Which Model Gets to Read the Chart?
Physicians are building their own AI tools, and patient data cannot leave the building. So which local model deserves to be in the room with a chart? Not the biggest one. I measured.
Yoram Friedman, MD
(121)Physician and enterprise product leader. Author of the Agentic AI for Product Leaders series. Writes on AI, healthcare, and what changes when software starts acting. 15+ years at SAP, Walmart, and regulated health tech.
Only the Developers Have Met the Agent
A PM who has never watched an agent work cannot write its spec, its evals, its boundaries, or its intent. The pod model assumes PM, UX, and engineering share an understanding of the product's behavior. For agentic products, only one of the three has it.
The Anticipated Question: Data Products Meet Their Second Customer
Data products were built on a quiet assumption: the designer could anticipate the questions. Agents break that assumption without breaking the discipline. What survives is not the shape we prepared but the honesty of the contract underneath it. Five questions for every data product owner.
The Hospital Wrote the Rulebook. Everyone Else Is About to Read It.
Healthcare asks who is accountable for a machine's mistake before it asks what the machine can do, because it has no choice. On 18 August a federal regulator applied that question to generative AI. The framework it produced is not new. It is clinical governance, written down for everyone else.
The Command Line
The word "prompt" was not invented for AI. It was on the screen in 1993, the C colon and the blinking line, waiting for you to learn the machine's language. We are teaching doctors DOS again. The interface will absorb the commands, the way it always does. What it cannot absorb is judgment.
Where the Time Came From
People keep asking how I found time to write four books in a year while doing my day job. The answer is not a tool and not discipline. It is a layered setup where every piece of work is tagged, linked, and ready for the next one. The tools are rented. The brain underneath is the asset.
The User Did Not Leave. She Split in Two.
The operator did not vanish when the agent arrived. She split. What she knew became the agent's context and bounds. What she checked before committing became the supervisory surface. The only casualty was the dashboard, and most agents are being built to generate it.
The Labels Came Off the Pyramid
A new paper rebuilds the DIKW pyramid for medicine and shows where AI actually breaks. Strong on the arrows nobody worried about, weak on the one the structure depends on. The pyramid always assumed you could tell which layer you were standing on. That assumption is gone.
Tesla Removed the Steering Wheel. That Was the Honest Part.
A Cybercab was filmed with no steering wheel. Removing it was the honest part. The rungs below it, hands on the wheel, eyes on the road, ready to take over, were never supervision. They were accountability. Most enterprise AI is still on rung three.
Your Engineers Are Writing the Policy Now
Engineering changed its unit of work to intent, policy, and boundaries. Product and design mostly still ship user stories and screens. So the policy gets written by the last person to touch the ticket, in a prompt, at midnight. The control plane has inputs, and nobody upstream owns them.
Earn Your Way Out of the Loop. Then Show the Receipt.
The Claude Code team ships most of its code through an agent that acts on its own. The six lessons everyone is copying are real. But in a regulated enterprise, earning your way out of the loop only counts when the receipt is an audit trail, not one person's memory.
The Sixth Archetype Anthropic Forgot
Anthropic mapped product work into five archetypes: Prototyper, Builder, Sweeper, Grower, Maintainer. All five stop the moment the thing ships. But an agent keeps deciding after you let go of it, and there is a sixth archetype no one drew: the one who watches it decide.