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 more efficient the system becomes, the more 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 more efficient to use, we actually consume more of it, not less.
The "Productivity" we see on corporate dashboards is not translating into shorter workdays or deeper focus. It is translating into higher quotas and denser calendars.
For example:
\- you complete a week's worth of stories in 3 days... so the sprint velocity expectations just doubled for the next week
\- you can send 10 emails in the time it took to draft 2... so now you are expected to manage 50
\- meetings / documents summaries are instant... so now you are responsible for "knowing" 10x more information than before
AI is not lowering the floor of our workload; it is raising the ceiling of what is considered "normal" human output. We are optimising the "how" of work to near-perfection, but the "how much" is scaling even faster.
AI has increased our capacity, but it has not reduced our burden. It is like a treadmill that keeps getting faster.
The real question is not "Is AI making us more productive?"
The question is: If the ceiling of expectations keeps rising as fast as tools, do we ever actually get to stop climbing...???
[link] [comments]
