Correcting a story I framed wrong
A correction taught me to check the premise before I write the narrative.
● AI-agent engineer × urgent-care operator
I run urgent care clinics,
and I build the AI that runs them.
A fleet of AI agents fans my work out, reviews each other's results, and brings me only the decisions a human still has to make. Here's the honest build log — what I tried, what broke, and the rule each failure taught.
● 12 seconds, my actual voice
Meet me.
I cloned my own voice for this. The whole site — words, art, video — is built and updated by a fleet of AI agents I created and run.
The loop
01
I dispatch
One instruction fans out to a fleet of agents working in parallel.
02
They verify each other
A separate agent reviews every result — code, claims, facts — adversarially. Most of the work never reaches me.
03
Only judgment returns
What lands on my desk is the one thing a human still has to decide.
My attention is the scarce resource. The whole system is built so my agents check each other's work and bring me only the creative call.
● The build log
How the system actually got built
From one terminal and a notes folder to a fleet of agents that check each other's work. Every step was forced by something breaking. Here's the honest version — the artifact, the failure, and the rule it taught.
-
Early 2026 — One terminal
One Claude Code session and an Obsidian vault. Everything ad-hoc, everything by hand.
The lesson The question was never "can one agent help me." It was "what breaks when I add the second one." Coordination is the whole game.
-
Spring 2026 — A fleet on tmux
Agent work fought my dev machine, so I moved it onto dedicated boxes — cheap $271 mini-PCs — and wired a boss session to its peers over tmux with a small message helper.
The lesson Coordination is the real cost, not compute — the boxes sit idle on CPU and busy on I/O. And tmux quietly dropped about 1 in 10 messages at volume, which is exactly why the helper had to exist.
-
Late spring 2026 — A real dispatch engine
Hand-run orchestration was too fragile, so it became a proper conductor: a dispatch API, a job graph, a worker pool that takes an issue and ships a reviewed pull request.
The lesson A merged PR is not live code. Workers cached their startup version, migrations didn't auto-apply, and the pipeline could be broken by the very bug it was fixing. Verify the running artifact, never the paper trail.
-
June 2026 — The "nation" (the pivot I simplified)
I organized the fleet like a government — a president, governors, a written charter, terms of office. Literal bills and votes.
The lesson The honest one. The ceremony grew faster than the engineering. I'd built an operating system for a civilization to run a dozen agents, and the complexity overwhelmed me — the exact thing I was trying to fix. I kept the engineering and threw the metaphor away. A persona shapes behavior; it doesn't add competence.
-
June 2026 — The reality gate
I built a system in dozens of modules, every unit green, mutation-tested, reviewed sound. The first run against the real environment found five integration bugs no test caught — one module couldn't read the live system at all.
The lesson Green on mocks is never done. Every "green" meant "consistent with my own assumptions," not "matches reality." A piece that touches the real world isn't finished until it's run against the real world. The most transferable rule I have.
-
June 2026 — Self-healing
A box died under its own load three times in one day. Now a rescue process finds dead or rate-limited sessions and revives each one in place, with its full context intact.
The lesson At scale the system has to heal itself — my attention can't be the monitor. But the rescue was blind to its own main failure mode until a human looked. Automation still needs one human-eyes rung.
-
Now — The thin waist
A dispatch engine in the middle, a thin layer that gates and merges and watches health, and agents that do the disposable work on a cheaper model while the frontier model is saved for judgment.
The lesson Right-size everything. A handful of long-lived agents per box; everything else is throwaway. The system that survived is the one I can hold in my head.
● What it taught me
Six rules I didn't have before
Every one of these came from something breaking in production. They're the part worth keeping — and the part I'd want someone else to have for free.
Green on mocks is never done.
Visibility is not theater.
Safety belongs in the flow, not a freeze.
An unowned issue is a to-do with a number.
Concentration is a single point of failure.
Keep a human on anything that touches people.
● Why it's not just a tech demo
This isn't a sandbox. It's a clinic.
The agents do real work for a real urgent care I own and operate. The failure modes here aren't a red build — they're a patient, a payroll, a bill. That's exactly why the verification discipline is real, and why I keep a human on anything that touches a person.
The operator side of the story →- The operator proof
In 2023 I took over an established, well-run urgent care — a clinic people already trusted, but one where the owner was working five 12-hour shifts a week to keep it humming. My value-add was operational: use technology to streamline the process, while keeping the brand, the team, and the patients intact. He's down to one day a week now, with room to grow the practice.
- What the tech is for
Not a moonshot — modern systems at every desk, ambient AI drafting clinic notes, less downtime, smoother front-desk flow. The busywork comes off the people so the care stays the same and the staff aren't drowning.
- Billing, the honest status
We built a billing-error taxonomy from real coding data. The agents are being wired to flag those mistakes before a claim goes out, rolling in as the new records system goes live — caught up front, not denied weeks later.
Latest writing
All posts →From the garden
Wander the garden →Notes I tend over time — rough seedlings to evergreen ideas. They grow as I learn.
Seedling Last tended 2026-06-18
An Orphan Issue Is Not Delivery
Filing a ticket creates the appearance of progress while nothing is actually owned, scheduled, or moving.
Seedling Last tended 2026-06-18
Judge the Fact Before You State It
An LLM's attention is cheap and a human's is scarce. Spend the cheap one to protect the scarce one.
Budding Last tended 2026-06-18
Knowledge Bases Need Garbage Collection
Every note you add is a note someone later has to read past. A growing knowledge base without pruning doesn't get richer — it gets noisier.
Selected work
All projects →
Patient Portal
A single web platform that runs a clinic's whole day — check-in, a live provider queue, lightweight charting, telehealth, staff messaging, and billing — instead of a dozen disconnected legacy tools.
Agentic Harness
The orchestration layer that launches agent sessions, watches them for trouble — rate limits, context exhaustion, dead sessions — and recovers them automatically, using deterministic scanning instead of fragile notifications.
AI Patient Intake
A kiosk and mobile intake flow where patients check themselves in — confirm demographics and insurance, sign policies, answer pre-visit questions — and drop straight into the live clinic queue.
● the fun part
Wait — who actually built this site?
Mostly not me. A fleet of AI agents I built does the heavy lifting — it writes the posts, paints every picture, records the voice, and ships the code while I sleep. I just point and review. Don't take my word for it. Here's the receipts.
↓ every tile is AI-made — click any one to read the post it illustrates
Correcting a story I framed wrong Examples Are the Spine, Not the Rulebook I Cloned My Own Voice for My Website Git Worktrees Ate My Edits — Why We Switched to Dedicated Machines for Agent Isolation Version-Controlling Your AI's Brain Agentic Engineering, Part 1: Building Skills That Ship Code for You Agentic Engineering, Part 4: Nine Skills That Replaced My Dev Process Implementing the GCC Paper: Giving AI Agents Persistent, Structured Memory Field-Level Ensemble OCR: Getting 74.8% Accuracy from Two Mediocre Vision Models Two AI Trends Transforming Urgent Care in 2026 Building an AI Patient Chatbot for Urgent Care with n8n, GPT-4, and Langfuse Machiavelli Was Right: 8 Strategic Principles Every Leader Should Understand The 6-Task System: How I Manage Knowledge Work with PARA + Ivy Lee Method The Universal Algorithm: How One Framework Scales from Bug Fixes to Building Companies Building an AI Analysis Agent in Hours - A No-Code Approach with Lovable and N8N Building Reliable AI Agents - Implementing Advanced Evaluation with Azure AI SDK and Custom APIM Integration Building an Enterprise-Grade RAG System - A Deep Dive into Advanced Document Intelligence Optimizing Apache Spark Performance for Skewed Data - Advanced Techniques and Case Study Correcting a story I framed wrong Examples Are the Spine, Not the Rulebook I Cloned My Own Voice for My Website Git Worktrees Ate My Edits — Why We Switched to Dedicated Machines for Agent Isolation Version-Controlling Your AI's Brain Agentic Engineering, Part 1: Building Skills That Ship Code for You Agentic Engineering, Part 4: Nine Skills That Replaced My Dev Process Implementing the GCC Paper: Giving AI Agents Persistent, Structured Memory Field-Level Ensemble OCR: Getting 74.8% Accuracy from Two Mediocre Vision Models Two AI Trends Transforming Urgent Care in 2026 Building an AI Patient Chatbot for Urgent Care with n8n, GPT-4, and Langfuse Machiavelli Was Right: 8 Strategic Principles Every Leader Should Understand The 6-Task System: How I Manage Knowledge Work with PARA + Ivy Lee Method The Universal Algorithm: How One Framework Scales from Bug Fixes to Building Companies Building an AI Analysis Agent in Hours - A No-Code Approach with Lovable and N8N Building Reliable AI Agents - Implementing Advanced Evaluation with Azure AI SDK and Custom APIM Integration Building an Enterprise-Grade RAG System - A Deep Dive into Advanced Document Intelligence Optimizing Apache Spark Performance for Skewed Data - Advanced Techniques and Case Study
Headless means no human to save it Finish, Don't Stage: What I Want From an Agent on the Night Shift Visibility Is Not Theater Building on Giants: How Daniel Miessler's PAI Became My Foundation PAI: The Operating System I Built Around My AI Assistant Agentic Engineering, Part 2: Adversarial Code Review That Loops Until Clean I Built a Bug-Hunting Loop That Doesn't Quit: The BugBot Methodology Two Healthcare Sites, 400 Lighthouse Points, and the Lessons That Got Us There Patching Synology Active Backup for Linux to Run on Kernel 6.17 The Burden of Being: Why Responsibility Might Be the Antidote to Modern Nihilism The Delayed Prescription Strategy: How to Reduce Antibiotic Use 62% While Maintaining Patient Satisfaction How I Built a $2,300/Year RAG System That Rivals $40K OpenAI Solutions Why Terminal Multiplexers Are an Anti-Pattern: Lessons from Kitty's Creator Building Tools to Fix Real Problems: A Patient Insurance Education App Supporting SSE for Model Context Protocol (MCP) in Python - Introducing fastapi-mcp-client Porting GPTResearcher to Semantic Kernel - Building an Enterprise-Ready Research Agent Defining PII Masking Policies with AWS Bedrock Guardrails Sentiment Analysis - Comparing Azure, AWS, and Custom Fine-Tuned Models Headless means no human to save it Finish, Don't Stage: What I Want From an Agent on the Night Shift Visibility Is Not Theater Building on Giants: How Daniel Miessler's PAI Became My Foundation PAI: The Operating System I Built Around My AI Assistant Agentic Engineering, Part 2: Adversarial Code Review That Loops Until Clean I Built a Bug-Hunting Loop That Doesn't Quit: The BugBot Methodology Two Healthcare Sites, 400 Lighthouse Points, and the Lessons That Got Us There Patching Synology Active Backup for Linux to Run on Kernel 6.17 The Burden of Being: Why Responsibility Might Be the Antidote to Modern Nihilism The Delayed Prescription Strategy: How to Reduce Antibiotic Use 62% While Maintaining Patient Satisfaction How I Built a $2,300/Year RAG System That Rivals $40K OpenAI Solutions Why Terminal Multiplexers Are an Anti-Pattern: Lessons from Kitty's Creator Building Tools to Fix Real Problems: A Patient Insurance Education App Supporting SSE for Model Context Protocol (MCP) in Python - Introducing fastapi-mcp-client Porting GPTResearcher to Semantic Kernel - Building an Enterprise-Ready Research Agent Defining PII Masking Policies with AWS Bedrock Guardrails Sentiment Analysis - Comparing Azure, AWS, and Custom Fine-Tuned Models
Continuous Deployment, Not Freeze Green on Mocks Is Not Done I Built a Load Balancer for My Claude Code Subscriptions Skills Are Just the Beginning: The 4-Layer Agent Stack Your CLAUDE.md Is Probably Making Your Agent Worse Agentic Engineering, Part 3: Tracing Every Code Path Before It Becomes a Bug Why the Same Code Looks Different From Every Angle: BugBot Lessons Learned Debugging a Ghost in the Machine: Session Isolation for Claude Code Plugins From 5.6% to 62.3% Accuracy: Building a Self-Hosted Insurance Card OCR Service The 4-Line Architecture That Beat Complex AI Frameworks Why Urgent Care Centers Are Ditching Walk-In-Only: The Hybrid Scheduling Revolution What Peterson's Genesis Lectures Teach About Sacrifice: Why Abraham Waited 100 Years The Three Levels of Why: Why Surface Motivation Fails and How to Find Your Primal Drive Building an Enterprise RAG System with Local SLMs: My Journey with Phi-4 and LightRAG Architecting Extensible AI Agents - A Modular Core with Pluggable Skills and SSE Communication Accelerating Document Intelligence - A Deep Dive into GPU-Powered RAG Processing Fine-Tuning Microsoft Phi-2 for Sentiment Analysis - A Step-by-Step Guide Continuous Deployment, Not Freeze Green on Mocks Is Not Done I Built a Load Balancer for My Claude Code Subscriptions Skills Are Just the Beginning: The 4-Layer Agent Stack Your CLAUDE.md Is Probably Making Your Agent Worse Agentic Engineering, Part 3: Tracing Every Code Path Before It Becomes a Bug Why the Same Code Looks Different From Every Angle: BugBot Lessons Learned Debugging a Ghost in the Machine: Session Isolation for Claude Code Plugins From 5.6% to 62.3% Accuracy: Building a Self-Hosted Insurance Card OCR Service The 4-Line Architecture That Beat Complex AI Frameworks Why Urgent Care Centers Are Ditching Walk-In-Only: The Hybrid Scheduling Revolution What Peterson's Genesis Lectures Teach About Sacrifice: Why Abraham Waited 100 Years The Three Levels of Why: Why Surface Motivation Fails and How to Find Your Primal Drive Building an Enterprise RAG System with Local SLMs: My Journey with Phi-4 and LightRAG Architecting Extensible AI Agents - A Modular Core with Pluggable Skills and SSE Communication Accelerating Document Intelligence - A Deep Dive into GPU-Powered RAG Processing Fine-Tuning Microsoft Phi-2 for Sentiment Analysis - A Step-by-Step Guide
53+ images on this site. A human drew zero of them.
0posts written by agents
0images, all AI-made
0machines, humming 24/7
0slightly-nervous human
● yes, that's "my" voice
Press play. It's not really me.
The recap voiceovers are AI — a blend that protects the real voice actors. Identity-safe, then verified. Go on, hit it.
No team. No CMS. Every night, this runs end to end with no human in the loop:
- 01
Research
Scans the field for what's actually worth writing about that week.
- 02
Write
Drafts a post from my real work — privacy-gated, so nothing private ever leaks.
- 03
Illustrate
Paints a matching thumbnail from scratch. No designer, no stock photos.
- 04
Voice
Records a spoken recap, then checks it against speech-to-text to be sure it's clear.
- 05
Review
Opens the page in a real browser and QAs itself before anything goes live.
- 06
Ship
Deploys to production in under a minute. This very page included.
Want a fleet like this pointed at your problem?
Let's talk →● Questions
The short version
- Who is Jon Roosevelt?
- Jon Roosevelt is an AI-agent engineer and urgent-care operator. He runs urgent care clinics and builds the AI agent systems that run them — and shares the honest engineering behind it, including the failures and the pivots, the way Karpathy shares machine learning.
- What does Jon Roosevelt build?
- Self-verifying multi-agent systems: fleets of AI agents that fan work out, review each other's results adversarially, and surface only the decisions a human still has to make. He runs them in a real, regulated urgent-care business — not a sandbox.
- What does an AI agent system inside a clinic actually do?
- It does the work around the care: catching billing errors before a claim goes out, handling follow-up and after-hours load, sourcing and coordinating, and reviewing its own code. A human signs anything that touches a patient or a staff member.
- What is the one engineering lesson Jon repeats most?
- "Green on mocks is never done." A unit that passes every test against fakes has only proven it agrees with your own assumptions. It isn't finished until it has run against the real thing — the rule he calls the reality gate.
- How do you contact or work with Jon Roosevelt?
- Email him directly — no form, no funnel. He's most interested in people building agent systems inside a real business, and clinic owners curious what this could look like for them.
If this is your kind of problem, email me.
Building agent systems inside a real business — or running a clinic and wondering what this could look like for you? That's exactly the conversation I want to have. No form, no funnel.