AI Promised Less Work. I Have Never Worked More.
- AI tools promised to reduce work and save time, but in practice, they often cause developers to take on substantially more tasks simultaneously.
- The shift from tackling sequential, focused problems to overseeing multiple parallel AI agents creates continuous, exhausting context switching.
- Each agent occasionally requires human review, decision-making, or plan approvals, constantly interrupting deeper focus.
- Analysis of the author's GitHub contributions reveals that in 2026, activity surged back to early-career levels despite serving in a leadership role (VP of Technology).
- GitHub contributions in 2026 surpassed every full year between 2021 and 2025, driven by the low marginal cost of initiating new tasks and experiments.
- Work patterns have devolved into cycles of intense bursts followed by acute cognitive burnout and fatigue.
- New model capabilities spark rapid project starts ("cheap implementation"), until reviewing overhead forces an eventual retreat.
- The capacity to generate and implement tasks has far outpaced the mental capacity to think through and manage them.
Hacker News Discussion
- Many commenters strongly resonated with the fatigue, framing the issue as an epidemic of "side quests" that distract from core objectives.
- Commenters emphasized strict task discipline as a countermeasure:
- Planning finite daily to-do lists away from the keyboard and strictly resisting unvetted tangent projects.
- Discussion drew parallels to Cory Doctorow's "Reverse Centaur" concept, where humans become overwhelmed assembly-line assistants servicing autonomous machine output.
- Others pointed out that delegating mundane work removes low-stress mental breathers, concentrating the entire workday into high-friction, consequential decision-making.
- A split emerged on adoption: some view AI agents as essential superpowers worth the discipline overhead, while others argue juggling agents mimics bad management and leads directly to burnout.