The Reps You’re Handing to the Robot
I learned to write a decent headline by writing about four hundred bad ones.
Nobody told me that was the deal at the time. It just felt like grunt work — the junior job of cranking out variant after variant so someone more senior could pick the one that didn’t embarrass anyone. I resented it a little. It was slow, it was repetitive, and most of what I produced went straight in the bin. What I didn’t notice, because you never notice it while it’s happening, was that the bin was the point. Every headline that didn’t work taught me something about the three or four that eventually would. The reps weren’t in the way of the skill. The reps were the skill.
I’ve been thinking about that a lot lately, because AI is very good at exactly the part I once resented. Ask it for forty headlines and you’ll have them before your coffee’s cool. Ask it for the audience segments, the first-draft brief, the variant testing matrix — all the humble, repetitive, character-building drudgery that used to be how a marketer’s judgment got built — and it hands them over instantly, cheerfully, and mostly fine.
Which is wonderful. And which is worth being a little uneasy about.
Because here’s the thing the productivity pitch skips over: judgment isn’t downloaded, it’s accumulated. The psychologist Anders Ericsson spent his career on this, and his 1993 study with Krampe and Tesch-Römer — the one that later got flattened into the pop-culture “10,000 hours” idea — made a subtler point than the myth remembers. It wasn’t that experts just log more time. It’s that they log more time doing the specific, effortful, often tedious reps that build the pattern-recognition underneath skill. Expertise, Ericsson argued, is deliberate practice compounding into intuition. You don’t get the intuition without the practice. There’s no shortcut, and — this is the part that keeps me up a little — there’s no shortcut you can buy for someone else, either.
So what happens to the next generation of marketers when the four hundred bad headlines get written by a machine?
I don’t think the answer is “make them suffer through busywork for its own sake.” That would be nostalgia dressed up as principle, and it’s a bad look on anyone over forty. The reps I did weren’t sacred. Plenty of them genuinely were a waste, and if a tool had freed me from them I’d have grabbed it with both hands.
But some of them weren’t waste. Some of them were the only way I was ever going to develop a gut. And I can’t always tell, looking back, which was which — which is exactly the problem. When you hand the whole repetitive layer to AI, you don’t get to keep the reps that mattered and skip the ones that didn’t. You get to skip all of them. The bathwater and the baby leave together, quietly, and nobody sends a memo.
HBR has been circling this from the management side lately — a run of pieces on redesigning mentoring and entry-level work for the AI era, all worrying the same bone: as the tools take over the tasks that used to build expertise, where does the expertise come from now? I don’t have a clean answer. I’m not sure anyone does yet. But I think the honest move for those of us who did get our reps is to stop pretending the question is rhetorical.
Maybe it means being more deliberate about which reps we protect — keeping a few effortful things human on purpose, not because the machine can’t do them but because doing them is how a person becomes someone worth listening to. Maybe it means the value of a mentor goes up, not down, because the reps still have to happen somewhere and now they have to be designed rather than assumed. Maybe it just means noticing.
I still write bad headlines, by the way. Fewer than four hundred now — the ratio improved — but I still do it by hand more often than is strictly efficient. Not out of principle. Out of a suspicion that the day I outsource all the bad ones is the day I stop being able to spot the good one.
That’s the whole thought. No framework this time, no download. Just a question I’d genuinely like your answer to, if you’ve got one: which reps are you keeping?
Sources: K. Anders Ericsson, Ralf Th. Krampe, and Clemens Tesch-Römer, “The Role of Deliberate Practice in the Acquisition of Expert Performance,” Psychological Review 100, no. 3 (1993): 363–406.
superwired is a fractional CMO practice helping founder-led and growth-stage companies build marketing that’s wired for the AI era — real strategy and senior leadership, without the full-time cost.
