Why I Left Research to Chase Growth

I spent years chasing hard problems in a lab. Deep learning, published papers, the kind of work where the reward is a smaller error bar and a slightly less confused advisor. I loved it. I also didn't know a single thing about getting someone to care that I existed.

That changed when I founded The Voice Box, my own AI agency, and ran directly into the oldest problem in business: I needed leads, and nobody was going to hand them to me because my math was elegant.

So I did the thing every researcher secretly hates. I posted on LinkedIn. Badly, at first. Then less badly. I started treating every post like an experiment: a hypothesis about what would resonate, a small sample, a result I couldn't argue with no matter how much I wanted to. Forty thousand followers in eight months. Seven million people reached. A hundred thousand dollars closed in ninety days. All organic, no ad spend, no growth hack, just repeated experimentation on a distribution channel I'd been too proud to take seriously.

Here's the part that actually surprised me: growth was not a soft skill I was tolerating on my way back to real work. It was the same discipline as research, pointed at a messier target. Hypothesis, test, measure, iterate. The variables were psychology and taste instead of gradients and loss functions, but the method was identical. And increasingly, the tooling is converging too: growth is turning into engineering, agents doing the testing at a scale no single human ever could.

That's the thread this whole newsletter pulls on. I'm now the founding growth engineer at Swan, building what we call the first autonomous business, and every week I'm going to tell you exactly what that looks like from the inside: what worked, what broke, what I'd do differently if I were starting today.

Still figuring it out, in public. Let's go.