Darwin or the Leap

AI is driving evolutionary change in how work gets done, but the popular story mixes up two different things. The tools spread gradually through firms, the way Darwin said traits spread. The work they touch has barely moved, and when it does change, it changes in bursts that come with redesign. My hypothesis is that the burst comes when a firm moves decisions to a new venue.

The popular punctuated story gets two things wrong: the leap is happening inside incumbents, not to them, and it doesn't start in a walled-off unit at the edge. I lay out six falsifiable hypotheses and a 24-month, 100-firm test built to tell a ramp from a step. I also add a labeled proposal, governed autonomy, for moving decisions without losing the name on the outcome. None of it has been tested on AI yet.