Love, Skill and AI

1 · AI already took production

Be honest about how much AI already does. Marking up door and window dimensions on a plan — deciding what needs a note, how to show it — used to take three or four hours. The model now does the plan and its annotations in about ten minutes. A drawing set that took three days is heading toward half a day.

And it isn't only speed. AI writes, at volume. It never tires, never loses focus, and takes in more than any of us can. On the production side, it wins.

2 · The work moved from making to judging

So production isn't where we add value anymore. AI can write ten versions; it can't tell you which one is good. It can draw ten schemes; choosing the right one is still ours.

We encode that judgment. We built a review we feed our own projects into — not generic design rules. It reads proportion off the image and makes specific calls: whether an opening belongs here, not a vague "add emphasis." It helps. The scarce skill isn't making. It's judging.

But we won't pretend judgment is a safe harbour. For a long time, taste and judgment looked like the last things a person was still needed for. AI is coming for those too — its output already beats an average design, and beats a fresh graduate's easily. Taste isn't a fixed moat. We have to keep raising our own just to hold its value.

3 · Interest keeps you in — but it isn't enough

None of this lasts without genuine interest — no one puts in the years it takes to get good at something they don't care about. But interest alone isn't a moat either. AI out-produces any one of us ten to one — ten options to our one. Passion doesn't close that gap by itself.

4 · The moat is the intersection

The best work here doesn't come from mastering every tool. It comes from people who reach technology through their own door — bringing judgment, taste, and real interest, and letting AI carry the execution no one wanted to master by hand.

That's the intersection, and it's the whole point. The moat is where three things meet: what a person loves, what they're genuinely good at, and AI in the loop. Miss one and it collapses. Interest without skill stays a hobby. Skill without AI gets out-produced. AI without judgment and taste produces what anyone else's would. Put the three together and you're hard to replace. Love, skill, AI — that's the only moat that holds.

*That's one half of how we see it. The other half — why we integrate at all, and why "AI-native" is the wrong goal — is in* [**Towards AI Integration**](/ai/towards-ai-integration/).