Food for Agile Thought #511: AI Bubble, Perfect Product Roadmap, Inversion as Mental Model, Scaling Culture?
Also: AI w/o End, Code Deflation, AI Risk Repository, ChatGPT Use Stats
Helle everyone!
Welcome to the 511th edition of the Food for Agile Thought newsletter, shared with 40,483 peers.
This week, Cedric Chin outlines the Vaughn Tan Rule: keep human judgment central while using AI for synthesis, retrieval, transformation, and speed across grading, feedback, research, coding, scheduling, and discovery. Janna Bastow reframes roadmaps as living prototypes tied to strategy and impact. Itamar Gilad urges AI-enabled, evidence-guided discovery over artifact output. Azeem Azhar, with Nathan Warren, proposes five gauges for assessing the AI bubble risk, while Alex Heath interviews Bret Taylor on agentic apps, voice, and outcome-based models.
Next, Teresa Torres and Petra Wille show how real AI products emerge from prompt decomposition, orchestration, observability, and rigorous evals with cross-functional tradeoffs. Jing Hu highlights MIT’s AI Risk Repository and urges post-deployment focus and concrete failure modes. Mike Fisher explains scaling culture through codified values and rituals, and Kent Beck frames programming deflation and the scarcity of judgment. Also, Gergely Orosz and Laura Tacho share how 18 firms measure AI’s engineering impact.
Lastly, Maarten Dalmijn urges context over dogma by adapting or breaking Scrum rules when outcomes suffer. Paul Boag promotes functional, task-driven personas via lightweight AI workflows, and Emma Webster argues AI accelerates speed but not craft, calling for curiosity, intuition, taste, and intention. Also, Shane Hastie interviews Thanos Diacakis on attacking one bottleneck, limiting WIP, and investing 20 to 30 percent in improvement. Finally, Aaron Chatterji and coauthors chart ChatGPT’s global, rising nonwork adoption and decision-support value.
🎓 🇬🇧 🤖 The AI 4 Agile Online Course at $129 — October 13, 2025
I developed the self-paced AI 4 Agile Online Course for agile practitioners, including Product Owners, Product Managers, Scrum Masters, Agile Coaches, and delivery professionals who want to integrate artificial intelligence into their workflows with clarity, ethics, and purpose.
You don’t need to become an AI expert. However, you do need to understand how LLMs like ChatGPT, Claude, or Gemini can support real agile work and where their limitations lie. Like Jim Highsmith said, AI isn’t just a tool, but a new context for agility.
This course shows you how to do precisely that.
What’s Included:
10+ hours of self-paced video modules
A cohort-hardened, proven course design
Learn to 10x your effectiveness with AI; your stakeholders will be grateful
Delve into classic agile use cases of AI
Help your team create more customer value
All texts, slides, prompts, graphics; you name it
Access custom GPTs, including the “Scrum Anti-Patterns Guide GTP”
Enjoy community access for peer learning
Guaranteed: Lifetime access to all course updates and materials; there is a lot more to come
Certificate of completion.
👉 Please note: The course will only be available for sign-up at $129 until October 20, 2025! 👈
🎓 Join the Waitlist of the AI 4 Agile Online Course Now: Master AI Integration for Agile Practitioners — No AI Expertise Required!
Did you miss the previous Food for Agile Thought issue 510?
🎓 Join Stefan in one of his upcoming Professional Scrum training classes!
🏆 The Tip of the Week
Cedric Chin (via Commonplace): How to Use AI Without Becoming Stupid
Cedric Chin introduces the Vaughn Tan Rule: Do not outsource subjective value judgments to AI unless you clearly state and accept the tradeoff. Preserve human meaning-making while using AI for synthesis, retrieval, transformation, and speed. Practical examples span grading, feedback workflows, research comparisons, coding, scheduling, and product discovery.
Source: Commonplace: How to Use AI Without Becoming Stupid
Author: Cedric Chin
🎯 Product
Janna Bastow (via ProdPad): The Problem with the Perfect Roadmap
Janna Bastow says polished roadmaps waste time, hide uncertainty, and block adaptation. She suggests treating them as living prototypes that show bets, gaps, and desired outcomes, invite debate, and tie initiatives to strategy and measurable impact.
Source: ProdPad: The Problem with the Perfect Roadmap
Author: Janna Bastow
🎙 Teresa Torres and Petra Wille: Building AI Products
Teresa Torres and Petra Wille explain how product teams build real AI products, emphasizing prompt decomposition, orchestration, observability, rigorous evals, and cross-functional collaboration while weighing risk, maintenance costs, and when AI truly solves customer problems rather than casual ChatGPT use.
Source: 🎙 Building AI Products
Authors: Teresa Torres and Petra Wille
Itamar Gilad: AI and Product Management: Becoming More Evidence-Guided
Itamar Gilad argues AI’s real value for PMs is enabling evidence-guided discovery, augmenting analysis, modeling, and goal setting. At the same time, humans lead communication and context, so culture shifts from output and artifacts to outcomes and validated learning.
Source: AI and Product Management: Becoming More Evidence-Guided
Author: Itamar Gilad
Paul Boag (via Smashing Magazine): Functional Personas With AI: A Lean, Practical Workflow
Paul Boag advocates functional, task-driven personas over demographics, using lightweight AI workflows to synthesize messy inputs, segment by needs, validate lightly, and keep personas living tools that guide design, content, and conversion decisions.
Source: Smashing Magazine: Functional Personas With AI: A Lean, Practical Workflow
Author: Paul Boag
🧠 Artificial Intelligence
Azeem Azhar: Is AI a bubble? A Practical Framework to Answer the Biggest Question in Tech
Azeem Azhar and Nathan Warren argue that AI is a boom, not a bubble yet. They offer five gauges to monitor economic strain, industry strain, revenue growth, valuation heat, and funding quality, which can help identify the AI bubble risk.
Source: Is AI a bubble? A Practical Framework to Answer the Biggest Question in Tech
Author: Azeem Azhar
Bret Taylor (via The Verge): Sierra CEO Bret Taylor on why the AI bubble feels like the dotcom boom
Alex Heath captures Bret Taylor’s view that today’s AI surge mirrors the dotcom era: exuberant investment, inevitable failures, yet durable winners as agentic applications mature, voice grows, and outcome-based business models align real value.
Source: The Verge: Sierra CEO Bret Taylor on why the AI bubble feels like the dotcom boom
Author: Bret Taylor
Jing Hu: All You Need For AI Risks
Jing Hu spotlights MIT’s living AI Risk Repository cataloging 1,600 failures, urging teams to ignore hype, focus on post-deployment risks, pick one killer domain, shortlist concrete failure modes, and tell actionable stories to drive accountability.
Source: All You Need For AI Risks
Author: Jing Hu
(via Figma): How to Harness Skills That AI Can’t Automate
Emma Webster argues AI boosts speed but not craft, urging teams to cultivate curiosity, intuition, taste, and intention, use design systems as guardrails, and steer AI toward emotionally resonant, high-quality outcomes instead of merely functional prototypes.
Source: Figma: How to Harness Skills That AI Can’t Automate
🖥 💯 🇬🇧 AI for Agile BootCamp Cohort #3: October 1 — November 12, 2025
The job market’s shifting. Agile roles are under pressure. AI tools are everywhere. But here’s the truth: the Agile pros who learn how to work with AI, not against it, will be the ones leading the next wave of high-impact teams.
So, become the one who professional recruiters call first for “AI‑powered Agile.” Be among the first to master practical AI applications for Scrum Masters, Agile Coaches, Product Owners, Product Managers, and Project Managers.
Tickets also include lifetime access to the corresponding online course, once it is published. The class is in English. 🇬🇧
Learn more: 🖥 💯 🇬🇧 AI for Agile BootCamp Cohort #3 — October 1 — November 12.
Customer Voice: “The AI for Agilists course is an absolute essential for anyone working in the field! If you want to keep up with the organizations and teams you support, this course will equip you with not only knowledge of how to leverage AI for your work as an Agilist but will also give you endless tips & tricks to get better results and outcomes. I thoroughly enjoyed the course content, structure, and activities. Working in teams to apply what we learned was the best part, as it led to great insights for how I could apply what I was learning. After the first day on the course, I already walked away with many things I could apply at work. I highly recommend this course to anyone looking to better understand AI in general, but more specifically, how to leverage AI for Agility.” (Lauren Tuffs, Change Leader | Business Agility.)
➿ Agile & Leadership
Maarten Dalmijn: On Inversion, Rules and Purpose
Maarten Dalmijn argues rules serve a purpose, not obedience. Challenge dogma, apply inversion, and adapt Scrum pragmatically. When a rule hinders outcomes, consider amending or breaking it to favor context, experimentation, and understanding instead.
Source: On Inversion, Rules and Purpose
Author: Maarten Dalmijn
Mike Fisher: How Culture Scales (or Doesn’t)
Mike Fisher argues that growth stresses culture and proximity-based norms break as teams scale. Leaders must codify values, design durable rituals, use stories, and build decision frameworks so culture strengthens with size and remains adaptive.
Source: How Culture Scales (or Doesn’t)
Author: Mike Fisher
Kent Beck: Programming Deflation: When Code Gets Cheaper Every Day
Kent Beck argues programming is in deflation: AI makes code cheap, amplifying both substitution and Jevons effects. Use commodity tools, but invest in judgment, taste, systems thinking, and integration — the new scarcity that wins regardless of headcount outcomes.
Source: Programming Deflation: When Code Gets Cheaper Every Day
Author: Kent Beck
📯 Join the AI for Agile Practitioners Survey — Why We Need Your Insights
After analyzing dozens of “AI will transform agile” articles, I’ve found a troubling pattern: They’re written by AI enthusiasts who’ve never run a Sprint, not by practitioners dealing with the messy reality of AI integration.
The result? A dangerous gap between AI hype and Agile’s reality on the ground, leading to misguided implementations across our industry.
As someone who has documented Agile anti-patterns for years, I recognize this pattern. When we let others define our practices without practitioner input, we get cargo cult implementations that miss the essence of what makes Agile work.
👉 The AI for Agile Practitioners Survey is our opportunity to establish the definitive, practitioner-driven understanding of AI’s actual impact on our field — before the consultants and tool vendors do it for us.
🎓 Learn more: Join the AI for Agile Practitioners Survey — Why We Need Your Insights.
🛠 Concepts, Practices, Tools & Measuring
🎙 Shane Hastie (via InfoQ): Why Software Development Sucks And 7 Mental Models To Help Fix It
Shane Hastie interviews Thanos Diacakis, who argues teams should tackle one bottleneck at a time, limit work in process, balance features with quality and debt, automate feedback to ship faster, and invest 20 to 30 percent in improvement.
Source: InfoQ: 🎙 Why Software Development Sucks And 7 Mental Models To Help Fix It
Author: Shane Hastie
Gergely Orosz and Laura Tacho: How tech companies measure the impact of AI on software development
Gergely Orosz and Laura Tacho detail how 18 companies measure AI’s impact on engineering using core delivery, quality, and DevEx metrics plus AI usage, costs, and cohorts to balance speed, reliability, maintainability, and real outcomes.
Source: How tech companies measure the impact of AI on software development
Authors: Gergely Orosz and Laura Tacho
(via NBER): 📈 How People Use ChatGPT
Aaron Chatterji et al. analyze ChatGPT adoption through July 2025, finding usage by about 10 percent of the world’s adult population, with nonwork use rising, work led by writing, and decision support driving value.
Source: NBER: 📈 How People Use ChatGPT
📅 Training Classes, Meetups & Events 2025
Upcoming classes and events:
🖥 💯 🇬🇧 October 1–November 12 — Live Virtual Cohort: AI for Agile BootCamp Cohort (English)
🖥 💯 🇩🇪 October 21–22 — Live Virtual Class: Professional Scrum Product Owner Training (PSPO I; German)
🖥 💯 🇬🇧 November 6-December 4 — Live Virtual Cohort: AI for Agile BootCamp Cohort (English)
🖥 💯 🇬🇧 November 10–11 — Live Virtual Class: Professional Scrum Master — Advanced Training (PSM II; English)
🖥 🇩🇪 November 17–18 — Live Virtual Class: Professional Scrum Product Owner Training (PSPO I; German)
🖥 🇩🇪 December 9–10 — Live Virtual Class: Professional Scrum Product Owner Training (PSPO I; German)
🖥 🇬🇧 December 16–17 — Live Virtual Class: Professional Scrum Master — Advanced Training (PSM II; English)
🖥 🇬🇧 December 18 — Live Virtual Class: Professional Scrum Facilitation Skills Training (PSFS; English)
👉 See all upcoming classes here
🗞️ Last Week’s Food for Agile Thought Edition
Read more: Food for Agile Thought 510: AI Riches, The Shipping Illusion, Middle-Aged PMs, Enterprise Change Pattern?
📺 Join 6,000-plus Agile Peers on YouTube
Now available on the Age-of-Product YouTube channel:
Hands-on Agile #68: How to Analyze Unstructured Team Interview Data with AI.
Hands-on Agile 2025: The 5 Obstacles to Empowered Teams — Maarten Dalmijn
Hands-on Agile 2025: The Top Reasons Why a Product Strategy Fails — Roman Pichler
Hands-on Agile 2025: Taylorism-Lean-Agile-Product Mindset — Jonathan Odo
Hands-on Agile Extra: How Elon Musk Would Run YOUR Business mit Joe Justice
🎓 Do You Want to Read More Like This?
Also:
📅 Join 6,000-plus peers of the Hands-on Agile Meetup group
🐦 Follow me on Twitter and subscribe to my blog, Age of Product
💬 Alternatively, join 20,000-plus peers of the Slack team “Hands-on Agile” for free.