Marketing is splitting into two halves that can’t share a strategy. I build AI systems for a living and I’m betting my own brands on craft, so this isn’t theory from the bleachers — I run tools from both sides every working day.
The slop half
AI-generated landing pages, AI-generated emails, AI-generated thought leadership. Volume is up 100x and trust is down by about the same. Anyone running that playbook is racing a hundred thousand people running the identical one, and the marginal value of the output is falling toward its marginal cost: zero.
That’s not a moral objection. It’s an operator one. Slop works on Tuesday and dies by Friday, because by Friday everyone with a credit card has the same machine.
The craft half
Brands that are recognizably made by people. Slower output, stronger trust, a smaller funnel up top, much higher conversion at the bottom. The economics work because conversion compounds and attention doesn’t.
You can tell when writing has a real person behind it — the specific weird example, the willingness to hedge, the sentence that risks something. Your audience can tell too. Models don’t produce those tells on their own.
The part people get wrong
The split isn’t “AI versus humans.” I build AI phone agents. I use these models every day, on almost everything. The split is about who’s the voice.
AI underneath a human voice is leverage — it makes my hands faster and my output more mine, not less. AI as the voice is slop — it makes everyone’s output identical. Same tools. Opposite bets.
Why you still have to pick
You can’t run the ten-posts-a-day engine and still defend the post that took six hours. Once volume becomes the metric, the slop eats the craft from the inside. And the audiences sniff each other out in about a minute.
Pick one. If you don’t pick, the market picks for you, and it picks slop — because slop is what most people in the room are quietly defaulting to.
Where I might be wrong: maybe more operators than I think will pull off scaled craft — AI trained on a real voice, drafted fast, edited hard. I doubt it, because the editing is itself a craft most marketers never built. But I’m watching.
I picked craft, with AI doing the heavy lifting underneath. Ask me in two years whether the line held.