Format Online paper / PDF "Developers in the METR experiment believed AI made them approximately 20% faster after completing tasks, although their measured completion times were 19% longer." View Source Location PART 3 - THE THIRD AXIS CHAPTER 15: DEPLOY NOW, PRAY LATER Gambling Builds No Skill Source title Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity Source authors Joel Becker; Nate Rush; Beth Barnes; David Rein Source year 2025 Confidence high Importance high
Format Online paper / PDF "In a randomized study of experienced open-source developers, allowing early-2025 AI tools increased task-completion time by approximately 19%." View Source Location PART 3 - THE THIRD AXIS CHAPTER 15: DEPLOY NOW, PRAY LATER Gambling Builds No Skill Source title Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity Source authors Joel Becker; Nate Rush; Beth Barnes; David Rein Source year 2025 Confidence high Importance high
Format Online paper / PDF "When software teams fail to preserve the intentions, goals, constraints, and rationale behind their systems, they accumulate what Storey calls intent debt." View Source Location PART 3 - THE THIRD AXIS CHAPTER 15: DEPLOY NOW, PRAY LATER Gambling Builds No Skill Source title From Technical Debt to Cognitive and Intent Debt - Rethinking Software Health in the Age of AI Source authors Margaret-Anne Storey Source year 2026 Confidence high Importance high
Format Online paper / PDF "AI-assisted software development can accumulate cognitive debt when code grows faster than developers’ shared understanding of the system." View Source Location PART 3 - THE THIRD AXIS CHAPTER 15: DEPLOY NOW, PRAY LATER Gambling Builds No Skill Source title From Technical Debt to Cognitive and Intent Debt - Rethinking Software Health in the Age of AI Source authors Margaret-Anne Storey Source year 2026 Confidence high Importance high
Format Online paper / PDF "AI-assisted developers produced more complete code while showing weaker understanding of their own implementations." View Source Location PART 3 - THE THIRD AXIS CHAPTER 15: DEPLOY NOW, PRAY LATER Gambling Builds No Skill Source title More Code, Less Understanding? On the Impact of AI Assistants on Developers’ Productivity and Code Ownership Source authors A. Martín-López; R. Tufano; E. Guglielmi; A. B. Sánchez; A. L. Sanz; S. Scalabrino; R. Oliveto; S. Segura; G. Bavota Source year 2026 Confidence high Importance high
Format Online paper / PDF "AI-assisted learning outcomes depend on how the assistant is used, with some cognitively engaged interaction patterns preserving learning." View Source Location PART 3 - THE THIRD AXIS CHAPTER 15: DEPLOY NOW, PRAY LATER Gambling Builds No Skill Source title How AI Impacts Skill Formation Source authors Judy Hanwen Shen; Alex Tamkin Source year 2026 Confidence high Importance high
Format Online paper / PDF "Using AI to complete tasks requiring knowledge of a new Python library reduced skill formation in a controlled experiment." View Source Location PART 3 - THE THIRD AXIS CHAPTER 15: DEPLOY NOW, PRAY LATER Gambling Builds No Skill Source title How AI Impacts Skill Formation Source authors Judy Hanwen Shen; Alex Tamkin Source year 2026 Confidence high Importance high
Format Online paper / PDF "People can misremember whether they or an AI system contributed particular ideas or text, especially in mixed human-AI workflows." View Source Location PART 3 - THE THIRD AXIS CHAPTER 15: DEPLOY NOW, PRAY LATER Gambling Builds No Skill Source title The AI Memory Gap - Users Misremember What They Created With AI or Without Source authors Tim Zindulka; Sven Goller; Daniela Fernandes; Robin Welsch; Daniel Buschek Source year 2026 Confidence high Importance high
Format Online paper / PDF "The LLM fallacy describes a gap between perceived personal capability and actual unaided capability when people treat AI-assisted outputs as evidence of their own competence." View Source Location PART 3 - THE THIRD AXIS CHAPTER 15: DEPLOY NOW, PRAY LATER Gambling Builds No Skill Source title The LLM Fallacy - Misattribution in AI-Assisted Cognitive Workflows Source authors Hyunwoo Kim; Harin Yu; Hanau Yi Source year 2026 Confidence high Importance high
Format Online paper / PDF "AI-assisted participants reported higher overall confidence than participants reasoning without AI, even though faulty AI recommendations reduced accuracy." View Source Location PART 3 - THE THIRD AXIS CHAPTER 15: DEPLOY NOW, PRAY LATER Gambling Builds No Skill Source title Thinking—Fast, Slow, and Artificial - How AI is Reshaping Human Reasoning and the Rise of Cognitive Surrender Source authors Steven D. Shaw; Gideon Nave Source year 2026 Confidence high Importance high
