Writing
Long-form thinking on healthcare AI, emergency medicine, clinical safety, and how to read evidence honestly. Filter by field below.
AI Scribes Are Not the Endgame
AI scribes solve a real documentation problem. But calling them co-pilots confuses transcription with clinical reasoning — and the gap matters.
Why Vague Symptoms Are Where Medicine Gets Dangerous
Vague presentations defeat protocols, degrade triage, and exploit pattern recognition's blind spots. How emergency medicine actually reasons under ambiguity.
Clinical Safety Is Not a Checkbox
The test of safety work is whether it changed the product. If the hazard log was written and nothing was redesigned, what happened was compliance theatre.
How to Read a Medical Paper Without Being Hypnotised by the Abstract
Read methods first, numbers second, the authors' opinions third, and the abstract last — because the abstract is the paper's advertisement, not its evidence.
Why Doctors Need to Understand AI (Before AI Reorders the Profession)
AI now operates on clinical language and reasoning, and the defaults are being set now — understanding it is how doctors get to set the terms instead of inheriting them.
Screening Is Not Always a Gift: The Arithmetic That Flatters Early Detection
"Early detection saves lives" is the most intuitive sentence in medicine — and one of the easiest to prove without proving anything at all.
The Confident Wrong Answer: Safety Thinking for Clinical AI
Traditional clinical software fails in ways you can anticipate. AI fails differently — fluently, confidently, and most dangerously when it is wrong. Safety thinking has to change to match.
Who Is Your Clinical Safety Officer — and Why "Nobody, Really" Is the Wrong Answer
Many digital health products have a named clinical safety officer and no real one. The gap between the title and the function is where safety quietly stops happening.
Compensation Hides the Crash: Why Normal Vital Signs Are the Most Dangerous Reading in the Room
The body is built to mask its own emergencies — and the numbers are the last thing to tell you it has run out of road.
Hazard Is Not Risk — and Confusing Them Is How Digital Health Ships Harm
Two words that digital health teams use interchangeably mean very different things. The confusion isn't pedantic — it's the mechanism by which real harm gets reasoned away.