These twenty-eight references let you find the ideas the book draws on, and their context. They do not validate the builder test.
Choose a reading according to the question you want to go into. For learning: Norvig and Hamming. For the conditions of collective work: Edmondson, Grove or Deming. For a need or a delivery: Levitt, Fitzpatrick, Cagan, Singer and Accelerate. For working with AI: Mollick, then the studies by Brynjolfsson, Dell’Acqua and METR. For building under constraint: The Prosperity Paradox, Ushahidi and Masakhane.
The notes below separate the reference from the use proposed here. Connections to other crafts are lines of reading, not results established by these authors. Some links give the full text, others a summary, an extract, or a library record. The dates refer to the edition or release named; the linked pages may show a later edition.
1911. Frederick Winslow Taylor, The Principles of Scientific Management. A historical text on splitting the preparation of work from its execution. It helps you examine where certain prescriptions came from; its conception of authority is not a model to take up uncritically. Full text.
1954. Peter Drucker, The Practice of Management. A bearing for putting the work of management, its goals and its responsibilities up for discussion. Here, the connection to the cards is there to raise questions about what work is for, not to reduce every activity to a commercial enterprise. Library record.
1960. Theodore Levitt, Marketing Myopia. An invitation to define the business by the needs it serves rather than by the product that exists. The reissue, at the publisher.
1968. Melvin Conway, How Do Committees Invent?. The relation it proposes between how an organisation communicates and how its systems are designed can help you investigate a dependency. It is not enough to explain every architecture by the org chart. Introduction and text by the author.
1975. Fred Brooks, The Mythical Man-Month. Essays on organising software development and on the cost of coordination. Adding people to a late project means examining learning and communication; the famous formula does not replace that examination. Anniversary edition at the publisher.
1977. Michel Crozier and Erhard Friedberg, L’Acteur et le Système. A bearing for reading the constraints of collective action and the room people have to decide. The connection proposed here is to examine concrete dependencies without assuming anyone’s intentions. Record for the original edition.
1979. Donald T. Campbell, Assessing the Impact of Planned Social Change. An article on evaluating social change. It invites you to examine how a measure used for deciding can influence what it claims to observe. A work metric therefore also calls for a discussion of how it gets used. Article and DOI.
1982. W. Edwards Deming, Out of the Crisis. MIT Press dates the origin of this book to 1982. Its management proposals give a place to working conditions and to learning. They do not establish a universal percentage of problems attributable to the system. At the publisher.
1983. Andrew S. Grove, High Output Management. Management practices drawn in particular from experience at Intel. You can examine support, decisions and a team’s means there, separating what a manager can decide from what a peer can propose. At the publisher.
1986. Richard Hamming, You and Your Research. An account of choosing problems, of research conditions, and of working habits. It is a situated talk, not a measure of talent or a general rule about careers. Transcript of the talk.
1992. Bruno Latour, Aramis, or the Love of Technology. An inquiry into an abandoned transport project and the relationships that were supposed to hold it together. It lets you question a purely technical explanation of failure, without automatically naming a culprit. The French edition at the publisher.
1997. Eric S. Raymond, The Cathedral and the Bazaar. An essay on modes of software development and on cooperation around open projects. The chances to learn from shared work do not imply that everyone can publish their professional work. The author’s text, revised since its first release.
1999. Amy Edmondson, Psychological Safety and Learning Behavior in Work Teams. The article studies 51 teams in a manufacturing company and reports an association between psychological safety and learning behaviours. It is not the study of hospital teams sometimes summarised under this reference. That association does not guarantee the effect of any particular workshop. Article in the Harvard repository.
Peter Norvig, Teach Yourself Programming in Ten Years. An online text, kept updated, on learning through practice and feedback. The title does not set the same timescale for every person or every craft; Norvig himself points out the limits of a single number. The author’s text.
2009. Ory Okolloh, Ushahidi, or ‘testimony’: Web 2.0 tools for crowdsourcing crisis information. One of its founders tells how a tool was built in Kenya within days during the 2007-2008 post-election crisis, then released as open source and reused worldwide. What is built here, under constraint, serves elsewhere: that is card 10-08. Article at the publisher.
2011. Rich Hickey, Simple Made Easy. A talk distinguishing simple from easy in software design. It gives you a way of examining what is intertwined; extending it to other activities stays an analogy to check. Talk and notes.
2013. Rob Fitzpatrick, The Mom Test. Practices for preparing conversations and limiting biased feedback. Looking for facts rather than approval stays an aid to judgement, without making one interview a validation of the need. The book’s site.
2017. Marty Cagan, Inspired, second edition. This edition develops the work of product teams and the discovery of useful solutions. The distinction between a problem handed over and features prescribed serves as a bearing; it does not remove a beginner’s need for framing or for learning. The author on this edition.
2018. Nicole Forsgren, Jez Humble and Gene Kim, Accelerate. The book presents research on software delivery performance. It helps you examine delivery, stability and working conditions together; it does not close the speed-versus-quality debate in every craft and every context. Introduction and extracts at the publisher.
2018. Annie Duke, Thinking in Bets. A former professional poker player separates the quality of a decision from its outcome, and names the mistake of judging one by the other. It is the mechanism of card 06-05. Introduction at the publisher.
2019. Clayton Christensen, Efosa Ojomo and Karen Dillon, The Prosperity Paradox. Innovation that creates a market for people existing offers leave out, with African examples such as Celtel and its prepaid cards. The book puts cases behind the real needs of card 00-03. Extract.
2019. Ryan Singer, Shape Up. The 37signals method: fix the time, let the scope vary, and decide first how much effort a problem deserves. It is the trade-off of card 05-02. Book online.
2020. Nadia Eghbal, Working in Public. An investigation into how open-source software is really made and maintained: who contributes, who maintains, who pays. It sheds light on cards 02-07 and 10-05. Introduction at the publisher.
2020. Iroro Orife et al., Masakhane: Machine Translation for Africa. An open research community, led from Africa, building machine translation for African languages. Participants without research training contribute published results. Building here, at world level, and publishing it. Article.
2023. Erik Brynjolfsson, Danielle Li and Lindsey Raymond, Generative AI at Work. 5,179 customer support agents: an AI assistant raises issues resolved per hour by 14% on average, and by 34% for novices. The tool spreads the practices of the best agents. NBER working paper.
2023. Fabrizio Dell’Acqua et al., Navigating the Jagged Technological Frontier. 758 BCG consultants: with AI, they work faster and better on tasks it handles well, and make more mistakes on the one it does not. AI multiplies your judgement, and your error: that is card 08-02. Summary of the study by MIT Sloan.
2024. Ethan Mollick, Co-Intelligence. Four rules for working with AI, written for every trade: invite it to every task, stay the person who checks, tell it what role to play, and assume this is the worst AI you will ever use. Introduction at the publisher.
2025. METR, Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity. A controlled trial with sixteen experienced developers on their own projects: 19% slower with AI, while they believed they were faster. That is why card 10-02 separates what was measured from what was assumed. Introduction to the study.
This selection is still marked by management and software, and writing by builders about their own working conditions is still rare. To propose another source, or to contest a connection, name the passage in question and what the source would help someone understand better. The ways to get in touch are open.