Where I learned to be wrong

The small jobs went first. A landing page for a dentist, an internal tool nobody would ever demo, a rebuild of a five-page site — the work I used to take to fill a slow week. For years those were the easiest calls to win. Now the same client opens a chat window and builds a version themselves over a weekend.

I don't miss the invoices. They were never the point. I miss what those jobs actually were, which I only understood once they were gone: they were where I learned to be wrong without it costing anyone much.

The cheap work was the classroom

Everything I now call judgment — knowing which feature to cut, when an architecture will fall apart under its first real users, what a client means rather than what they wrote — I picked up on jobs where being wrong was cheap. A bad call on a dentist's landing page teaches the same lesson as a bad call on a bank's, minus the lawsuit. The stakes were low enough to experiment on, real enough to sting.

That practice is what's disappearing. On the freelance platforms, entry-level project share fell from 15% to under 9% in a single year; what's left is either the cheapest machine-grade volume or the expensive, high-risk work. The good money went where someone is accountable for the decision. Ken Rutkowski's point: you now pay more for someone to own a call, not just make it — the gap between what a machine outputs and what a person will put their name on.

But that accountability doesn't come from nowhere. It's built from years of small, cheap mistakes — and those are exactly the jobs vanishing.

I still had somewhere to practice

Chris Dunlop, who runs a consultancy, wrote about the same shift: his question is where junior developers practice now that the small jobs are drying up. Mine is quieter, because I work alone. I got through that school; whoever starts today finds it closed.

That leaves me somewhere good and uneasy at once. This judgment got harder to acquire right as it became the thing worth paying for. And I didn't close the school behind me — the market did.

I build Phora partly to keep a place where I can still be wrong on my own dime. A bad call there costs me a weekend, and only mine. The practice I used to get paid for, I now pay for myself.

You cover the tuition now

Good judgment became the most valuable thing in the trade the same year the cheap school that taught it shut down. If the market won't pay you to practice, the tuition is yours to cover.

Mistakes were always the curriculum. Someone else just used to cover the tab.