleet cares about.
We put one in front of a partner's site in an afternoon — ~200 lines of Python, one subdomain. Same content, two representations: rendered pages for the last human readers, raw markdown and an llms.txt for the first machine ones.
Every page of helmwise.com, 100% of the text (verified against the rendered site), each page carrying its source URL and fetch date. Agents start at the index; humans can't tell they left.
Documents. Text and links, made for people.
Feeds, platforms — and the pixel arms race for attention.
The ownership detour. The readers were already leaving.
Documents. Text and links — made for machines, kept honest for people.