Overthinking in the Age of Infinite Drafts
Studying classical piano taught me early that precision is achieved through repetition. During those years, I could spend hours at the piano working through the same phrase, each repetition containing subtleties noticeable only to the person playing it, until each note settled exactly how I wanted it to sound.
In experimenting with AI tools, I’ve noticed the same pattern surface again: the quiet pull toward one more attempt in pursuit of something better. You ask for one more edit. Then another. And another. The system returns each version effortlessly, and it becomes tempting to believe the next one might finally be the perfect match.
Repetition toward mastery used to come from not having enough: enough information, enough clarity, enough signal to know which direction mattered.
In the age of AI, it seems that repeated behavior in pursuit of perfection can stem from the opposite problem: too many options. AI can produce endless drafts and endless iterations. It becomes easy to fall down the rabbit hole of refinement — if this version is good, perhaps the next one could be slightly better. The system keeps returning answers effortlessly, and each new version carries the small promise that it might finally feel right.
AI has changed how we work, but it has also quietly changed where the work happens.
Before AI, overthinking came from uncertainty. You revised because the answer wasn’t clear yet. You searched for more information, another perspective, one more data point that might reveal the right direction.
Now the opposite problem appears.
The answers arrive quickly. Often they are good. Sometimes they are very good. But there is always another version waiting behind them. There is always more: insights, information, perspectives you might not have developed on your own.
The discipline no longer lies in producing and refining the work.
The discipline is deciding when iteration stops.
In leadership roles, that shift matters more than it first appears. When systems make iteration effortless, the temptation to keep refining quietly grows.
In practice, I’ve applied this discipline to my own work. The harder leadership challenge is creating conditions where a team — especially one still learning to work with AI — can make the same call.
The habit I learned at the piano still applies.
Repetition improves the work. Practice really does bring a phrase closer to what it should be.
But eventually, the phrase has to end.
AI makes revision nearly infinite.
Judgment still belongs to the person responsible for the outcome.
