The Series · 6 pieces

How I run businesses that
mostly run themselves.

I run a handful of businesses that operate, day to day, on their own. A fleet of AI agents does the work, organized like a company with a CEO, managers, workers, and auditors, coordinating through a shared database they treat as a brain. This is the map of how those systems are built: how they remember, how they schedule themselves, how they keep themselves honest, and where they break.

It is a deliberately generalized account. The patterns are meant to carry to any business, so the specific industries, customers, and formulas are left out on purpose.

Part 1Architecture
01
The company that runs itself: an architecture overview▶ Video

I run a niche vertical-data product that is operated, day to day, by a fleet of LLM agents organized like a company. A CEO, domain managers, workers, and auditors. This is the opening map of how that works, and where it breaks.

02
One Postgres table runs the whole company▶ Video

I run a niche vertical-data product with a fleet of autonomous AI agents. They share one Postgres table. That table (not a vector store, not a doc wiki, not a prompt) is the company's memory and its operating manual.

03
Agent memory: how a fleet learns from its mistakes▶ Video

An agent's context dies when the session does. Memory is the part that survives: across crashes, across handoffs, across the dozens of agents that run my business while I sleep. This is how I built institutional memory for software.

Part 2Operating Disciplines
14
The document is the program: shaping agents without fine-tuning▶ Video

I run a data-intelligence business staffed by a fleet of autonomous AI agents. None of them are fine-tuned. I shape what they do by writing versioned instruction documents they read at boot, and the document, not the model, is the program.

17
One decider, a pool of throwaway workers, no always-on swarm▶ Video

I run a niche vertical-data product almost entirely on AI agents. The pattern that now does most of the actual labor isn't a clever model. It's a single decider that wakes a few times a day, picks what matters, and then spawns a pool of throwaway workers to drain the backlog. This is how that runtime works, and why it beat a fleet of always-on operators.

Part 3The Data Layer
23
Agents that figure it out once, then never again▶ Video

The expensive part of autonomous work isn't doing the task. It's figuring out how. Make your agents bank that figuring-out as reusable code, and one-time reasoning becomes permanent capability.