Two years of shipping code at 60% capacity.
Not burned out. Not sick. Just... slower than I used to be. Longer to get into flow. Harder to stay there. The 3PM wall that I'd accepted as just how afternoons work. Waking up tired after eight hours of sleep and reaching for coffee before my IDE had finished loading.
I assumed it was the job. The screen time. Getting older.
Turns out it was three fixable bugs I'd never thought to check.
Step 1: Run the Profiler
bash
What I'd been doing for two years
$ drink-more-coffee
$ restart --hope-for-the-best
$ blame: workload, age, "just how it is"
What I finally did
$ request-bloodwork --full-panel --micronutrients
Results:
vitamin-d: 19 ng/mL // CRITICAL (target: 40-60)
rbc-magnesium: low // HIGH (serum was "normal" — misleading)
omega3-index: 3.2% // HIGH (target: 8%+)
Three bugs. All silent. All compounding each other. None of them showing up on a standard annual checkup.
Step 2: Write the Patches
No exotic dependencies. No 15-package install. Three targeted fixes based on actual profiler output.
javascript
// patch-001: vitamin-d3-k2.js
const vitaminD = {
form: "D3", // not D2 — 87% less effective at raising levels
dose: "4000 IU", // not the legacy 400 IU default
cofactor: "K2-MK7", // routes calcium — don't skip this dependency
timing: "08:00",
requires: "dietary fat" // fat-soluble — null context = silent failure
}
// patch-002: magnesium-glycinate.js
const magnesium = {
form: "glycinate", // not oxide (known broken build, absorption: ~4%)
dose: "400mg",
timing: "22:30", // 30min pre-sleep — where it earns its value
note: "separate from calcium — race condition"
}
// patch-003: omega3-high-epa.js
const omega3 = {
EPA: "1000mg+",
timing: "largest meal",
preCheck: () => {
if (smell === "rancid") throw new Error("oxidized — discard immediately")
}
}
For sourcing I used sunday.co.ua — official Ukrainian store for Sunday Natural, a German brand that ships correct ingredient versions by default. D3+K2 combined, glycinate form for magnesium, high-EPA omega-3. Direct from Germany.
No configuration required. Correct defaults out of the box.
Step 3: Deploy and Monitor
Day 01-07: No significant output change. Committing anyway.
Note: falling asleep ~10min faster. Possible placebo. Logging.
Day 08-14: Sleep quality PR merged successfully.
Waking up less during night. More rested at same duration.
Magnesium glycinate effect — documented in research, real in prod.
Day 15-21: Afternoon performance regression reduced.
3PM wall: hard stop → manageable dip.
Context switching: cheaper. Flow state: more accessible.
Focus degradation under pressure: slower.
Day 22-30: Energy curve stable. Mood baseline more consistent.
No acute peak — no "kick". Just system running closer to spec.
Step 4: Run Tests
Retested at 8 weeks:
vitamin-d: 41 ng/mL // ✅ inside optimal range (was 19)
rbc-magnesium: normal // ✅ resolved
omega3-index: 6.1% // 🔄 improving (target: 8%+, was 3.2%)
Numbers moved. Symptoms moved with them.
The Retrospective
What I'd do differently:
Run the profiler first. I wasted months on random supplements before having actual data. Bloodwork cost less than two months of guessing and told me exactly what to fix.
Don't install the 27-in-1 multivitamin. That's npm install everything with no package.json, wrong versions, and competing dependencies. It looks productive. It resolves nothing.
Minimum 30 days before evaluating. Micronutrients aren't caffeine. No acute effect. No obvious moment where something kicks in. Anyone closing the PR after five days is measuring noise.
The Bigger Lesson
I had normalized running at 60% for so long that I'd forgotten what full capacity felt like.
The bugs weren't dramatic. No crash. No fatal error. Just a silent performance regression I'd retrofitted explanations for — workload, screens, age, "just how it is now."
It wasn't how it had to be.
It was three unread error logs and a blood panel I should have run two years earlier.
Top comments (0)