Key points
- The author spent five hours managing three projects with AI agents at a Munich coworking space.
- Fiona Fung of Anthropic described kicking off multiple agents simultaneously on Lenny’s Podcast.
- The author cites Gene Kim and Steve Yegge’s book Vibe Coding to explain the workflow change.
A developer reported feeling exhausted after spending five hours at WERK1 in Munich using AI agents to advance three projects simultaneously, raising questions about whether this workflow constitutes harmful multitasking.reported
Evidence: reported
Why working with AI agents is so tiring (and what to do about it) – Tech, work, and what comes nextaifleet-lobsters
The claim
The author described reviewing results, making decisions and writing new instructions for agents handling a blog post, an open source contribution and analytics redesign. This process generated a specific type of mental fatigue similar to that experienced during back-to-back job interviews or customer workshops.
The response
Fiona Fung, who manages the Claude Code and Cowork teams at Anthropic, described on Lenny’s Podcast how she kicks off multiple agents at once and navigates meetings while they work. This behavior contrasts with traditional productivity advice that warns against task switching.
What the evidence means
The author argues that decades of psychological research on task switching may not apply here because the switching occurs at natural hand-offs rather than mid-thought. The author cites Paul Graham and Cal Newport to highlight the tension between this new workflow and established views on deep focus.
In practice
Users adopting this workflow act as head chefs directing line cooks, according to concepts from Gene Kim and Steve Yegge’s book Vibe Coding.reported The limitation is that users must still review outputs and make decisions, creating a cognitive load that the author describes as buzzing tiredness.
Why it matters
- Enterprise buyers must evaluate if agent workflows reduce deep work capacity or simply shift the type of cognitive labor required from developers.
- The comparison to professional kitchen management suggests organizations need clear hand-off protocols to prevent the fatigue described by the author.
- If this workflow is distinct from traditional multitasking, training programs may need to update guidance on context switching for technical teams.
- The reported exhaustion indicates that managing multiple concurrent agents requires significant mental energy, potentially limiting scalability without tooling improvements.
Topics: AI agents, productivity, Anthropic, workflow, cognitive load
Updated
Then / now / next
Productivity experts like Cal Newport advocated for uninterrupted deep focus and warned against task switching.
Developers are using AI agents to run multiple projects simultaneously, describing it as a head chef workflow.
It remains unclear whether this agent hopping model will become a standard enterprise practice or if the fatigue limits its adoption.
Story timeline
- Then Productivity experts like Cal Newport advocated for uninterrupted deep focus and warned against task switching.
- 05 Nov 2026 Why working with AI agents is so tiring (and what to do about it) – Tech, work, and what comes next
- Now Developers are using AI agents to run multiple projects simultaneously, describing it as a head chef workflow.
- Next It remains unclear whether this agent hopping model will become a standard enterprise practice or if the fatigue limits its adoption.
The strongest doubt
Counterpoint from the sources
Traditional psychological research consistently shows that task switching degrades performance far beyond the time taken for the switch itself.
For readers
Developers
- The workflow requires reviewing agent outputs and making decisions, which the author found mentally exhausting after five hours.
- Switching between agents happens at natural seams between tasks, unlike interrupting deep work.
Business
- Anthropic team members use this multi-agent approach during meetings, suggesting a shift in how technical work is scheduled.
- The fatigue described may impact productivity if teams adopt this workflow without accounting for mental load.
Researchers
- The author questions whether decades of research on task switching apply to AI-assisted workflows.
- The comparison to kitchen management offers a new framework for analyzing cognitive load in AI interactions.
Ask the article
Local keyword search of this article; no generated answers.Did this article give you the evidence you needed?
Your choice stays on this device. No public tally.Questions readers ask
Is working with multiple AI agents considered multitasking?
The author argues it is not multitasking but rather agent hopping, where switching happens at task hand-offs rather than interrupting deep thought.
Who else uses this multi-agent workflow?
Fiona Fung at Anthropic described kicking off multiple agents simultaneously while attending meetings, according to the author’s report.
What causes the fatigue described in the article?
The author attributes the exhaustion to the constant review of results and decision-making required to guide multiple agents across different projects.
Connected stories
Reading receipt
- The author spent five hours managing three projects with AI agents at a Munich coworking space.
- Fiona Fung of Anthropic described kicking off multiple agents simultaneously on Lenny’s Podcast.
- The author cites Gene Kim and Steve Yegge’s book Vibe Coding to explain the workflow change.
Still open: Does this workflow actually improve productivity compared to single-threaded deep work, or does the fatigue negate gains?
Still open: Will tooling evolve to reduce the decision-making load that causes the reported exhaustion?