01-projects/health-and-longevity

marchione systems personal os

2026-05-11·reference·source: Max Marchione, "Systems" page on personal site·by Max Marchione

Systems (Personal Operating System) - Max Marchione

Why this is in the vault

Marchione publishes his personal operating system as a live document - the instrumentation, daily targets, supplementation stack, and behavioral rules that he runs his life on. He has been wearing a CGM since age 16 and tracking sleep with an EEG headband for 12 months. That cadence + duration of self-instrumentation is rare in the founder ecosystem and makes his stack high-signal as a reference point for what a maximally-instrumented personal-OS actually looks like in practice.

The founder's health-and-longevity project is, at its core, the question "what should Ben's personal operating system look like?" The project README sketches the targeting-system framework. Marchione's Systems page is one concrete instantiation of that framework - useful both as a what-good-looks-like reference and as a stress-test for which parts are genuinely high-signal vs which are quantified-self LARPing.

The core content

The Systems page is a continuously-updated personal-OS document, not a single essay. The categories and contents (as of 2026-05-11):

Instrumentation stack

Sleep stack (titrated against the EEG data)

Magnesium glycinate, glycine, theanine, inositol, GABA. Targeting 7 hours/night (he claims 7 hours is less sleepy than 8.5 for him, which is one of his listed contrarian beliefs). The stack is conventional in the biohacker community - the differentiator is that he's titrated it against an objective sleep-quality sensor over 12 months, not just stacked it by reputation.

Daily metrics

7k steps, 4x weekly exercise, 10-40 supplements (range, depends on phase / experiment), environmental controls (light, temperature, air quality - exact protocols not specified on the page).

Behavioral / belief layer

His Contrarian Beliefs page surfaces these load-bearing positions:

The last one is borderline-crank but the cluster as a whole reads as "what someone discovers when they instrument themselves obsessively for years and notice patterns no one writes about." Worth filing as primary-source observations to pressure-test, not consensus claims.

Reading / consumption stack (from his "Faster" essay)

Speed-listens podcasts at 2-3x via Owltail (which goes up to 32x). Batches content by source (e.g. 20 podcasts from one creator alongside their book). Reads multiple books in parallel for cross-domain transfer. Adjusts speed based on HRV / recovery state - i.e. the sensor stack feeds back into the information-consumption protocol.

Mapping against Ray Data Co

Targeting-systems thesis convergence

Marchione's personal-OS is the same shape as RDCO's targeting-system, applied to one human:

Layer RDCO bets Marchione's personal-OS
Targeting Niche + bottleneck Health-span + cognitive-performance + energy
Sensors Notion + vault + watch loops CGM (continuous), EEG headband (sleep), HRV / cognitive-state wearable, scale
Actions Skills, content, decisions Supplementation stack, sleep timing, exercise, breathing protocol, content consumption
Tools MCP servers, scripts Specific apps, Owltail, supplement brands, CGM hardware
Feedback loop /self-review, /improve Titrating sleep stack against EEG, glycemic response against CGM, listening speed against HRV state

The structural identity here is the proof point. The targeting-system frame generalizes - Marchione independently arrived at the same architecture without using RDCO's vocabulary. That's evidence the frame is real, not RDCO-specific.

Where his stack diverges from the founder's project plan

Personal-OS analog to RDCO operating system

The most directly portable idea: Marchione's stack updates continuously and is published as a live document, not a static protocol. This mirrors how RDCO's skills + SOPs + memory files evolve continuously rather than being frozen at v1. The founder's health-and-longevity README should be treated the same way - a live document, updated quarterly with what the sensors are saying and which actions have been added/dropped.

The other portable idea: sensor → action → re-measure with the same sensor → re-titrate action. Marchione's sleep-stack titration against EEG data is the canonical example. The project's Whoop-based weekly health-pulse skill should do the same job: measure → action → re-measure → adjust.

Specific actions for the project

  1. Treat Marchione's contrarian-beliefs list as a candidate-experiments pool. Each is a cheap individual test. Files: surface 2-3 candidates per quarterly review.
  2. Steal the EEG-titrate-sleep-stack pattern: once Whoop is wired, use HRV + sleep stages as the feedback layer for any supplement intervention. Don't add supplements without a sensor that can detect their effect.
  3. Defer the CGM decision until bloodwork comes in. Marchione's default-CGM stance is the strongest pro-CGM argument we've seen; if HbA1c, fasting insulin, or triglycerides come back stuck, the case opens up.
  4. Don't copy the 10-40 supplement range. Stay supplement-minimal as project README specifies.

Related