Six Months of Vibe-Coding, No Revenue

I built a real product by talking to machines for half a year. The thing I would not build was the one that pays: a sales call.

For six months I built, and I charged no one. Not a slow ramp, not a cautious soft launch. Zero revenue, and if I am honest, zero on purpose, without ever letting myself say the word "purpose" out loud.

Six months of building produced a real inventory of capability and one number that never moved. The problem was never the left column. It was that I kept adding to it to avoid the right one.

Here is what got built in that time, because I want to be fair to the work. Six agents that hand tasks to each other. A multimodal rule engine. Forty-two proprietary computer-vision models. And, late, almost grudgingly, a three-step sales kit. I am not a traditional coder. I build by running AI, by sitting with task windows and describing what I want until it exists. People have started calling that vibe-coding, and for six months I was very, very good at it. The product is real. It works. And it earned nothing.

I want to write down why, because the reason is not the one I told myself, and if you build things, it is probably coming for you too.

Sources

  1. Founders must do unscalable things, including recruiting users by hand (Paul Graham, Do Things That Don't Scale, 2013)paulgraham.com
  2. Poor product-market fit named in 43% of startup failure post-mortems (CB Insights, Top Reasons Startups Fail)cbinsights.com