https://edward-schaefer.com/field-notes/ Field notes from the actual work — what holds up, and what breaks, when a large organization takes on AI. https://edward-schaefer.com/field-notes/the-bug-that-taught-me-about-ai-cost-governance/ https://edward-schaefer.com/field-notes/the-bug-that-taught-me-about-ai-cost-governance/ An ordinary bug burned through millions of tokens in a day. The spend cap that finally stopped it was only the smoke alarm. The real fix was the sprinkler system underneath. https://edward-schaefer.com/field-notes/the-steadiness-test/ https://edward-schaefer.com/field-notes/the-steadiness-test/ A leader goes quiet under bad news, and people often read that as composure. More often it's broicism: real anxiety held rather than examined, tending to leak out sideways onto whoever is standing closest. https://edward-schaefer.com/field-notes/human-in-the-loop-is-a-coaching-problem/ https://edward-schaefer.com/field-notes/human-in-the-loop-is-a-coaching-problem/ Approving every move an agent makes feels responsible, until the approver becomes the bottleneck. This essay reframes human-in-the-loop as a coaching relationship, where agents earn autonomy the way people do, instead of a gate every action has to clear. https://edward-schaefer.com/field-notes/nobody-is-wip-limiting-the-human/ https://edward-schaefer.com/field-notes/nobody-is-wip-limiting-the-human/ Agents never sleep, so work in progress no longer has the natural ceiling that human labor used to put on it. This essay borrows the oldest idea in kanban, limiting work in progress, and applies it to the one constraint that never got faster: human attention. https://edward-schaefer.com/field-notes/comprehension-debt-is-a-debt/ https://edward-schaefer.com/field-notes/comprehension-debt-is-a-debt/ None of the code here gets written by hand anymore: agents write every line, and the job left is pointing the work and judging what comes back. Comprehension debt is the gap between what a system actually is and what anyone can still account for, a debt that can be managed but never fully paid down. https://edward-schaefer.com/field-notes/coaching-agents-isnt-coaching-people/ https://edward-schaefer.com/field-notes/coaching-agents-isnt-coaching-people/ A new model made every other agent Ed runs better too, and the credit belonged to the setup around them, not the model itself. This essay asks how much of the coaching craft actually transfers to agents, and where improvement still has to be built by hand. https://edward-schaefer.com/field-notes/before-you-schedule-a-skip-level-ask-this-first/ https://edward-schaefer.com/field-notes/before-you-schedule-a-skip-level-ask-this-first/ Senior leaders often feel an itch to sit down one-on-one with someone well below them on the chart. This essay treats that itch as a signal pointing at a specific problem, not a meeting to schedule before you know which one. https://edward-schaefer.com/field-notes/emotional-acuity-the-leadership-capability-most-vps-never-learned/ https://edward-schaefer.com/field-notes/emotional-acuity-the-leadership-capability-most-vps-never-learned/ Why leaders need to read emotional signals fast: the case for emotional acuity at the VP level, and a practice ladder for building it. https://edward-schaefer.com/field-notes/tone-in-digital-communication-the-leadership-skill-most-managers-never-practice/ https://edward-schaefer.com/field-notes/tone-in-digital-communication-the-leadership-skill-most-managers-never-practice/ Why digital messages get read as attitude, and what leaders can actually control about it, now that AI drafts so many of them. https://edward-schaefer.com/field-notes/hard-lessons-from-the-worst-manager-ive-ever-had/ https://edward-schaefer.com/field-notes/hard-lessons-from-the-worst-manager-ive-ever-had/ What a self-important first boss taught me about respect, developing people, and why performance conversations should help, not punish. https://edward-schaefer.com/field-notes/hard-lessons-from-the-worst-performance-review-ive-ever-received/ https://edward-schaefer.com/field-notes/hard-lessons-from-the-worst-performance-review-ive-ever-received/ A performance review that blindsided me after years of strong ratings taught me the one rule managers should never break: no surprises, ever. https://edward-schaefer.com/field-notes/surround-yourself-with-courageous-truth-tellers-not-sycophantic-yes-men/ https://edward-schaefer.com/field-notes/surround-yourself-with-courageous-truth-tellers-not-sycophantic-yes-men/ Why leaders end up surrounded only by agreement, what it costs them, and the specific habits that make it safe for someone to tell them the truth. https://edward-schaefer.com/field-notes/the-immense-risks-of-superficial-knowledge-gained-from-frameworks-like-safe/ https://edward-schaefer.com/field-notes/the-immense-risks-of-superficial-knowledge-gained-from-frameworks-like-safe/ Why fluency in a framework like SAFe isn't the same as skill, what that gap costs teams, and what Shu-Ha-Ri explains about the trap of stopping there. https://edward-schaefer.com/field-notes/the-career-delta-file/ https://edward-schaefer.com/field-notes/the-career-delta-file/ What a Career Delta File is, how a decade of keeping one has paid off, what the gaps cost, and why agent memory has to solve the same problem. https://edward-schaefer.com/field-notes/the-review-is-not-a-demo/ https://edward-schaefer.com/field-notes/the-review-is-not-a-demo/ Most sprint reviews are only demos. What the review is actually for, and how to get the real one back. https://edward-schaefer.com/field-notes/expertise-shifted-left/ https://edward-schaefer.com/field-notes/expertise-shifted-left/ Experts increasingly put their discernment into designing and validating the systems agents run, not into the work itself. This essay asks where the next generation of experts will develop that discernment, now that agents do the formative work people used to learn on. https://edward-schaefer.com/field-notes/ai-adoption-was-the-easy-part/ https://edward-schaefer.com/field-notes/ai-adoption-was-the-easy-part/ AI adoption already happened in most organizations, so the real risk moved from the people who won't start to the ones who won't wait, the sprinters. This essay argues you can't govern that fast tail with a communication plan, only by putting the practice into the substrate the agents actually run on. https://edward-schaefer.com/field-notes/the-work-still-got-done/ https://edward-schaefer.com/field-notes/the-work-still-got-done/ When an agent takes over the labor, the same change lands as a win for one person and a loss for another, and the loss is now measurable. This essay looks at what decides which one a person gets, and why relocating your identity from the task itself to what survives it is what converts the loss into gain. https://edward-schaefer.com/field-notes/how-your-culture-deals-with-failure/ https://edward-schaefer.com/field-notes/how-your-culture-deals-with-failure/ When something breaks, whether an organization hunts for someone to blame, follows the rulebook, or asks what let the failure happen reveals which of Ron Westrum's three cultures it actually runs, and what accountability really means there. https://edward-schaefer.com/field-notes/what-happens-to-messengers/ https://edward-schaefer.com/field-notes/what-happens-to-messengers/ Whether an organization shoots, neglects, or trains the person who brings it bad news reveals which of Ron Westrum's three cultures it actually runs, and what everyone else learns about speaking up next time. https://edward-schaefer.com/field-notes/how-your-culture-handles-bridging/ https://edward-schaefer.com/field-notes/how-your-culture-handles-bridging/ Every organization handles cross-team help somewhere between discouraged, tolerated, and encouraged. How to tell which one you run, and how to make the favor economy visible without punishing the people who keep it working. https://edward-schaefer.com/field-notes/who-is-watching-your-agents/ https://edward-schaefer.com/field-notes/who-is-watching-your-agents/ The human oversight capacity gap is the distance between the oversight your agents now require and the oversight your organization can actually supply, and closing it starts with mapping where that gap is thinnest. https://edward-schaefer.com/field-notes/the-culture-underneath-the-work/ https://edward-schaefer.com/field-notes/the-culture-underneath-the-work/ Where Westrum's three-cultures model came from, how it reached software, what the evidence honestly supports, and the permission it grants: to change your culture, accept it, or finally have it confirmed. https://edward-schaefer.com/field-notes/how-your-culture-shapes-cooperation/ https://edward-schaefer.com/field-notes/how-your-culture-shapes-cooperation/ Cooperation runs low, modest, or high depending on the culture around it. Why organizations that say they want cooperation still restrict it, what the warm silo costs, and when swarming teamwork backfires. https://edward-schaefer.com/field-notes/does-coaching-work/ https://edward-schaefer.com/field-notes/does-coaching-work/ What four meta-analyses actually show about coaching, why the headline number needs a footnote, what predicts a good outcome, the side effects, and why the ROI multiples cannot be trusted https://edward-schaefer.com/field-notes/how-to-evaluate-a-coach/ https://edward-schaefer.com/field-notes/how-to-evaluate-a-coach/ How to evaluate a coach in three moments, before the fit call, during it, and after the first sessions, and why the working alliance predicts a good outcome better than credentials or personality match https://edward-schaefer.com/field-notes/the-decision-before-you-hire-a-coach/ https://edward-schaefer.com/field-notes/the-decision-before-you-hire-a-coach/ Whether coaching is the right kind of help at all: what a coach actually does, when a mentor, advisor, or consultant is the better purchase, and the two questions to answer before you interview anyone https://edward-schaefer.com/field-notes/the-agent-standup-practice-note/ https://edward-schaefer.com/field-notes/the-agent-standup-practice-note/ What the agent standup is, what it isn't, and how to run it every day https://edward-schaefer.com/field-notes/when-the-standard-moves/ https://edward-schaefer.com/field-notes/when-the-standard-moves/ When agents do the work, your standard moves. How to tell which standards were priced by scarcity and can move, which were priced by consequence and must hold, and how to write the standard where the work runs https://edward-schaefer.com/field-notes/where-your-culture-puts-the-risk/ https://edward-schaefer.com/field-notes/where-your-culture-puts-the-risk/ When a dependency slips, who is exposed? Westrum's responsibilities row, from shirked to narrow to shared risk, and the measurement move that puts the owner's name next to the waiting team's date https://edward-schaefer.com/field-notes/consistency-isnt-the-models-job/ https://edward-schaefer.com/field-notes/consistency-isnt-the-models-job/ AI models are built to vary. When you need the same result every time, have an agent build the script that does the work, and keep the model for the work that needs judgment https://edward-schaefer.com/field-notes/where-your-culture-sends-new-ideas/ https://edward-schaefer.com/field-notes/where-your-culture-sends-new-ideas/ An approved idea can still die in process. Westrum's novelty row, from crushed to problems to implemented, and how to design a first try so a rough week reopens the plan and the idea keeps its yes https://edward-schaefer.com/field-notes/more-of-a-guideline-than-a-rule/ https://edward-schaefer.com/field-notes/more-of-a-guideline-than-a-rule/ A sprint rule or a say/do ratio is often an easy way to say no without thinking it through. Why frameworks get used that way, what it costs teams, and what it looks like when teams handle it well