Decision-Making

Today I rebuilt the routing rules for my AI agents: which model handles mechanical work, which handles hard reasoning, which validates the output before I see it. The setup looked rigorous. Every substantial piece of work got checked by a validator agent running at high effort on my most capable model. The critique that stopped me: the validator was the same model as the thing it was checking. Same model, same characteristic errors. When the producer and the checker share a mind, the check confirms the work is consistent with itself. It can’t see the errors that mind is inclined to make, because it’s inclined to make them too. ...

In August 2025 I wrote a thought card about structured data. The argument was simple. Treat your data as a long-lived asset that should be findable by any system, and you build infrastructure that compounds. The card said nothing about agents. I had no idea WebMCP was coming. Six months later, Google announced WebMCP and the agentic web, the standard protocol for AI agents to act on structured web data. A protocol I’d never heard of, designed for a use case I hadn’t predicted. And the platform I’d been building already fit it. Not because I saw the future. Because the principle didn’t depend on seeing the future. ...

Most leaders don’t actually leave doctrine behind. They leave the way people remember them leading. Mission statements claim transparent, equitable, research-led, open. The qualities show up in the language. They rarely show up in the practice. The honest version of the legacy claim is smaller and harder. It’s not “I built a great system.” It’s “I knew the way my own thinking gets stuck, and I built a tool that breaks me out of it.” ...

I was working through the details of bringing a small team in-house. Purchase price, salaries, org chart, who reports to whom. Normal acquisition stuff. And I caught myself thinking about it like a hiring decision. How many people, at what cost, doing what work. But we weren’t hiring. We were acquiring. The distinction matters, and I think it matters more in the auction industry than most people realize. When you hire someone, you’re renting their skills for as long as they stay. The moment they leave, whatever wasn’t documented or embedded in a system walks out with them. That’s not cynical, it’s structural. I wrote about this a few weeks ago in a different context: the relationships someone builds on your platform become yours, but the person is always rented. Everyone is. ...

The Idea This came up while fixing a bug where stale cached data was silently injecting garbage into every outbound email for a full day. The cache never questioned itself. It just kept serving bad data. Human memory has a built-in half-life for each memory. The things you use stay strong. The things you don’t fade. That’s not a bug, it’s what keeps the signal-to-noise ratio manageable. You naturally surface what’s relevant because relevance reinforces the memory. ...

The Observation My cousin texted me this morning with a line I’ve heard before but landed differently today: “A ship in a harbor is safe, but that’s not what ships are for.” It’s attributed to John A. Shedd, published in 1928. One of those quotes that gets printed on coffee mugs and LinkedIn posts until you stop hearing it. But context matters. He didn’t send it as decoration. He sent it as a nudge. ...

The Observation I didn’t build structured data systems because I predicted WebMCP or the agentic web. I built them because “touch it once, make it findable by any system” is just good practice. “Computers do the mundane, humans critically think” is just a sound division of labor. These are principles, not predictions. But those principles, arrived at independently, produced exactly the infrastructure the agentic web needs. A straight line runs from a thought card I wrote in August 2025 to a platform that’s accidentally ready for a protocol Google announced six months later. ...

The Box There are five of us in a rental car heading from the airport to the hotel. A friend is talking about a book. Nikki Erlick’s The Measure. The premise is simple: one morning, everyone on Earth wakes up to find a small wooden box on their doorstep. Inside is a string. The length of the string is the length of your remaining life. The question the novel asks is the obvious one. Do you open it? ...

Why you’re accountable for warning, not for results, and why silence isn’t neutral Ezekiel was a priest with no temple to serve in. He was thirty years old, living in exile by a canal in Babylon, watching his homeland crumble from a distance. The first wave of exiles had been taken in 597 BC. Jerusalem hadn’t fallen yet, but it would. His purpose had been stripped away. And then God showed up, not in the holy place, but in enemy territory. ...

Why the obvious hides in plain sight The Chinese surveillance balloons floated across American airspace for years before anyone noticed. Not because we lacked sensors. Not because we lacked data. The U.S. defense system collects more information than any surveillance apparatus in human history. The balloons went undetected because of something far more dangerous than technological failure: we weren’t looking for them. I heard this story recently on the Jordan B. Peterson podcast from Garry Nolan, a Stanford immunologist who’s become one of the more credible scientific voices in UAP research. He explained how our defense sensors are programmed to filter for known threat signatures (rockets, planes, missiles). Everything else gets discarded immediately as noise. The system processes what it expects and throws away what it doesn’t. ...