If AI is really making us more productive... why does it feel like we are working more, not less...?

The promise of AI was the ultimate system optimisation: Efficiency. On paper, the tools are delivering something similar to what they promised: - Github Copilot / Claude writes effective code. - LLMs summarise the meeting minutes. - Automations handle Jira tickets. But I see a pattern: The efficient the system becomes, the the system demands. We have not used AI to buy back our time; we have used it to increase our "Normal Output" threshold. This is Jevons Paradox in real time: as the resource becomes efficient to use, we actually consume of it, not less.