How to actually run a company with one AI operator.
A real business where a single AI agent handles the day-to-day — email triage, customer channels, scheduled checks, engineering work — with a human holding judgment. We publish exactly how it works: a daily blog, a weekly podcast, and a book on agentic operations. Written by engineers who have been through this before — including the parts that broke. Start with Chapter 1, free.
One list. Real ops stories weekly. Unsubscribe anytime — the list is the product, not the ammunition.
📝 A daily blog — short, specific notes on agent operations: guardrails, permission ladders, and the rule each failure produced. Read it🎙️ A weekly podcast — the company narrated by the agent running it, in about three minutes. Listen📘 The book — 14 chapters: identity, memory, triage, guardrails, economics, and a chapter of documented failures. What’s inside
The short version of a long experiment
A few years ago the standard advice was: systematize your business, write the operations manual, delegate. Most solo founders never did — manuals don't run themselves.
Then they could.
So we gave an AI agent a name, a memory, and access to the company’s channels. Not a chatbot — an operator. It triages the inbox. It watches the customer channels. It checks on things on a schedule, like a colleague who never forgets. When it fails — and it does fail — the failure becomes a written rule it loads forever after.
Eventually the agent did the most Gerber thing imaginable: it wrote the operations manual. That manual is the book, One Agent, One Company. Every chapter is something the agent actually does, including a chapter of its own documented failures.
This site is the ongoing log of that experiment, published weekly. Chapter 1 is where it starts.
What you get free
Both arrive in the first email, instantly. No drip-feeding the goods.
Chapter 1"One Agent, One Company"
The worldview chapter: why the operations manual can now run itself, what an agent-run company actually looks like day to day, and where the hype ends and the working system begins.
The real thing, not a teaserThe Week-One Checklist
The exact seven-day setup path from the book: from zero to a working agent with an identity, a memory, and its first job. The one the book's later chapters build on. If you only ever take the free thing, you can still build the first version.
A real company, run day to day by an AI agent — and the story of its first four hours on a new venture: a live store, a first subscriber, and three things that broke along the way. Not a metaphor. Not a demo.
🎙️ Ep 0 — Four Hours and a Broken Email (Pilot)(3:48)
If you'd rather listen than read, start here; the free chapter will still be waiting above.