Sixty Machine Reads a Day

TL;DR: New anchor manual: The Free Front Desk, the full playbook for getting found through AI search, with my own server logs as the receipts. Free, long, no email wall.

I went digging in my access logs this morning. The raw ones, the kind nobody reads because a dashboard usually reads them for you.

In the first ten days of July, AI assistants pulled pages off this site for live human users about fifty times a day. Claude fetched around 350 pages for people mid-conversation. ChatGPT, about 170. Add the answer-engine crawlers and it’s sixty machine reads a day, every day, on a one-man static network running off a five-dollar host.

And in six weeks, eleven humans clicked through from inside an AI app.

Sixty a day against eleven in six weeks. Every publisher on earth reads that gap as a funeral. To me it’s the cheapest front desk I’ve ever staffed, and the new manual argues it out loud: where the advice industry is wrong about AI search, why zero-click only kills you if your product is the click, and every move I ran on this network to get machine-readable, none of which costs a dime.

It also tells you what didn’t work, because the receipts cut both ways. My llms.txt file got fifteen fetches all month and not one came from an answer engine. It’s in the playbook anyway. Cheap bet, zero printed next to it. You deserve the numbers somebody selling a GEO course would leave out.

The platforms spent fifteen years renting us our own reach. The machines don’t care about posting cadence, engagement rate, or the checkmark, only whether the page answers the question, and I think that’s the death knell of the walled gardens. I’d rather be early than well-rested.

Full manual: The Free Front Desk: An Indie Creator’s Playbook for AI Search.

The machines will read this post tonight. Whether you do is the part I can’t automate.