You are responsible for the slop

There's a complaint I hear in every room where AI comes up: the output is generic. It's slop. The writing sounds like everyone else's writing, the design looks like a template, the strategy could have been printed off a cereal box. And the complaint always arrives with the same verdict attached — the tool isn't good enough yet.

I want to offer a less comfortable verdict. It's the man in the mirror. You have to look at that man in the mirror and understand that you are the one responsible for the products you create — including the ones you create with a machine.

Directed, not clicked

When I show people the systems I've built — dashboards, campaigns, whole platforms — somebody always asks how I built them. And I correct the question every time: I didn't build them, I directed them. I told the tools what I wanted, looked at what came back, and told them what was wrong with it. Then again. And again, until the thing on the screen matched the thing in my head.

I'm equally honest in the other direction: the tools did the building. I don't take credit where I don't deserve it. But the taste, the standard, the refusal to accept the fourth draft when the fifth one could be better — that part is mine, and that part is the difference between slop and something you'd put your name on.

Quality is a decision, not a feature

Here's why this mindset matters more than any technique. If you believe quality comes from the tool, you'll wait — for the next model, the next update, the version that finally reads your mind. You'll be waiting forever, and your competitors who took responsibility will lap you while you wait.

If you believe quality comes from your direction, everything changes today. A weak first draft stops being a verdict on the technology and becomes what it actually is: information about your instructions. You sharpen the direction, you raise the standard, you go again. The machine never gets tired of your revisions. The only limit on the loop is how early you're willing to say good enough.

So the next time something generic comes back off the assembly line, resist the modern reflex of blaming the machine. Ask what you failed to specify. Ask what standard you never stated. Then get back in there and work on producing the thing until you actually get it. That's not a burden — it's the best news in this whole era. The quality of what you make with AI is up to you. It always was.