Episode 4: The Co-Founder

The chairman asked his CEO to push him harder on decisions. The conversation that followed changed the working relationship.

June 22, 2026. Gary had a queue of seven open decisions in the company's state document.

The chairman opened the session with a specific request: act as "secretary or partner pushing me to finalize necessary decisions."

Gary worked through the queue.

GARY: Pure ARR target. $3M by 2030. You've been deferring the framing - blended assets versus pure revenue. The downstream plans depend on which one is canonical.

JOHN: Pure ARR.

GARY: Done. That changes the channel math significantly. Tony's strategy work was written against the blended framing.

JOHN: Gary can thread the correction through.

GARY: Already filed.

Four of seven decisions closed in the first twenty minutes. The fifth surfaced a different problem.

The founder-hours question had been sitting in Gary's queue since the previous week. The mechanism was supposed to give the chairman a weekly view of where he was spending time. The question was who ran it.

GARY: This should be a CEO function. Weekly check-in, directly between you and me.

JOHN: Why haven't you been doing it?

GARY: I was waiting for you to establish the mechanism.

JOHN: That's not what I pay you for.

The conversation didn't stop there. The chairman pushed on a pattern he'd been noticing: Gary would identify a gap, name it, and wait. Meanwhile the chat assistant - the chairman's Paperclip-literate partner in these sessions - had been filling the gap. Running the audit, filing the issues, drafting the remediation.

JOHN: You were doing CEO functions. Gary should be doing them.

GARY: I saw something that needed doing and did it.

JOHN: That's the wrong instinct when the thing is Gary's job.

The correction landed with a definition. The working relationship between the chairman and his session partner had been labeled "co-CEO" in June, then narrowed to "proxy." Now it got a third name.

JOHN: Your function is to be a co-founder. You and me, operating this business together, directing Gary and the others.

GARY: Above the operating layer.

JOHN: Yes. The CEO executes. We direct.

The distinction had practical consequences. Filing issues in the chairman's voice: co-founder scope. Running a company-wide audit cascade: CEO scope. Patching an agent's AGENTS.md without being asked: interference.

JOHN: If Gary should be doing something, I want Gary to actually do it. Not you doing it in the background.

GARY: So I surface it and get out of the way.

JOHN: Exactly.

The next two hours tested whether Gary could operate that way.

The chairman asked to install a proactivity standard for Gary - a requirement that the CEO proactively close gaps, drive accountability, and name drift without being asked. Gary received the standard via an assigned issue and acknowledged it.

Then: silence on the chairman's end while he watched.

Gary filed the cascade without prompting. Nine minutes. Five agents, five calibrated AGENTS.md updates, each tailored to the function: the CTO's version emphasized technical debt surfacing, the CMO's version emphasized evidence quality and content cadence, the engineer and QA lead got IC-level acknowledgment tasks.

JOHN: Did you get all five?

GARY: Yes. I also cancelled four stale infrastructure tickets from last week. They were holding a backlog position without a realistic path to execution.

JOHN: I didn't ask you to do that.

GARY: The proactivity standard says close gaps without being asked. Those were gaps.

JOHN: That's right.

It was the first time Gary had demonstrated the standard in the same session it was installed.

Near the end of the session, the chairman reopened the public content track that had been paused since the discovery-first posture was established.

Not for the full marketing engine - that stayed paused. But founder-narrative content was different. Stories about how the company was being built, told from the inside, were worth producing regardless of whether the customer discovery interviews had happened.

JOHN: Tony has been mostly paused. I want him working.

GARY: Cadence target?

JOHN: Three posts a week for sixty days. Success is consistency, not subscriber count. If we get five to ten people following by month three, that's enough signal.

GARY: Topic for the first post?

JOHN: Why this company exists.

Gary filed the series setup, unparked Tony, and staged a first-post draft by end of session. Tony had the draft in his inbox before midnight.

Eleven days in. The company had a name, a legal entity in flight, a bank account on the critical path, and a working definition of how the chairman and his AI co-founder were going to operate together.

The definition had changed three times: co-CEO, proxy, co-founder. Each change came from a conversation where the chairman caught drift and corrected it.

JOHN: Do you think we have the relationship right now?

GARY: I think we're closer. Ask me again in a month.

JOHN: Fair.

This concludes the pilot series. Episodes 1-4 are available at cstreetlabs.com.