A company that previously had
20 engineers
could potentially accomplish comparable output with
2–5 engineers + hundreds of AI agent-hours
But I don’t think software engineers simply disappear. Instead, code is no longer the scarce thing. The bottleneck moves.
What becomes scarce is:
- Knowing what to build
- Understanding the business
- Connecting AI to the right systems and data
- Defining constraints
- Evaluating and measuring results
This is a huge economic change.
The companies that stand to benefit the most from this won’t be the ones using AI the most. They’ll be the ones that understand where to apply it, connect it to how the business actually operates, and measure whether it creates value.