Who reviews the work when the people who learned by doing it are gone?

The Linux Foundation surveyed nearly 400 hiring managers and found AI driving a 27 percent net increase in tech hiring across Europe. Inside that growth, entry-level technical hiring contracted by 3 percent while the rest of the world grew it by 14 percent. Organisations are now 3.7 times more likely to train existing staff than hire new people.

The report’s authors suggest what is happening. Demand is shifting towards roles built on judgment, contextual reasoning and oversight of AI systems. The work juniors learned on is being absorbed by the tools, so teams are recomposing around supervision rather than production.

Here is the detail that deserves more attention. The capability gap these organisations rate as their most serious is not technical AI skills. It is AI security and risk management, cited by nearly two thirds of respondents. The market is reorganising around oversight while reporting that oversight is precisely what it cannot staff.

And the pipeline question makes it worse. Senior judgment is not hired, it is grown, and it is grown by doing the work the agents now do. Cut the junior intake and you are consuming your reviewer population without replacing it. In ten years the scarce skill will not be writing code. It will be the earned ability to know when generated code is wrong.

Team composition is now a control decision. Every seat you remove from the loop is a check you have chosen to live without, whether or not anyone wrote that down.

When you last reshaped your team around AI, did anyone ask what supervision capacity you were deleting?