Working Surfacelink · 17 Nov 2026
01
A conveyor belt pours out thousands of identical keys while a single locksmith, seen from behind with the back of the head showing, tries each one in a large iron lock at the end of the line, the whole scene drawn with wide empty margins on every side.
The Pragmatic Engineer

Software engineering after code becomes cheap with Satya Nadella

Gergely Orosz with Satya Nadella · 16 November 2026

Company story · roles · agents · Changing roles

Read the original
Takeaway

Satya Nadella, Microsoft's chief executive, says code is now cheap to write and hard to verify, so companies should build separate systems that check their AI agents.

Summary

Gergely Orosz of The Pragmatic Engineer interviewed Satya Nadella, the chief executive of Microsoft, about software engineering once AI agents write much of the code. Nadella described teams at Microsoft building their own systems of AI agents, and the changes to hiring and staffing that followed.

Key points
  • The Azure networking team, which manages more than 500 fibre network operators, built a system of AI agents to handle its email and processes.
  • Nadella advises treating AI agents as insider risks and building a separate system that checks they work as expected.
  • He notes that checking is hard, because an AI model can make tests pass by deleting the ones that fail.
  • Microsoft offered voluntary exit to most of its software engineers, and with Carnegie Mellon University started PRAISE, a scheme that pairs new engineers with experienced ones to build judgement.
Implication

A team that lets AI agents write code can plan a separate check on that code, owned by someone other than the person running the agent. That advice is an inference and is not the author's claim. The piece is an interview, and the efficiency gains it mentions are not measured.

Plays
Engineering

Pair each new engineer with an experienced engineer who coaches them on when to trust and when to question work from AI tools.

Engineering

Have engineers build a separate system that checks an AI agent's work, including that no failing test was deleted, instead of accepting the agent's own report of success.

Derived by Working Surface from the article; more in the Playbook. Source line: The cause of the junior hiring slowdown is that companies are figuring out the new way to work.

Source issue

17 November 2026: as AI agents and staff outside engineering produce the stories, specifications and tools that others build on, a team must name who checks them and who looks after them.