Productivity Theatre

Security Theatre is the practice of implementing security measures that are considered to provide the feeling of improved security while doing little or nothing to achieve it (Wikipedia).

Productivity Theatre follows the same script.

AI agents and orchestrators are sold as productivity multipliers. Spin up a swarm of agents, watch the terminal scroll, and feel productive. Deliverability, questionable…

A coding agent that opens 47 PRs in an afternoon looks impressive. How many of those PRs are coherent? How many survive review? An orchestrator that spins up five sub-agents to tackle a task feels powerful. How many of those agents are duplicating work, fighting each other, or solving the wrong problem? I’m still uncertain the value add that people coding with multiple Claude Max subscriptions are actually delivering.

Activity is not output. Motion is not progress.

The same pattern plays out everywhere: dashboards that track lines of code written, meetings about productivity frameworks, tools that optimise the process of planning work instead of doing it. The theatre runs on the feeling that something is happening.

Next time you see claims about AI-enhanced productivity, look past the demo. Try to find out what’s actually shipping.