Thinking
On trust infrastructure, portable proof, and making intelligence real. Subscribe via RSS.
Target Fixation: Why AI Chatbots Can Make People More Stuck
Target fixation collapses a rich field of signals into a single line. This essay connects skydiving, AI psychosis, and chatbot sycophancy to explain why people can get pulled deeper into harmful loops.
Why Seven Keeps Appearing in Biology, Music, and Proof
Why does seven keep showing up in vertebrae, musical scales, working memory, and LAKIN receipts? A structural argument from geometry, cognition, and Assembly Theory.
The Survey Is Not a Receipt
A survey asks how you felt about what may or may not have happened. A receipt records what actually happened — and that difference is the missing accountability layer in maintenance, service, and coordination.
Receipts Are Proof of Assembly
Assembly Theory measures whether something was built or happened by chance. Receipt data applies the same principle to coordination — revealing patterns no other dataset can show.
The Guy at the Window
Surveillance moved from outside the window to inside the device. Your record of your own life should live on your device — not on someone else's server.
Are You Real?
Your professional reality is platform-dependent. Remove the institutions, and you can't prove your own history. Receipts change that.
The Cost of Real
It is cheaper to make something up than to prove something happened. LAKIN exists to change that equation — just as clay tablets did for the ancient world.
The Three Shapes of Trust
Triangle, square, circle — aim, wall, story. The minimum structure where ambiguity collapses and trust becomes provable.
Friction Is Testimony
Friction — the resistance from independent systems — is the only reliable testimony that something actually happened. AI removed the last natural friction. Here's what replaces it.
AI-Generated Resumes Are Indistinguishable From Real Ones. Now What?
AI can generate perfect, tailored resumes in seconds. The resume as a trust signal is over. Credentials need receipts, not stories.
The Status Update Problem: Why "On Track" Means Nothing
Status updates are the most common form of business communication and the least verified. Here's what "on track" should actually mean.
AI in Financial Compliance: Why Summaries Aren't Evidence
AI-generated compliance documents look identical to verified ones. Financial institutions need verification infrastructure, not better summaries.
Why AI Detection Is a Losing Game
AI detection accuracy is degrading with every model improvement. The question that survives isn't "who made this?" — it's "did it touch a wall?"
Touched vs. Untouched: The Only Question That Matters in the AI Era
The human-vs-AI binary is decaying. The distinction that lasts is whether a claim touched an independent system and came back with proof.
What Is AI Slop?
AI slop isn't about quality — it's about whether output ever touched reality. A new definition for the defining problem of the AI era.