Distribution first, code second

A friend told me about his project recently. Three months of building. No user feedback, no validation — just a technical solution he hoped people would want.

I see this constantly with developer-founders.

The problem isn't the code. The code will probably work. The problem is they're solving the wrong thing first. Before you write a line, you need to know if anyone actually wants what you're building. That's the only question that matters at the start.

When I said this to my friend, he pushed back: "But I'm a developer. I don't know how to do marketing."

This is where it gets interesting. Even with everything AI can do today, people still believe marketing is some foreign skill they simply don't have. I don't think that's true anymore.

Sit down with an AI agent and just talk through your idea. You'll have a list of validation experiments in twenty minutes. Run some ads and see how many people sign up. Set up a waitlist. Send a short questionnaire to potential users. None of this requires a marketing background — it requires a willingness to find out if you're wrong before you've spent three months being wrong.

The technical solution is increasingly the easier part. Frameworks, tools, AI assistants — building has never been more accessible. Distribution is still hard. Repeatable distribution is even harder.

If you're a developer starting something new, flip the order. Validate first. Build second.