GreyBrain Pulse · Published 23 Sept 2026
Make it argue against your working diagnosis
The clinician · What’s possible · F3
A rehearsal of the argument against yourself, and the name of the trap you were walking into.
Why this matters for your role
The clinician’s skill: ask which system, population, comparator and outcome were actually studied.
Your worked example
A fictional teaching exercise compares a task-specific imaging system evaluated in a screening programme with a general-purpose chatbot that has not been evaluated in that programme.
Put it into practice
Build a checklist covering intended use, population, comparator, outcomes and validation. Explain why evidence for the imaging system cannot establish the chatbot’s clinical performance. Do not diagnose or recommend treatment.
Clinician prompt
Make it argue against your working diagnosis
A rehearsal of the argument against yourself, and the name of the trap you were walking into.
This is a teaching exercise on a fictional case. You are not diagnosing anyone and neither am I. Fictional case: [PASTE A TEACHING CASE — presentation, findings, initial results. Invent it or use a published teaching case.] My working diagnosis is: [STATE IT] Do four things: 1. Build the strongest possible argument AGAINST my working diagnosis using only the findings given. 2. Give the three alternatives that best explain the same findings, and for each, the single finding that most argues for it and the one that most argues against. 3. Name the cognitive trap my working diagnosis is most vulnerable to here — anchoring, availability, premature closure or another — and say exactly where in the case I would have fallen into it. 4. List the discriminating tests in order of how much they would change the picture, not how easy they are to order. End with a short section titled "Your check" listing what I must verify myself before using any of this.
Synthetic practice example. Run it as-is and check the output against the supplied facts.
Your check
This is reasoning practice on fictional material. Clinical judgement stays with the clinician; nothing here is a diagnostic tool or a reason to change management.
Teaching cases only. Never enter a real patient’s presentation, findings or identifiers into a public AI tool.
Sources inform this fieldnote; their publishers do not endorse GreyBrain. This exercise does not confer clinical competence or a certificate.