[onoht]

This piece hit something I've been trying to articulate for months.

The part about the identity shift from builder to reviewer - that's the real thing nobody's talking about. I spent years getting good at turning thoughts into code. That's a craft. There's a rhythm to it, a kind of flow state you hit when the problem and the solution start locking together.

Now I spend most of my time evaluating code I didn't write, catching issues I didn't create, in systems I didn't design. The volume is higher. The satisfaction is lower.

The HBR study numbers track with what I'm seeing around me. 83% saying AI increased their workload. That's not a bug, that's the whole point. We made code production faster, so now we produce more code. Nobody stopped to ask if that was actually the bottleneck worth solving.

The thing that gets me is the pretense. Everyone talks about AI making engineers more productive. But if you look at what's actually happening, we're not producing better software. We're just producing more of it, faster, with the same number of people. That's not productivity - that's volume.

What's being lost is the time to think. To sit with a problem long enough that you actually understand it before you start implementing. The old friction of writing code manually gave you that thinking time by default. Now you have to fight for it.