What AI can't do for a solo developer
AI writes code fast. It still can't decide what's worth building, notice the bug you haven't hit yet, or care about your users.
By Nzubechukwu Cyprian · Pharmacy student at UNN, full-stack developer
There's a version of the AI hype where solo developers suddenly ship like teams of twenty. I live on the other side of that hype, building products mostly alone, and I can tell you exactly what AI did and didn't change.
What it changed: the speed from idea to first working version. Scaffolding, boilerplate, migrations, config files, that annoying utility function you'd otherwise write by hand — all of it compresses from hours to minutes.
What it didn't change:
- deciding what to build, and harder, what to delete
- noticing the bug in the flow you assumed worked
- the taste to know that a screen is ugly even though it functions
- the patience to support a real user on a bad day
- the responsibility when something you shipped breaks someone's data
The real bottleneck of one
When I audited my own week, coding was never the bottleneck. The bottlenecks were deciding, verifying, and caring. AI amplifies throughput of the thing that was already cheap. It barely touches the things that were always hard.
So my honest summary: AI made the easy part instant and left the hard part hard. The people who feel like they "10x'd" were bottlenecked by typing. Most of us weren't.
Keep reading
- AI
The model is not the product
Everyone debates which model is smartest. Almost nobody talks about the ninety other things standing between a prompt and a shipped feature.
- AI
My AI workflow is boring, and that's the point
No autonomous agents rewriting my codebase. No prompt wizards. Just a disciplined loop between a draft, a review, and a test suite.
- AI
Prompting is just requirements gathering with worse vocabulary
The teams that get good results from AI aren't better at talking to models. They're better at saying what they want.