AI hasn’t made me a faster coder. It’s made me a better thinker. And that’s a much more interesting outcome.
Before AI, I spent roughly 10% of my time on design and 90% writing code. That ratio has inverted. Now I spend 90% of my time on design and about 10% on code. The AI handles most of the implementation. I handle most of the thinking.
Here’s the thing though: I’m not sure the old ratio was ever right.
We told ourselves the coding was the thinking; that you couldn’t fully design a system without building it; that the act of writing code was inseparable from the act of understanding the problem.
There’s some truth in that, but there’s also a lot of motivated reasoning from an industry that spent decades optimising for the wrong thing. What actually happened is that implementation drag consumed design time.
The friction of writing code meant we committed to solutions before we’d fully understood the problem. We’d get halfway through building something and discover the design was wrong, but by then the cost of changing direction felt higher than the cost of carrying on with something slightly broken.
AI removes most of that friction. And when you remove implementation friction, the design phase expands to fill the space it always deserved.
The result? Better thought through systems. Cleaner separations of concern. Fewer “we’ll fix that later” architectural decisions that become permanent. More time spent asking “is this the right thing to build?” before asking “how do we build it?”
The inversion isn’t just a change in how I spend my time. It’s a rebalancing of where intellectual effort belongs in the development process.
I suspect this is what software development was always supposed to look like.
Has AI changed how you spend your time, or just how fast you spend it?