Team of One.

On April 12th I became the only engineer at TwinTone.

I'd love to say it was a strategic decision made from a position of strength. It wasn't. It was a decision made because the way we were building wasn't working, and I had a choice: hire again and hope, or find out how far one person with AI could actually go.

I chose to find out.

The first realisation

The first thing I noticed was how much of a "team" is coordination. Standups. Handoffs. Waiting on someone else's branch. Explaining context you've already explained twice. When it's just you and your agents, most of that disappears. There's no one to wait on. There's only the next decision.

What I shipped. In the first month alone I launched a site for Aria, our first AI streamer. I built a pipeline that researches every person who books a call with me before the call happens, so I walk in already knowing their business. And I took apart 32 of the best-performing AI ad formats on the internet and published every one as a playbook on this site. Fake podcasts. Phone-in-hand demos. "What I eat in a day." Formats doing millions of views, broken down to the structure underneath.

A year ago that's a quarter of work for a small team. It was one month for me.

What it actually cost

I won't pretend it was easy. When you're the only one, every bug is yours. Every outage is yours. Every bad decision has nobody to share it with. The work got lighter. The weight didn't.

But here's what I learned in month one. The bottleneck was never the number of people. It was the number of clear decisions being made per day. Agents multiply whatever you point them at. Point them at clarity and you get leverage. Point them at confusion and you get confusion, faster.

A team of one isn't a smaller team. It's a different thing entirely.

‹ Back to home