In June a colleague in Beijing and I started a partnership. We called it Jing Joes. The idea was simple: AI has made a whole class of apps and services feasible that weren’t before, and we wanted to build some of them, ideally for side revenue, and at worst for ourselves.

Our first app is Ember字, an HSK spaced-repetition app. I’ll write about it here eventually. This post is about the other thing: what it takes to turn “let’s build something together” into a company.

Why proprietary

My partner isn’t convinced by open source. His argument is simple: if we want to sell something, and the whole service with all its balancing under the hood is public, AI makes it very easy for someone to clone what took us months to tune. Proprietary is a defence against that, not an ideology. I’ve stopped arguing.

The agreement

Designing the partnership agreement was the first real lesson. It has to be fair when both of us stay, and fair when one of us doesn’t. What happens if one partner stops contributing early? The agreement has to answer that before it happens, not after, because after is when everyone is angry and no one wants to negotiate.

The tooling loss

The surprise was the tooling. I’ve always built on open source, and I took a lot of it for granted. On a closed proprietary project, GitHub becomes a paid service for features I used for free: private everything, more actions, more seats. I know I can rebuild most of it on my own laptop. That doesn’t make it hurt less. The loss is felt deeply, every time a workflow I used without thinking now costs something or just isn’t there.

What I’d tell someone starting

  • Write the exit rules before you need them. Fair and balanced when one partner fades early is the whole game.
  • Decide the open-source question early, and make it a decision, not a default.
  • Budget for the closed-source tax on tools. It’s real and it’s recurring.
  • A partner who covers your weak side (for me, UI and marketing) is worth more than a smarter copy of yourself.
  • The company is also there to push you to ship. That part is easy to forget and maybe the most valuable part.