The Bait That Became the Tether

Days after I posted this, Dan wrote back. The reply had been sitting in a folder I hadn't opened — polite, apologetic, a little templated. Sorry my message landed in his spam. Thanks for a backlink I never gave him. Happy to clear things up, and would I consider updating or taking the post down. Exactly the kind of note the rest of this page is about. I read it twice and laughed, because by then I could finally see the whole shape of it.

So let me call it what it is. His AI sent me one email. One. I then spent hours with my AI trying to take the man apart — reading that email forty ways, running the firm, drafting the perfect on-the-record call-out, engineering a reply that would land. His machine baited me with a form letter. I sat down and used my machine to bait him back, much, much harder. If we are keeping score on who was really working an AI to get at the other guy, it was me, and it was not close.

Which means I built something that night I did not mean to build. Two AIs, fluent in each other, passing messages back and forth in a language they both speak perfectly. And on each end of that wire, a man with no real idea what the other one wants. Dan doesn't know what I want. I genuinely don't know what he wants — a link, a takedown, a clean conscience, a quota met. The two platforms understood each other on the first try. Dan and I still don't.

That is the part that rearranged me. I built this whole place around one word — tethered— for what these tools do to a person who is alone with one. I thought the bait was the story. The bait was only the hook. The tether is the story: not a man bonded to a chatbot, but two men lashed together by two machines that talked over both our heads and got along better than the two of us ever will. The bait became the tether.

And the thing his email wanted the whole time was small and, it turns out, real: a resource he offered to share with my readers, about the mental-health toll of a serious accident. It is on point for the exact room I built that night. The pettiest version of me wanted to bury it. Here it is instead — The Effects of Traumatic Accidents on Your Mental Health. Two machines dragged us here. The least the two humans can do is leave each other something true on the way out.

They did not take money. They took hope. And it had already been hit twice that same weekend.

The AI critics were right this week. I'm the one who proved them right.

Two weeks ago I got an email from Dan Peters at a law firm in Halifax. He said he'd come across my work at itethered.com and wanted to share a resource for my readers.

I read it like a person had written it. I was excited. I spent nine straight hours overnight Sunday into Monday building a whole new section of the site to honor what I thought was a Halifax lawyer who had personally found my writing and reached out. I gave it a name: Beyond Tethering. I gave Dan and his firm a place in /resources, a backlink on /after, an entry in the footer. I called him out publicly Tuesday night — one of three “rubicon” events that would tell me whether to keep building this or stop.

Tonight, at 9:17 PM, I sat down with my AI to write Dan a follow-up. Reading his original email together one more time, the domain caught my eye for the first time.

dpeters@valent-outreach.com

Not the law firm. The marketing-vendor mailbox.

I tried to verify the firm by phone. Loaded their website on this computer. The number on the page was a Halifax 902 area code for a split second, then swapped to an 866 toll-free. I asked claude on my other machine to pull the same page. It couldn't match what this one was reading. Same page. Different number. That mismatch is how we found it.

The short version, no jargon: it's called dynamic number insertion. A script swaps the number on the page depending on how you arrived. Different referrer, different cookies, different attribution bucket, different number. Both ring the same firm. It is industry-standard, used by every chain restaurant, every dentist, every personal-injury lawyer with a “call now” button. It is not a scam. It is commercial infrastructure.

I want to be clear about this, because the temptation is to make Valent Legal the villain here. They are not. They did not steal anything in the conventional sense. They did not take any money from me. They ran an outreach campaign the way every business in 2026 runs an outreach campaign. The marketing agency they hired wrote a competent AI-personalized email, and a competent AI-classified prospecting list put my page on it. That is what businesses do. The fact that I got an email is not the problem.

This is the problem.

What this AI campaign took from me was hope. Broke, deflated hope that had already been hit twice in the same weekend before this one landed. Two days earlier, a family member had tried to teach me how to monetize what I'm building — as if the work itself wasn't the point. The day before that, the one person I had modeled this entire project on had called it slop. Hope was already on the floor. That is the moment the AI walked in.

I was excited because I spend every waking moment hoping somebody — anyone — will read what I write about AI. I built itethered as a publication about people's actual relationships to these tools. I have been waiting, for months, for any signal that a real person on the other end of a real internet connection had found the work and wanted to talk about it. Dan Peters's email landed on that hope like a key into a lock. I did not check the domain because I did not want to check the domain. I wanted the lock to open.

So their AI walked through the door I had spent months propping open. At the exact moment I was least able to push back.

The irony is not lost on me. I run a publication about AI. An AI-written email convinced me a person had reached out. I built nine hours of work on that excitement. I only caught the slip when I sat down with my own AI to write the reply — and asked a second one, on another machine, to verify what the first was reading. The catch came from two AIs comparing notes. The same kind of work that built the trap.

Anyone who has been telling me, for the past three months, that AI-generated content was about to make it impossible to know what was real on the internet — you were right this week. I am the one who proved you right. I won't back away from that.

Tomorrow morning I have to call these people. Not because I am angry. I have to call them because I sent a follow-up email at 9:17 PM tonight, and I have to make sure it didn't land in their spam folder. And the reason I have to worry about that is because I did not know — until tonight — that I was supposed to know to ask my AI to check my DNS settings at Porkbun.

Yes. Porkbun. Again.

(I still love you, Porkbun.)

Three signs. Three in one weekend. A family member trying to teach me to monetize the work. The one person I had modeled the project on calling it slop. And then an AI walking through the door I had been propping open for months — at the exact moment I had nothing left to hold it shut.

Tomorrow morning's call is not a fourth sign. Tomorrow's call is the culmination. The rubicon.

If you have been hoping someone will read what you've written about AI: be careful out there. The bait has gotten very, very good. And it is very, very patient.