The Reversed Bicycle

There's a video that made the rounds a few years back: an aerospace engineer who understood the physics of bicycles perfectly well got on one where the steering had been reversed — turn the handlebars left, the front wheel goes right. He fell over immediately. It took him eight months of daily practice to learn to ride it. His own son picked it up in two weeks. Not because the boy was more talented. Because he had far less to unlearn.

I think about that video every time I read about a company doing what Klarna did this year. Their AI support agents were, by every normal metric, a real success — two-thirds of tickets resolved, response time down from 11 minutes to 2, a projected $40 million a year in profit. So the company did what years of running a support org had trained its leaders to do the moment a cost center gets more efficient: they cut headcount, from 5,000 down to 3,500.

Then the ambiguous cases started arriving — the ones with real emotion attached, the ones that don't sort into a category. Klarna's own CEO has since admitted publicly that the company cut too aggressively, and is now rebuilding the human expertise it let go.

I don't think Klarna misread the technology. I think their instincts were the reversed bicycle. Everything they knew about running an efficient support operation — correct for years — told them exactly the wrong thing to do the moment the terrain changed underneath them. Expertise wasn't the advantage here. It was the obstacle.

Where has your own expertise been telling you exactly the wrong thing to do?

From Dancing in the Wind: An Adaptive Leadership Framework in the Age of AI.