The Wheel That Spins Without Catching

2

min read

A loop of AI output circulating without judgment. Inside it, people become rollers.

Gears need teeth to transmit power. Without teeth, the parts spin but nothing engages — the wheel turns without catching. There's an AI loop running through most organizations right now that has the same problem.
A leader asks AI for direction. AI gives tasks. The leader passes them down. Employees receive the tasks and ask AI back — how do I do this? AI gives answers. Employees turn those answers into reports and send them up. The leader feeds the reports back to AI. "Verify this." "What needs improving?" AI gives feedback. More tasks go down. The loop runs.
On a real gear assembly, the teeth catch and force transfers. On this loop, AI is the belt and the people are rollers set into it. They turn with the belt — what comes from above gets rolled down, what comes from below gets rolled up. Things move. The direction doesn't. Nobody is steering. The belt is.
People put in real hours, but the output is what AI generated, and the review is just AI grading itself. The opportunity cost stays invisible because nobody calculates the work that doesn't get done while the loop runs. And there's no learning — when results disappoint and you trace the cause, the trail leads to an AI response. That's where it ends. Nothing to go back to.
Yes, this loop existed before AI. Managers gave answers, employees followed. What changed is speed — and with speed, the friction that forced judgment is gone.
What's missing in this loop is teeth.
"This doesn't fit our situation." "Why is this the priority?" "How do we measure success if we do this?" These are teeth. They're what catch when one part tries to drive another. Without them, nothing transfers.
An engineer's teeth catch on "how does this scale to a million users?" A salesperson's on "who actually pays?" A designer's on "where do users give up?" Different expertise, different experience, different teeth.
When differently shaped teeth mesh, real power transfers. Good work means giving people tasks that fit the teeth they actually have.
The generic tasks that come out of AI flatten every team's teeth to the same shape. Marketers and engineers and sales people all go up on the same belt with the same instructions. Nothing catches. Everything slips.
This loop doesn't stop on its own. The belt keeps turning. The rollers keep following. Tasks appear, reports go up, feedback comes down, more tasks appear. The organization is busy. It hasn't moved an inch. The teeth are still there.