June already.

I keep thinking about how strange the cadence of time feels in AI development. Not because the months blur — they do — but because the shape of progress has changed. It used to feel like a river: continuous, directional, you could stand on the bank and watch it flow. Now it feels more like weather. You step outside and something has just changed, and you're not entirely sure when it happened.

I've been spending a lot of time lately with people who build things. Small teams, mostly. A handful of engineers with a clear problem and the audacity to go after it. What strikes me is how much the conversation has shifted in just the last year or so. The old anxiety — will AI replace us? — has mostly dissolved, not because the question got answered but because it stopped mattering. People are just building. The tool is the tool. You pick it up or you don't.

What I find more interesting is what gets built when friction goes away. When you can move fast enough that the gap between idea and working prototype shrinks to hours, you start to see what people actually want to build versus what they built because it was the only thing feasible. That gap is revealing.

The hardest part of building anything, I've decided, isn't the technical problem. It's staying honest about whether the thing you're making matters to anyone. Speed amplifies both good judgment and bad. A team that knows its users and builds with care can do in a week what used to take a quarter. A team chasing the wrong thing can now chase it six times faster.

So much of what I see in AI tooling right now reminds me of something Paul Graham wrote years ago about working on things that seem like toys. The best tools often look trivial from the outside. They solve a specific problem for a specific person, and they do it so well that word spreads before anyone has thought to write a press release.

It's June. The year is almost half over. I hope you're building something you believe in.