I Need To Stop Coding

There's a specific kind of insanity that happens when a founder learns to code with AI. You go from "I should hire an engineer" to "I'll just build it myself this weekend" to waking up three weeks later having shipped an entire product nobody asked for while your actual business stalls out.

I know because I'm living it.

The tools are genuinely incredible. Claude writes better code than most junior devs. Cursor autocompletes entire features. You can go from idea to deployed app in a single afternoon. The dopamine hit of shipping something real—seeing it live, watching people use it—is orders of magnitude more satisfying than sending another cold email or updating a pitch deck.

The problem isn't that AI coding tools are bad. The problem is that they're too good. They make the wrong work feel like the right work.

Here's what I've been telling myself: "I'm being productive." And technically, I am. The codebase grows. Features ship. The product gets better. But the business? The actual company-building work of talking to customers, closing deals, building partnerships? That stuff requires a different kind of energy—the uncomfortable, unstructured, no-dopamine-hit kind.

Every hour I spend refactoring a component or adding a new API route is an hour I'm not spending on the things that actually move the needle. And the worst part is, the coding feels more productive. You get immediate feedback. Tests pass or they don't. The deploy works or it doesn't. There's a clean loop.

Business development has no clean loop. You send 50 emails, get 3 responses, have 2 meetings, and maybe—maybe—one of them turns into something six months later. There's no npm run build for relationship-building.

So here's my new rule: code before 10am or after 6pm. Business hours are for business. The AI tools will still be there tonight.

— Ian