Gut
IndividualProtocol layerSpreadsheet ledger

Finding physical ease through quiet weekly self experiments

How an architect traded restrictive food anxiety for disciplined weekly observation.

3 min readWellness & AI editorial

A 38-year-old architectural designer in Northern Europe spent his late afternoons nursing familiar, sour abdominal heaviness. Desk work meant quick sourdough sandwiches eaten over blueprints, followed by dull, distracting distension that blurred his concentration until dinner. Over two years, he had eliminated dairy, re-introduced it, and accumulated five half-finished jars of broad-spectrum digestive enzymes. Nothing settled the predictable 4 p.m. discomfort. His meals felt cautious, joyless, and entirely decoupled from any intelligible pattern.

He stopped categorising whole food groups as hostile and began noticing the quiet mechanics of preparation and timing. Instead of feeling trapped by an erratic body, he treated his afternoons as an ongoing dialogue, discovering that gentle changes in food temperature and pacing carried more weight than radical exclusions.

He set up a modest digital ledger to capture meals and sensations, then asked a reasoning model to audit the raw notes at the end of each week. Rather than following automated recommendations, he used the model merely to surface candidate variables he had overlooked. Each subsequent week became an intentional, single-variable trial. Throughout the process, the machine served strictly as an indexing mirror, leaving the pacing, hypotheses, and dining choices entirely in his own hands.

Colleagues stopped seeing him stretch backward over his drafting chair mid-afternoon to relieve abdominal pressure, and he began accepting invitations to shared studio lunches again without hesitation.

Adapt the shape to your own stack

Vendor-neutral steps. Use whichever AI tools you already trust — the shape of the work matters more than the brand.

  1. 1

    Log without immediate interpretation

    Capture simple meal components, approximate timing, and post-meal sensations in a plain spreadsheet ledger for seven days before drawing any conclusions.

  2. 2

    Audit for recurring correlations

    Supply the raw ledger text to a reasoning chat tool and ask it to highlight non-obvious pairings between specific meal textures and mid-afternoon discomfort.

  3. 3

    Design a single-variable trial

    Select one candidate variable to adjust for a full week, keeping preparation styles and dining intervals otherwise consistent.

  4. 4

    Evaluate observable physical ease

    Review the week’s qualitative notes to determine whether the adjustment created repeatable relief, iterating gradually rather than overhauling your entire diet.

Three things to read next.

See all →

Suggested for you

Based on what you've been reading — always learning.

See all →