For a long time, we tended to confuse being smart with making good decisions. We assumed that if we had enough data and thought long enough, the right answer would eventually show up. That model breaks down when answers become easy, but decisions don’t.

Intelligence can scale now. Decision quality can’t scale. When a constraint disappears, it doesn’t vanish, it moves. And whatever can’t scale becomes the bottleneck.

When intelligence becomes abundant, it’s natural to ask what replaces it as the scarce resource. Creativity. Trust. Emotional intelligence. Those may all matter. But before any of them can compound, something more basic has to work.

That’s where judgment shows up. Judgment is the ability to tell what matters from what doesn’t, and to act on that distinction. It sounds obvious. Which is why it’s easy to miss. But it explains a lot of stalled progress that otherwise looks mysterious.

I notice this most clearly when I get stuck myself. It’s not because I lack options. It’s because I have too many. AI is very good at producing them: more directions, more arguments, more things that could work. What it doesn’t give me is a reason to choose one. So important and unimportant decisions start to receive the same attention. Everything moves. Nothing advances.

Judgment is mostly about subtraction. Progress comes from deciding what not to do. Once enough noise is removed, what remains is often clear. Not elegant. Just clear. The difficulty isn’t finding viable paths. It’s letting go of paths that could work, sound reasonable, and still don’t matter.

This is uncomfortable. A real decision creates consequences. It closes doors. It makes later explanations harder. Asking an AI for another option doesn’t do any of that. So it’s tempting to stay in motion, refining, exploring, iterating, long after a decision should have been made. It feels like work. It isn’t judgment.

In an AI-driven world, this turns judgment into leverage. People with good judgment don’t outperform by doing more. They outperform by wasting less. They start fewer things. They reverse course less often. Over time, this compounds in ways that raw intelligence no longer does.

Judgment can’t be outsourced. Not to AI. Not to process. Not to consensus either. At some point, someone has to decide and live with the consequences.