The Right Medium

Take-aways

• Prompt for the understanding dialogue: "Ask me 3 questions about this change, one at a time. Wait for my answer after each question. And only at the end, tell me where I was wrong." Office version: "... 3 questions a critical colleague would ask about this document ..."
• Claude Code: switch the output style with /output-style explanatory or /output-style learning (Claude Code v2.1.269 or later), or run /config and select "Output style", or set "outputStyle": "Learning" in .claude/settings.local.json. Explanatory adds short Insight blocks; Learning leaves a TODO(human) spot at real design decisions and waits for you. https://code.claude.com/docs/en/output-styles
• Diagrams as text: GitHub renders Mermaid diagrams (for example sequenceDiagram) in issues, pull requests, wikis and Markdown files. https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/creating-diagrams

Studies and sources

Mentioned in the episode

• Larkin, J. H. & Simon, H. A. (1987): Why a Diagram is (Sometimes) Worth Ten Thousand Words. Cognitive Science 11(1), 65-100. Informationally equivalent diagrams and texts differ in computational effort: diagrams group information by location, save search and support perceptual inferences. https://doi.org/10.1111/j.1551-6708.1987.tb00863.x (free copy: https://mechanism.ucsd.edu/bill/teaching/F12/cs200/Readings/larkin.whyadiagramissometimesworth.1987.pdf)
• Roediger, H. L. & Karpicke, J. D. (2006): Test-Enhanced Learning: Taking Memory Tests Improves Long-Term Retention. Psychological Science 17(3), 249-255. Experiment 1: 120 students; after five minutes rereading was ahead (81% vs. 75%), after one week retrieval was ahead (56% vs. 42% of idea units). https://pubmed.ncbi.nlm.nih.gov/16507066/
• Chi, M. T. H., Bassok, M., Lewis, M. W., Reimann, P. & Glaser, R. (1989): Self-explanations: How students study and use examples in learning to solve problems. Cognitive Science 13(2), 145-182. Observational study, small sample. https://doi.org/10.1207/s15516709cog1302_1
• Chi, M. T. H., de Leeuw, N., Chiu, M.-H. & LaVancher, C. (1994): Eliciting self-explanations improves understanding. Cognitive Science 18(3), 439-477. Eighth graders, text on the circulatory system, 14 prompted to self-explain, 10 controls. https://doi.org/10.1207/s15516709cog1803_3
• Shen, J. H. & Tamkin, A. (2026): How AI Impacts Skill Formation. Preprint, callback to episode 1. https://arxiv.org/abs/2601.20245
• Stripe: Webhooks, section "Handle duplicate events" ("Webhook endpoints might occasionally receive the same event more than once."). Checked October 16, 2026. https://docs.stripe.com/webhooks#handle-duplicate-events

More on this (mentioned without names)

• Gravino, C., Scanniello, G. & Tortora, G. (2015): Source-code comprehension tasks supported by UML design models: Results from a controlled experiment and a differentiated replication. Journal of Visual Languages and Computing 28, 23-38. Students; the benefit depends on prior familiarity. https://doi.org/10.1016/j.jvlc.2014.12.004 . See also the aggregation of twelve experiments: Scanniello, G. et al. (2018), Empirical Software Engineering 23(5), 2695-2733. https://doi.org/10.1007/s10664-017-9591-4
• Brown, S.: The C4 model for visualising software architecture. A four-level notation from system context down to code; practical, no evidence of effectiveness. https://c4model.com/
• Ainsworth, S. (2006): DeFT: A conceptual framework for considering learning with multiple representations. Learning and Instruction 16(3), 183-198. Conceptual framework, not an effect study. https://doi.org/10.1016/j.learninstruc.2006.03.001
• Mayer, R. E. (ed.): chapter "Principles for Reducing Extraneous Processing in Multimedia Learning: Coherence, Signaling, Redundancy, Spatial Contiguity, and Temporal Contiguity Principles". In: The Cambridge Handbook of Multimedia Learning. Redundancy principle. https://www.cambridge.org/core/books/cambridge-handbook-of-multimedia-learning/principles-for-reducing-extraneous-processing-in-multimedia-learning-coherence-signaling-redundancy-spatial-contiguity-and-temporal-contiguity-principles/CD5B7AE1279A9AB81F8EEBB53DBEC86E
• Adesope, O. O. & Nesbit, J. C. (2012): Verbal Redundancy in Multimedia Learning Environments: A Meta-Analysis. Journal of Educational Psychology 104(1), 250-263. 57 independent studies, mostly university students: spoken plus written on average equal to written alone, better than spoken alone. https://doi.org/10.1037/a0026147
• Trypke, M., Stebner, F. & Wirth, J. (2023): Two types of redundancy in multimedia learning: a literature review. Frontiers in Psychology 14. 63 studies; content redundancy vs. modality redundancy, positive effects of content redundancy depend on prior knowledge. https://doi.org/10.3389/fpsyg.2023.1148035
• Kalyuga, S., Ayres, P., Chandler, P. & Sweller, J. (2003): The Expertise Reversal Effect. Educational Psychologist 38(1), 23-31. https://doi.org/10.1207/S15326985EP3801_4
• Leahy, W. & Sweller, J. (2011): Cognitive load theory, modality of presentation and the transient information effect. Applied Cognitive Psychology 25(6), 943-951. Two experiments with elementary school children (n = 24 and n = 64). https://doi.org/10.1002/acp.1787
• Agarwal, P. K., Nunes, L. D. & Blunt, J. R. (2021): Retrieval Practice Consistently Benefits Student Learning: a Systematic Review of Applied Research in Schools and Classrooms. Educational Psychology Review 33(4), 1409-1453. 50 experiments, 5,374 learners, 57% of effect sizes medium or large. https://doi.org/10.1007/s10648-021-09595-9
• Slamecka, N. J. & Graf, P. (1978): The generation effect: Delineation of a phenomenon. Journal of Experimental Psychology: Human Learning and Memory 4(6), 592-604. https://doi.org/10.1037/0278-7393.4.6.592
• Hunt, A. & Thomas, D. (1999): The Pragmatic Programmer. Addison-Wesley. Section on "Rubber Ducking" (explaining the problem to someone else). https://pragprog.com/titles/tpp20/the-pragmatic-programmer-20th-anniversary-edition/

Tools

• Anthropic: Claude Code documentation, Output styles (Explanatory with Insight blocks, Learning with TODO(human) spots, /output-style command). Checked October 16, 2026. https://code.claude.com/docs/en/output-styles
• GitHub Docs: Creating diagrams (Mermaid in issues, pull requests, wikis, Markdown). Checked October 16, 2026. https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/creating-diagrams

Foundation of the series

• Healthy AI First Software Development, version 2 (October 1, 2026), chapter 6 "Visual models and conversations" and chapter 1, finding 7 ("Active explanation deserves a trial"). The understanding dialogue and the diagram working rule are explicitly labeled as hypotheses there. The team agreement in the chapter "One Picture Instead of Twelve Pages" is a narrative proposal, not a study result.