THE NEXT PATIENTAI-driven health care intel

Technology architecture · tools serve the workflow · static snapshot

The technology stack

The stack is organized by function: acquire evidence, do the work, preserve state, store assets, deliver the product, and observe what actually happened. Any individual AI model can be replaced without losing the business.

Architecture principleState outside the modelsKeep state, provenance, and approvals outside the models. Use APIs for repeatability, the browser for setup and visual verification, and Steve for the consequential gate.
Models think. Systems remember. Live surfaces prove.

The redesign. GitHub is the control record, not the binary bucket. R2 is the working binary store. Beehiiv is the publication engine. The browser is an operating surface—not a durable integration.

05 / The board · architecture as a circuit

The working surface is not the memory layer

Inputs arrive on edge connectors, execution is the main IC, the durable record is memory with the manifest as its bridge, delivery leaves on the bottom edge, and reader evidence comes back on the blue return trace. The held Grok lane is an unpopulated socket. Select a component to read its contract.

Board pans sideways on small screens; the components below are the controls.

  • Signal trace · production flow
  • Return trace · observation and learning
  • Unpopulated · held lane
  • Memory · durable store
  • Main IC · execution

Preferred order: connector or API for repeated work; browser for configuration and visual proof; manual action only at the real gate

The blue return line matters as much as the green production line. Without received-form evidence, the system knows what it sent—not what the reader saw. Evidence returns to the work layer and the durable record deliberately—not as inferred rules.

Recorded in
tech-stack.html and schematics.js stack map (visual-operating-system v1.3.2) · spine/PROTOCOL.md §1
Layer
Architecture and recorded integration state · not a fresh account audit. Intel Filter and Reading Room migration belong to Build, outside TNP.
What this stack cannot do
Certify that a live integration is healthy. A tool is not an owner; ownership sits with the layer's contract, state, and proof requirement.

Management view

What each layer owns

A tool is not an owner. Ownership sits with the layer's contract, state, and proof requirement.

Layer
Primary system
Contract
Failure test
Acquisition
Grok, Gmail, web, APIs
Evidence package with source, date, relevance, and uncertainty.
Can another worker verify the claim without the chat?
Orchestration
ChatGPT / Codex
Named job, named inputs, inspectable output, and explicit gate.
Could the work resume tomorrow from files alone?
Durable state
GitHub master + Obsidian view
Current routing, versions, protocol, and provenance.
Does NOW.md point to one current artifact?
Binary storage
Cloudflare R2
Stable object and manifest pointer. Drive remains a static archive.
Can the named image or proof be retrieved directly?
Delivery
Beehiiv + Cloudflare
Email, site, ops, and reading surfaces as the reader sees them.
Did the delivered product arrive, render, and link correctly?
Observation
Gmail, analytics, replies, logs
Real-world evidence tied back to the named run.
Are corrections based on observation rather than model confidence?

Sequencing

Build now · do not optimize yet

The dotted migration lane—Intel Filter, Reading Room, recurring health routines—is acknowledged but not designed here; it belongs to Build, outside TNP.

Build now
  • Prove Beehiiv/Stripe checkout and the received-email form.
  • Keep GitHub and R2 roles sharply separated.
  • Preserve one named run package per issue.
  • Return live evidence to the next run.
Do not optimize yet
  • Multi-agent autonomy before the line repeats.
  • Automated editorial promotion.
  • A new platform to replace the current spine.
  • The legacy routines until the issue factory is stable.