Disagreeing with a Decision — Steelman First

P62.productive-disagreement.01 · Audience: guest, it-ml, language-pro

Strengths & Weaknesses Without Clichés Escalation Without Arson

LLM grading (this page):Real gradingOff — no LLM calls

A data scientist watches her team commit to a large language model for a task she is convinced a simple retrieval baseline would handle better, cheaper, and sooner. She has the evidence. She raises it in the planning meeting — and within two minutes the conversation is no longer about baselines at all; it is about whether she respects the tech lead's judgement, because her opening line was "this is over-engineered". Same week, different building: a teacher who opposed a rigid new timetable opened with "I can see why we're doing this" and was the only critic the head actually invited back. The difference was not who was right. It was who made the disagreement about the decision, and who accidentally made it about a person. This module teaches the first — starting with a move that feels backwards: arguing the other side's case before you argue your own.

🗺️ So far you have built: the answer-architecture craft (P60): hearing the question behind the question, structuring evidence with STAR, and sizing an answer to the listener — plus a career narrative (P61) worth defending. All of it assumed you and your listener wanted the same thing. This track drops that assumption: from here on, the other person disagrees with you.

In this module you will: the core discipline of productive disagreement: separating the decision from the person who made it; steelmanning — restating the other position at full strength before opposing it; anchoring the whole exchange to a goal you both share; and disagree-and-commit, the professional move that lets you lose an argument without losing your standing or sabotaging the outcome. The ladder closes it: steelman a decision you are given, then tell your own disagreement story so the other side stays credible.

Step 1 / 6 — The decision is not the person

Most disagreements go wrong in the first sentence, because the first sentence targets the wrong object. "This is over-engineered" sounds like a claim about a system, but the person who chose the system hears a claim about their judgement — and once someone is defending their judgement, they are no longer evaluating your evidence. The repair is mechanical, not saintly: aim every sentence at the decision, its inputs, and its consequences, and none at the decider. "I think this approach costs us three months" survives scrutiny; "you always overcomplicate things" starts a war and settles nothing.

ⓘ Concept: Attack surfaces: decisions have them, people defend them

target = the decision, its inputs, its consequences — never the decider's competence, motives, or character

Why it matters — A decision can be examined, weighed, and reversed without anyone losing face — that is what makes professional disagreement possible at all. A person, once they feel examined, will spend the meeting protecting themselves rather than the project, and they will usually win that contest because defending oneself is a deeper instinct than evaluating a baseline. Framing is therefore not politeness layered on top of the argument; it decides whether the argument can happen. The practical test for any sentence you are about to say: could the other person agree with it without admitting they are a worse professional than you? If not, rewrite it — the information can stay, the aim must change.

✅ Key takeaways

Aim every sentence at the decision — its inputs, costs, consequences — never at the decider. The test: could they agree with your sentence without admitting to being a worse professional? If not, rewrite it; the information stays, the target changes.
Steelman first: state the strongest honest version of the position you oppose and get its owner's 'yes, that's fair' before you counter. It proves you listened, maps what your argument must beat, and often sharpens or dissolves your objection on the spot.
Name the shared goal early — 'we both want G' — so both positions become candidate answers to one agreed question. Inside that frame, evidence beats rank; without it, disagreement is a contest of wills, and wills are settled by seniority.
Disagree and commit: full-strength argument while the decision is open, full-strength execution once it closes — with, on high-stakes calls, an openly agreed revisit tripwire. Safety, legal, and ethical objections are the exception: those escalate instead.
If time proves you right, spend the vindication on the process ('what check would catch this earlier?'), never on the person. And in any retelling — including interviews — the other side stays a credible professional, or the story defeats its own point.

Next: P62.productive-disagreement.02 — Escalation Without Arson

None of this requires a technical field. The steelman works on timetables, translation briefs, seating charts, and model architectures alike — wherever a decision you doubt is owned by a person you must keep working with, which is to say: everywhere you will ever work.

Ask the mentor about this module

Ask a question about this content. The mentor explains and grounds its answer in what you are studying; asking is recorded as a learning signal, not a grade.

Images, PDF or text. Kept on this device only.

Keeping your files on this device

Keep a copy of the files I attach, in this browser

Off by default. The mentor always gets your file; this only decides whether your own copy stays here. Copies live in this browser only - they do not follow you to another device, and clearing site data removes them.

Ctrl/Cmd + Enter to send

🎓 Practice ladder

2 graded rungs · ~20 min

Two mentor-graded craft rungs. Rung 1 gives you a decision to disagree with — a school's rigid new timetable — and requires a confirmed-quality steelman before your counter-case. Rung 2 is your own disagreement story, told so the other side stays credible; a low-stakes story is always an accepted substitute. The mentor grades the telling, never your history.

Rung 1 — Steelman, then disagree (explorer)

Loading exercise…

Rung 2 — Your disagreement, their credit (practitioner)

Loading exercise…

Try it yourself

A scratch console for this page's ideas — ungraded, nothing you run here is recorded.

Scratch console

A scratch console with the scientific stack (pandas, numpy, scikit-learn). Runs on the server — no network, resource-limited and measured.

Output appears here.

My notes on this module

Loading your notes...

Where next?

Continue

Escalation Without Arson →

Productive Disagreement · P62

Later in Productive Disagreement

This module unlocks

Influencing Without Authority

Go up a level