Uber CEO Dara Khosrowshahi just described the exact moment companies stop hiring engineers. It’s closer than anyone wants to admit. Khosrowshahi: “About 90% of our coders are using AI.” But that’s not the number that matters. 30% of those engineers have become power users. And what’s happening to their output has no historical precedent. Khosrowshahi: “They are showing a clear differentiation in the number of diffs.” A diff is a code release. The purest measure of engineering productivity. Khosrowshahi: “It’s changing their productivity in a way that I’ve never, ever seen before.” Right now, the math still favors hiring. If an average engineer becomes 25% more efficient, Uber hires more engineers to go faster. But that equation has an expiration date. Khosrowshahi: “Maybe 5 years from now as the engineers get more and more productive, I may not decide to add engineering headcount.” The tipping point isn’t when AI replaces engineers. It’s when adding an AI agent and buying GPUs produces more output per dollar than hiring a human. Khosrowshahi: “At that point instead of adding an engineer, I should add agents and buy some more GPUs from Nvidia.” When the CEO of a company built entirely on software says that out loud, it’s not a prediction. It’s a planning assumption. Khosrowshahi: “The job of a coder is going to change from actually writing the code to orchestrating agents who are writing the code.” Not writing. Orchestrating. The engineer becomes the conductor. The AI becomes the orchestra. The most valuable asset in a tech company is officially shifting from human capital to pure compute. And once that math flips, it doesn’t flip back.