A Workday with Energy Left at Night

The short contract from the episode (copy it)

• Goal: a reminder the evening before, at 6 p.m. in the customer's local time.
• Out of scope: no text messages, no new email template, no rebuild of the appointments table.
• Examples: an appointment just after midnight on Wednesday gets its reminder on Tuesday at 6 p.m.; an appointment on the Sunday the clocks change also gets its reminder at 6 p.m. on Saturday, not an hour earlier at 5 p.m.
• Open: reminders for same-day bookings? Miriam decides.

Tool tip: what Claude Code checkpoints do not undo

A prompt that starts a new turn creates a checkpoint. Use /rewind, or press Esc twice with an empty input, to open the restore menu. Only supported file edits made with Claude's own file editing tools are captured; most subagent edits are not. Changes made through Bash commands (deleting, moving, copying) stay. Anything outside the files, such as database changes, calls to external services or emails that were sent, is not undone by a checkpoint. Checkpoints are not a replacement for version control. Source: https://code.claude.com/docs/en/checkpointing

Try this before episode 8

For one week, write a "cairn" at the end of every workday: what is done, what is open, what you start with tomorrow. Add your energy on a 5-point scale (5 = plenty of energy; episode 8 uses the same scale).

• Developer version: keep the lines in a notes file in your project, and write the 4-line contract once before your next task.
• Office version: put the 3 lines at the end of the document or chat you are working on with AI tools, like Miriam.

Sources

Basis of the episode

• Healthy AI First Software Development, version 2 (2026-10-01), chapter 10 "The proposed lifecycle" and chapter 11 "A sustainable workday". The lifecycle is explicitly marked there as a hypothesis to be validated.

Working with agents (vendor practice, not controlled studies)

• Anthropic: Best practices for Claude Code (explore, plan, implement; evidence instead of claims). Covered in depth in episode 3. https://code.claude.com/docs/en/best-practices
• Anthropic: Checkpointing in Claude Code (snapshot per prompt that starts a turn, /rewind, most subagent edits not captured, Bash changes not tracked, not a replacement for version control). https://code.claude.com/docs/en/checkpointing
• Anthropic (2025): Effective harnesses for long-running agents (progress file claude-progress.txt, git history). https://www.anthropic.com/engineering/effective-harnesses-for-long-running-agents
• DORA: Work in process limits. https://dora.dev/capabilities/wip-limits/
• Example of a daylight saving rule (Germany, UTC+1 in winter, EU summer time UTC+2), the basis of the time zone example in the German edition: PTB, time zones and summer time. https://www.ptb.de/cms/fachabteilungen/abt4/fb-44/ag-441/darstellung-der-gesetzlichen-zeit/zeitzonen.html

Interruption and resumption (covered in depth in episode 4)

• Leroy, S. (2009): Why is it so hard to do my work? The challenge of attention residue when switching between work tasks. https://doi.org/10.1016/j.obhdp.2009.04.002
• Leroy, S. & Glomb, T. M. (2018): Tasks Interrupted: How Anticipating Time Pressure on Resumption of an Interrupted Task Causes Attention Residue and Low Performance on Interrupting Tasks and How a "Ready-to-Resume" Plan Mitigates the Effects. Organization Science. Press release: https://www.washington.edu/news/2018/01/16/task-interrupted-a-plan-for-returning-helps-you-move-on/

Breaks and rhythm

• Albulescu, P. et al. (2022): Give me a break! A systematic review and meta-analysis on the efficacy of micro-breaks for increasing well-being and performance. PLOS ONE 17(8): e0272460. 22 samples, 2,335 participants; vigor d = 0.36, fatigue d = 0.35 (k = 9 each), performance d = 0.16, not significant; 13 of 22 studies in the lab; after very demanding work, 10 minutes may not be enough, according to the authors. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0272460
• Basic rest-activity cycle (Kleitman), overview including early criticism of the evidence for a waking cycle: https://en.wikipedia.org/wiki/Basic_rest%E2%80%93activity_cycle

Logging off, open goals, detachment

• Ghibellini, R. & Meier, B. (2025): Interruption, recall and resumption (meta-analysis on the Zeigarnik and Ovsiankina effects, University of Bern). Humanities and Social Sciences Communications 12. https://doi.org/10.1057/s41599-025-05000-w
• Smit, B. W. (2016): Successfully leaving work at work: The self-regulatory underpinnings of psychological detachment. Journal of Occupational and Organizational Psychology. 103 employees; the planning effect mainly for people highly involved in their work. https://doi.org/10.1111/joop.12137 · Summary: https://bps.org.uk/research-digest/heres-simple-way-improve-your-worklife-boundaries
• Wendsche, J. & Lohmann-Haislah, A. (2017, German Federal Institute for Occupational Safety and Health, BAuA): A Meta-Analysis on Antecedents and Outcomes of Detachment from Work. Frontiers in Psychology 7:2072. 91 samples, 38,124 employees; correlations, no causal evidence. https://www.frontiersin.org/articles/10.3389/fpsyg.2016.02072/full

Further reading (not mentioned in the episode)

• Masicampo, E. J. & Baumeister, R. F. (2011): Consider it done! Plan making can eliminate the cognitive effects of unfulfilled goals. Journal of Personality and Social Psychology 101(4), 667-683. Lab experiments. https://doi.org/10.1037/a0024192
• Sonnentag, S. & Fritz, C. (2007): The Recovery Experience Questionnaire. Journal of Occupational Health Psychology 12(3), 204-221. https://doi.org/10.1037/1076-8998.12.3.204

Callbacks to earlier episodes: the restaurant kitchen (episode 1), explore and plan, one-way doors (episode 3), attention residue and the resumption point (episode 4), a diagram as the answer to one question (episode 5), what a green test proves (episode 6).