Last Saturday I rewrote a component twice. Not because it was broken — it worked fine. I just wanted to see how it would feel with a different grid, a different rhythm of whitespace. Two hours, a dead end, an undo.
I hadn't done that in years. Not since the early days, when experimenting was the point.
The curiosity I'd priced out
For most of my 16 years in this work, curiosity had a cost I couldn't pay. A new stack for a side idea meant six weeks of learning before the first line of value. Seeing a feature in three different flows — three days I didn't have. Rewriting a button because the motion curve bothered me was homework for after the last deadline of the night.
So I stopped reaching. I chose the stack I knew. I built the flow I could defend. Curiosity became a luxury for weekends that never arrived.
Not what it took — what it returned
Every second article is about what AI takes: jobs, craft, judgment, jobs again. Fine. But the conversation keeps missing the other side of the ledger.
For me, it returned the Saturday afternoon I spent on that useless rewrite. It returned the week I gave to SwiftUI for Phora's iOS side — something I'd have easily postponed another year. It returned the budget to be wrong: to try things with no business case beyond "I wonder what happens if."
The speed isn't the gift. The gift is that being unsure costs less now. You can afford not to know, at least for a weekend.
The permission I'd misplaced
Sixteen years in, this feels like a permission I'd put down somewhere between the second and third client. Back when I'd stay up rewriting a button because the motion curve bothered me. Back when I'd pick a framework I'd never seen because the docs looked fun. None of that paid the rent. All of it made me better.
The speed is useful. But the best thing AI gave back wasn't time. It was room to be a beginner again.
There's always a next level.
If you like what you see — whether you're building a product or a team — I'd love to hear about it.