This page is a running record of how I spend my time outside of clinical work. Much of what I work on doesn’t turn into a product, post, or announcement. It looks like conversations, reading, sketching, experimenting and writing things that get cut, and ideas that are deliberately abandoned.
I keep this list for two reasons:
- To make that work visible
- To be transparent about the failures
BillBuddy (Hospital Billing Platform)
Nov 2025 → Present · Ongoing
Time investment: 20-30 h/week, sustained, multi-month focus
What it is
- A billing tool for hospital-based physicians that converts handwritten billing codes and printed patient census sheets into submission-ready billing entries.
- It takes a 20-30 minute manual and pain-staking process and automates it into a <5 min solution.
Why I'm doing it
- Billing workflows are archaic with no innovation in decades. It's not complicated, nor interesting, just repetitive friction that steals time and attention.
- Doctors already write the information down. Then they re-enter it later, carefully, when they’re tired and annoyed.
- That gap is a waste. I can close it, and people will pay for it.
What it took
- Writing and rewriting the end-to-end workflow so a developer can build without guessing: input, intermediate states, output.
- Writing and maintaining decision artifacts to prevent drift: TDDs for technical choices, ADRs for architectural decisions, and EXP records to capture what was tried, learned, and killed.
- Producing tens of pages of written documentation so requirements live outside my head and can be inspected critically.
- Enumerating edge cases before code exists: unreadable handwriting, missing identifiers, multiple codes per patient, admissions spanning days, late edits, cross-coverage notes.
- Defining v1 boundaries in writing so the product doesn’t sprawl: what is supported, what is rejected, what is deferred.
- Specifying architecture tightly enough to enforce accountability: capture, extraction, validation, review, export, and auditability.
- Producing documentation that locks decisions and creates accountability so the build doesn’t drift.
- Scoping phase two work separately: billing backend requirements, frontend needs, and who should own each piece.
- Developer search for phase two, where personality fit and judgment matter more than raw technical skill.
Status
- Specs are written. Scope is locked in.
- The open questions are now implementation details and edge-case handling.
- Need to learn front-end development myself and build that part myself.
Constraint
Using the app must be faster than doing it manually, even on a bad day.
What matters
Adoption without training. Accuracy that earns trust. Stickiness through time saved. Growth with low churn.
On-call escalation workflows in hospital settings
Mid 2025 → present · exploratory
Explored whether on-call noise and mis-escalation could be reduced with a structured intake layer for nurses before physician escalation. This involved mapping recurring failure patterns, clarifying what information is missing versus unused, and stress-testing whether the problem is tooling, training, or incentives.
Work included documenting escalation patterns, pressure-testing summaries that would be physician-usable, and identifying where responsibility actually breaks down.
Current state
Conceptually clear. Operational complexity remains high. No build decision.
What it clarified
Most escalation problems are not information deficits. They’re ownership and incentive failures.
Vital Shift direction reset
Late 2025 → present
Narrowed the newsletter away from broad longevity education toward helping readers interpret signals and make decisions. This involved writing pieces that never shipped, cutting sections that felt informative but hollow, and tightening the scope around what earns attention.
Current state
Direction locked. Fewer topics, more depth.
What it clarified
Subtraction creates more signal than expansion.
Longevity clinic model, revisited
Late 2025 · paused
Re-examined a clinic concept I had previously rejected, this time separating business quality from strategic optionality. The work was less about spreadsheets and more about understanding patient flow, physician time leverage, delegation limits, and whether the clinic functions as a profit center or a launchpad.
Current state
No build decision.
What it clarified
Some things are poor standalone businesses but useful strategic footholds.