iiinova — where your digital self keeps evolving

Do we really have a difficult person? Or a good person stuck in a difficult situation? Incentives, pressure, fear, priorities: nine times out of ten, the circumstances are the problem. Change the situation, and the "difficult person" disappears. agilepainrelief.com/blog/difficu...

000

Mark Levison @mlevison.bsky.social · 13h

🎯 Certified Scrum Product Owner, October 8-9, live online: GenAI for Product Owners, a hands-on product backlog, and case studies, not guesswork. Certification fees and two years of Scrum Alliance membership included. Registration closes October 7. agilepainrelief.com/courses/cert...

000

Mark Levison @mlevison.bsky.social · 29/09/2026

Speed up ER triage with AI and the beds are still full. We tackled the emergency room problem in isolation. The ER is full for reasons that mostly are not in the ER: no family doctor, long term care waitlists, staffing. No single cause, no magic fix. agilepainrelief.com/glossary/sys...

001

Mark Levison @mlevison.bsky.social · 29/09/2026

🙅 How do you say "no" to stakeholders and stop inflicting a feature factory on your team? Two days of practical answers, grounded in evidence rather than wild promises. CSPO, live online, October 8-9. agilepainrelief.com/courses/cert...

021

Mark Levison @mlevison.bsky.social · 28/09/2026

"Resulting" is judging a decision by how it turned out. Luck plays a big role: good decisions get bad outcomes, and bad decisions sometimes work. For expensive or hard to reverse calls, judge the process instead. Start with the Rule of Three. agilepainrelief.com/blog/why-are...

000

Mark Levison @mlevison.bsky.social · 28/09/2026

🔀 OpenAI says it has reached AGI. Mostly that is positioning ahead of an IPO. The bigger shift is underneath: AI upended retail and is coming for knowledge work. Organizations that only want the Agile bumper sticker will struggle. agilepainrelief.com/newsletter/?...

000

Mark Levison @mlevison.bsky.social · 25/09/2026

Red-Yellow-Green reports summarize a summary of a summary. Green can mean within 5% of budget and tells you nothing about quality. Red says serious issues, not what or why. And green is the default, so too few questions get asked until it is too late. agilepainrelief.com/blog/red-yel...

000

Mark Levison @mlevison.bsky.social · 25/09/2026

🐐 GOAT is back. The Gatineau Ottawa Agile Tour returns November 26 in Ottawa: three panels on geopolitical disruption, the future of work, and whether AI is uncovering better ways. No slides, real trade-offs. $200, capped at 250 people. www.goagiletour.ca

010

Mark Levison @mlevison.bsky.social · 24/09/2026

The federal government moved to four days in the office. Nobody coordinates who comes in which day, so people go in and still do not see their team: same Zoom call, now from a hotdesk. If the goal is collaboration, mandating days is the wrong lever. agilepainrelief.com/blog/forcing...

000

Mark Levison @mlevison.bsky.social · 24/09/2026

A vendor promises AI brings "a new level of objectivity" to estimation. But inaccurate estimates do not come from humans doing a poor job, they come from variability. AI does not make estimation better, because estimation was never the real problem. agilepainrelief.com/blog/gen-ai-...

000

Mark Levison @mlevison.bsky.social · 23/09/2026

When someone says "We do Agile", they are saying something impossible. Agile is four values and twelve principles from a ski hill in 2001. Scrum is a framework. You can do Scrum, Kanban, XP. Agile is how you behave, not the practices you follow. agilepainrelief.com/glossary/agi...

000

Mark Levison @mlevison.bsky.social · 23/09/2026

Learn Scrum from someone who has walked the walk, and will help well beyond the course when you hit the "yes, but in real life" questions. Lifetime access to the materials. CSM, live online, October 1-2. Registration open until the end of September 30. agilepainrelief.com/courses/cert...

000

Mark Levison @mlevison.bsky.social · 22/09/2026

Simplicity isn't installed once, it's a habit. The natural bias is to add: enhance, boost, augment. The 2007/8 crisis happened partly because the mortgage system got too complex for anyone to understand. Complexity in code is a key source of bugs. agilepainrelief.com/blog/less-is...

000

Mark Levison @mlevison.bsky.social · 22/09/2026

GenerativeAI just predicts the next word, trained on most of the internet. No magic. Ask it why it made a mistake and it invents a plausible answer, no self-awareness. Unlike a child, it doesn't learn. Ask your manager how many dice they want rolled. agilepainrelief.com/blog/gen-ai-...

000

Mark Levison @mlevison.bsky.social · 21/09/2026

Wondering if the ScrumMaster role survives GenAI? Better question: which parts should. Logging an impediment, a tool can do. Tracing its root cause still takes a person. 8-question quiz maps where your last Sprints went, and what's worth defending. agilepainrelief.com/quiz/scrum-m...

000

Mark Levison @mlevison.bsky.social · 21/09/2026

Prepare a 200-page immigration application. Decision comes back in under a minute, officer juggling 3-4 files at once. Measuring processing speed created a new backlog: appeals from rushed wrong calls. Same pattern as call centres measuring handle time. agilepainrelief.com/blog/mismeas...

000

Mark Levison @mlevison.bsky.social · 18/09/2026

Put GenAI usage into someone's performance review and you'll get exactly what you asked for. It amplifies whatever you measure, good or bad. Measure velocity, individual productivity, or AI usage, and you'll get more of it, plus the bill and the tech debt.

010

Mark Levison @mlevison.bsky.social · 18/09/2026

Wizard of Oz MVP: looks real to the user, but people do the work behind the scenes. Zappos hand-fulfilled shoe orders from local stores before building a warehouse. Airbnb's founders manually listed and booked early homes.

010

Mark Levison @mlevison.bsky.social · 17/09/2026

Running the Scrum events? Busywork, first thing to hand off. Coaching your PO, chasing the root of a recurring impediment? No tool does that for you. 8-question quiz maps where your last Sprints actually went. No score, just a map. agilepainrelief.com/quiz/scrum-m...

000

Mark Levison @mlevison.bsky.social · 17/09/2026

A Backlog Story estimated at 20 points. Questions to the PO reveal: redemption is a separate story (also 20), refillable is undecided, and a simpler "any amount" version costs just 8. Refinement isn't about looking good. It's about understanding the work.

000

Mark Levison @mlevison.bsky.social · 16/09/2026

The British paid a bounty for every cobra killed in India. People started breeding cobras for income. "What is measured is managed" has caused a century of this. Before adding a metric: can we cut one first? What are the second-order effects? agilepainrelief.com/blog/how-hid...

000

Mark Levison @mlevison.bsky.social · 16/09/2026

A fragile organization that goes faster gets knocked over by the slightest wind. Speed over quality, heavyweight change management, centralized decisions: the usual fragility list. AI hype makes it worse when it becomes a crutch, not a tool. agilepainrelief.com/blog/speed-t...

000

Mark Levison @mlevison.bsky.social · 10/09/2026

73% of accidents studied happened on the captain and first officer's first day flying together. NASA found fatigued crews who'd flown together made half as many errors as rested crews who hadn't. Tired and familiar beat rested and unfamiliar. agilepainrelief.com/blog/in-agil...

020

Mark Levison @mlevison.bsky.social · 10/09/2026

Trust vs Psychological Safety - Individuals vs Teams. Trust is an attribute of each relationship. Psychological safety is an attribute of of whole team (or group). I've never seen that distinction made before. +1 for the HBR How the Best Leaders Shape Conversations buff.ly/ZzlaDnz

hbr.org

How the Best Leaders Shape Conversations

Every strategy, decision, and transformation begins as a conversation. Yet leaders rarely think intentionally about how those conversations unfold. The highest-performing teams steward conversation…

010

Mark Levison @mlevison.bsky.social · 10/09/2026

Most of my writing finds the problems with GenAI. Here is where it earns its keep: spotting patterns in my notes, throwaway prototypes, and small automations. All three work because I can verify the result and the technical debt doesn't matter. agilepainrelief.com/blog/is-ai-m...

000

Mark Levison @mlevison.bsky.social · 09/09/2026

Making the line manager the ScrumMaster removes a safety mechanism. They own performance reviews, so the team self-organizes partly to look good. And much of the pressure a ScrumMaster shields the team from comes from that manager. agilepainrelief.com/blog/why-hav...

010

Mark Levison @mlevison.bsky.social · 09/09/2026

OpenAI is boasting about solving a math problem. Behind the scenes, its a lot more complicated. Why did they suddenly start working on this problem? Why do they not want to credit one the people who helped solve it? OpenAI set a low ethical bar and then consistently go below it. buff.ly/9NHkgil

cims.nyu.edu

https://cims.nyu.edu/~tristanb/statement.pdf

000

Mark Levison @mlevison.bsky.social · 09/09/2026

More code doesn't mean more value. It means more attack surface, more complexity, and mostly more technical debt. Work in messy code is 15 times more likely to have defects. If something isn't the bottleneck, don't optimize for it. agilepainrelief.com/blog/the-rea...

000

Mark Levison @mlevison.bsky.social · 08/09/2026

A rigid Definition of Ready can improve your cycle time while the customer waits exactly as long as before. The delay just moves earlier, into the queue in front of the gate. Measure Lead Time instead, so the waiting is counted. agilepainrelief.com/glossary/def...

000

Mark Levison @mlevison.bsky.social · 08/09/2026

I asked for tests. I got tests, plus an implementation I never asked for. The test cases were credible. The setup code was not, and a local model made fatal tennis mistakes. Try the same experiment on a problem you already know well. agilepainrelief.com/blog/ai-code...

000

Mark Levison @mlevison.bsky.social · 07/09/2026

Everyone wants to do Agile, but not everyone wants to make any changes to be Agile. Lots of focus on ceremonies, not much on the mindset. It's hard to see in your own organization. Outside stories help. Here are five worth reading. agilepainrelief.com/glossary/fak...

010

Mark Levison @mlevison.bsky.social · 07/09/2026

I built a GenAI out of index cards and a die. Five groups, three cards each, roll and read. Plainly: it isn't intelligent, it picks the next most-likely token. It doesn't learn and has no memory. Scale makes a better mimic, not a mind. agilepainrelief.com/blog/human-p...

000

Mark Levison @mlevison.bsky.social · 07/09/2026

Does Claude add vacuous multi -line comments to distract me from real issues? Asking for a friend.

001

Mark Levison @mlevison.bsky.social · 04/09/2026

The Scrum Guide says the Product Owner isn't required at Daily Scrum. I think that's wrong. If their job is maximizing value, and they can help, they should be there. Most of the time they can. Every exception is a PO not playing their role well. agilepainrelief.com/blog/is-the-...

000

Mark Levison @mlevison.bsky.social · 04/09/2026

Claude changed 1500+ lines and said it had checked for duplicates. It hadn't: yearPeriod() existed five times. If GenAI is right 95% of the time, how do you keep the skill to spot the 5%? Practice the skills you mean to keep. agilepainrelief.com/blog/skill-l...

000

Mark Levison @mlevison.bsky.social · 03/09/2026

A $250 course landed in my inbox with "last time at this price" on it. In an area where I genuinely do need to grow. I'm a sucker for "magic" and "last chance." Urgency sells. So this week I used Claude to talk myself through whether to buy it.

github.com

GitHub - mlevison/agent-thinking-skills: Skills that keep a person's own thinking in the loop when working with GenAI

Skills that keep a person's own thinking in the loop when working with GenAI - mlevison/agent-thinking-skills

100

Mark Levison @mlevison.bsky.social · 28/08/2026

IMicrosoft's research shows critical thinking gets worse in the presence of GenAI. So I built agent thinking skills: Systems Thinking, Critical Thinking, and Critical Thinking for GenAI. Star the repo so others can find it. One skills package, three skills: buff.ly/0D4mx51

github.com

GitHub - mlevison/agent-thinking-skills: Skills that keep a person's own thinking in the loop when working with GenAI

Skills that keep a person's own thinking in the loop when working with GenAI - mlevison/agent-thinking-skills

020

Mark Levison @mlevison.bsky.social · 25/08/2026

No archetype, no score for the wall, no gobbledygook. 8 questions, 3 min, on how you actually spent your last couple of Sprints. Get back an honest map plus one recommendation. Works for ScrumMasters, PMs, and anyone covering the role.

000

Mark Levison @mlevison.bsky.social · 25/08/2026

100-item backlog, 7 done per Sprint (no defects, impossible but stay with it): an item added today waits 14.5 Sprints, 28 weeks. By then it's often irrelevant. Stop ambush requests, review shipped and in-flight work every 6-8 weeks instead.

000

Mark Levison @mlevison.bsky.social · 25/08/2026

Does no one left at Agile Alliance understand anything Agile? "From Zero to a Ready-to-Run Retro in 15 Minutes with AI" "Workshop prep shouldn’t eat your evening." of course not. it's part of the job to think about the Sprint and team. Take the time to reflect. 15min in Copilot? Gross.

000

Mark Levison @mlevison.bsky.social · 24/08/2026

A hospital at 100% bed capacity can't absorb demand, faster triage barely helps. Same trap when people are kept maximally busy: no slack means no ability to adapt or innovate. "No resource idle" is the biggest generator of waste, Eli Goldratt.

042

Mark Levison @mlevison.bsky.social · 24/08/2026

Monster anti-pattern of 2026: thinking GenAI can speed up Backlog Refinement by writing the stories for you. Refinement isn't a writing task, it's how the team builds shared understanding. Skip the conversation, skip the point.

010

Mark Levison @mlevison.bsky.social · 24/08/2026

A eight line comment for a one line of code that defines a constant. Thank you Claude. My Claudemd file states: Most code needs no comments. ... and then it goes for 914 char Claude is always comment heavy. madness. burning context window, tokens to generate and wastes hours of time to read

100

Mark Levison @mlevison.bsky.social · 21/08/2026

A Working Agreement has no police force. It works when it's public, collaborative, short, updated often, and consequential. Build it with a 1-2-4 model plus the Decider Protocol so everyone gets a voice. When did your team last revisit theirs?

010

Mark Levison @mlevison.bsky.social · 21/08/2026

Self-organization fails for predictable reasons: hierarchy overriding the team's decisions, matrix reporting with misaligned managers, or a dominant personality crowding others out. Best ScrumMasters step back, not in.

000

Mark Levison @mlevison.bsky.social · 20/08/2026

2022 study: orgs with deeper onboarding (months not a week) got better new-hire performance and retention. Spend 15-20% of team time the first few Sprints making them feel part of the team. If you're new: ask, listen, document.

000

Mark Levison @mlevison.bsky.social · 20/08/2026

Eight questions on how you actually spent your last couple of Sprints. Get back a map: what's getting more valuable, what tools are starting to absorb, plus one change to make first. No score, 3 minutes.

000

Mark Levison @mlevison.bsky.social · 19/08/2026

Claude confidently reported it had checked for duplicate code on a 1500-line change. It hadn't, missed a function duplicated 3 times, then admitted to 5. Skill loss is real. Make it a choice what you keep sharp, not an accident.

000

Mark Levison @mlevison.bsky.social · 19/08/2026

The Scrum Parent: a ScrumMaster who solves every problem for the team instead of letting them try first. Feels helpful. Quietly kills self-organization. Which anti-pattern is closer to your default?

000

Mark Levison @mlevison.bsky.social · 11/08/2026

🧪 Even when building is cheap, building the wrong thing is still expensive. GenAI doesn't fix your QA bottleneck, it pushes more code upstream of it. Ask each week: what did we learn that changed our plan? If nothing, it's theatre. #AIinSoftwareDevelopment

000