Companies big and small are thinking about how AI will impact their coding workforces.
Uber CEO Dara Khosrowshahi recently offered a detailed and data-grounded view of how AI is already reshaping engineering productivity at one of the world’s largest technology companies — and where he thinks it’s all heading. His remarks are striking not just for their candor, but for the specificity with which he describes a future where the decision to add a human engineer might be replaced by the decision to add AI agents and GPUs instead.
“90% of our coders are using AI,” Khosrowshahi said. “There are probably 30% of them that are power users. They are showing a clear differentiation in the number of diffs. A diff is a code release that’s different from the last code release. So one of the measurements of productivity is just how many diffs are you putting to the code base?”
The metric is telling. At a company the scale of Uber, the volume and velocity of code changes is a direct proxy for how fast the business can move. Khosrowshahi made that connection explicit.
“Uber’s just a giant code base. That’s what we are. And so our engineers are literally the builders of the company — they are manufacturing the bricks that go into the system, and the architects who are thinking about what the system should look like. While 90% of our engineers are using AI tools of some sort, there’s about 30% of them that are using them at a completely accelerated pace, and it really is changing their productivity in a way that I’ve never ever seen before.”
He went on to describe how the nature of the engineering job itself is starting to change: “Ultimately, I do think that the job of a coder is going to change more and more from actually writing the code to, to some extent, orchestrating agents who are writing the code or building systems for you. It becomes more of an orchestration job versus a manual writing job, but the job will still be there, and my attitude is if my average engineer became 25% more efficient — which we haven’t gone there yet, but we will get there — I’m going to hire more engineers, because I want to go faster. There are still lots of unsolved problems that we haven’t solved.”
But then came the longer-range view — and the statement that will likely generate the most discussion in boardrooms and engineering departments alike.
“I can imagine, maybe five years from now, as the engineers get more and more productive, I may not decide to add engineering headcount, because at that point, instead of adding an engineer, I should add agents and GPUs from Nvidia. That may be the investment in the future. We’ll see.”
Khosrowshahi is not predicting mass layoffs — in fact, he explicitly frames near-term productivity gains as a reason to hire more people, not fewer. But the five-year horizon he describes points to a structural shift in how technology companies think about growth. Today, scaling engineering capacity means hiring engineers. In the future Khosrowshahi envisions, it may mean provisioning compute. The companies that figure out how to harness the productivity gains of AI-native engineering workflows will be able to do more with less — and the executives who are thinking about that tradeoff now are the ones most likely to be ready when that moment arrives.