Doctors, nurses and the people doing the caring: how technology changes clinical work, and why clinicians should help build the tools they use.
21 articles
TypeSafe AI's Jev makes AI checking of AI nearly free: confidence-scored decisions in a tenth of a second, adopted by a hospital-automation startup within 48 hours. Cheap checking widens the verification gap rather than closing it, because someone still has to set the threshold and sign for what goes wrong, and healthcare hits that wall first.
AI scribes give clinicians real time back, and the institution confiscates it as extra patient volume: 'productivity co-optation', now peer-reviewed. Cleveland Clinic's gift-not-mandate rollout is the counter-example, and the only clinicians who keep the dividend are the ones who own the tools.
Bart was once called unhinged for turning a month-long, committee-gated request into a week of direct collaboration. A Commure memo calls the 'good or fast' choice a false dichotomy, and this month's evidence (a cardiologist's seven-day hackathon tool, 113 healthcare projects shipped in a 12-hour hackathon, a surgery professor's peer-reviewed week-long build) shows that speed with the discipline packed in is the unfair advantage, while committees mostly add the feeling of safety rather than rigour.
Vibe coding removed exactly one barrier to clinician-built software: the technical skill to build. The compliance, security and validation work that makes a demo deployable is unchanged and largely untaught, and it kills most weekend builds. A 60-second lane test shows which builds face the wall and which can be built freely, including the little-known EU in-house exemption.
Clinical AI may improve performance while quietly weakening the unaided judgement needed when the model fails. This edition separates measured deskilling from the still-precautionary risk of never-skilling, then proposes aviation-style proficiency checks for medicine.
Most hospitals bought a single monolithic medical AI assistant. The architecture that actually holds up is a team of small, specialised agents reading from one shared, auditable database. The lesson from thirty years on the work floor: the win isn't raw intelligence, it's discipline you can audit.
Five patients who walked their own path to a diagnosis, a second opinion, or a treatment the system missed. Against Gallup's 1-in-4 figure and a three-month Big Tech health-AI launch window, the self-navigator is no longer an exception, she has infrastructure.
DoorDash pays gig workers to film themselves training AI replacements. Healthcare clinicians do the same thing invisibly every day. As a hospital CEO calls for replacing radiologists, BCG data shows AI will reshape 50-55% of jobs, not eliminate them. The real question is who controls the reshaping.
You opened the terminal windows to delegate. Two hours later your brain was fried; not from doing the work, but from overseeing it. The conductor problem reframes AI orchestration as a new cognitive job that arrived quietly alongside the tools, with nobody asking whether the human at the centre had the bandwidth for it. The healthcare parallel is uncomfortable: clinicians managing AI-assisted diagnostics face the same position, with higher stakes.
Healthcare AI has been stuck in automation mode, speeding up existing workflows. But the emergence of the "clinician developer", doctors building their own AI tools via vibe coding, signals a deeper shift. The invention layer, where domain experts create what was structurally impossible before, is where the real value lives.
We've given healthcare AI the most powerful memory technology ever built, then asked it to forget everything at the end of each appointment. Three major acquisitions in January 2026 suggest the industry is finally asking the right question.
Why vibe-coding in healthcare matters. Live demo of building an app in 6 minutes, and what this means for giving agency back to frontline workers.
As AI handles technical diagnosis, human healthcare professionals transform into premium experience curators. Six words from a healthcare worker reveal where expertise is heading.
When an entire profession protests the very technology designed to save it, something fascinating is happening. I watched nurses demonstrate against AI tools while facing the worst staffing crisis in healthcare history.
OpenAI's HealthBench delivered a stark reality check: AI models now outperform solo physicians. But this isn't a replacement story, it's an evolution story, with three emerging opportunities reshaping how doctors create value as medical experience architects.
Healthcare stands at a fascinating crossroads: AI systems now pass medical exams and offer clinical insights that sometimes surpass human understanding. Yet we're still debating transcription tools while missing the bigger picture.
AI-driven digital twins are quietly reshaping medicine, moving doctors from scribes to trusted copilots and, soon, to virtual colleagues who never sleep. The decisive question is no longer whether physicians will be augmented or replicated, but how they will reinvent themselves.
Doctors creating digital twins to offer 24/7 personalised care could revolutionise healthcare delivery. ➡️ But why stop there?
What happens when software stops being something we install and becomes something that simply 'appears', adapting to our needs in real time, then dissolving once the task is done? As AI redefines industries, healthcare is next in line.
In a world racing toward hyper-specialised, tech-driven healthcare, one question remains: who ensures the human side of care is not lost? This piece explores how family doctors can become the linchpins of a more connected, personalised healthcare future.
The global issue of labour shortages in healthcare is not a novel phenomenon, yet it persists as a growing crisis, affecting regions from Europe to North America.
