What the AI found
“Your 2pm cognitive slump correlates with carbohydrate volume at lunch (r=0.68), not with sleep duration or morning caffeine timing.”
Before
Three midday coffees, brain fog blamed on sleep
After
One 15-minute weekly protocol test she controls
The same system, three states — real screens, not a screenshot
- Oct 12: terrible focus after 2pm, bought oat latte
- Oct 14: brain fog again, slept 7.5 hrs though?
- Oct 18: try no caffeine past noon (failed by 2:30pm)
- Oct 22: foggy, maybe need magnesium?
Prompt
Here are 28 days of my cognitive ledger (sleep, caffeine timing, lunch macronutrients, 2-4pm subjective focus score 1-5). Cross-reference sleep duration and caffeine timing against the 2pm slump scores. What pattern emerges?
Here are 28 days of my cognitive ledger (sleep, caffeine timing, lunch macronutrients, 2-4pm subjective focus score 1-5). Cross-reference sleep duration and caffeine timing against the 2pm slump scores. What pattern emerges?
AI
Sleep duration (p=0.41) and morning caffeine timing (p=0.38) show no statistically meaningful correlation with your 2-4pm focus scores. However, carbohydrate volume at lunch strongly predicts the slump severity (r=0.68). On all 11 days your lunch exceeded ~45g carbs, focus dropped to 1 or 2 by 14:15, regardless of sleep.| Week 3 Protocol | Sub-30g carb lunches (Mon-Thu) |
| Slump severity (1-5 scale) | 1.2 average |
| Afternoon caffeine intake | 0 mg |
Solving an Afternoon Cognitive Slump Without Touching Caffeine
How a software architect isolated the real driver of her 2pm cognitive crash using the 3-Layer Method.
A 36-year-old software architect, Northern Europe
Tools used
The real tools used here — swap any for your own equivalent. Each links to how we’d set it up.
Starting state
Before anything was set up
Her desk bore the evidence of an unsolved pattern: two half-empty mugs, an espresso cup, and fragmented Apple Notes documenting afternoon brain fog. For eighteen months, she assumed her sluggish two o'clock code reviews stemmed from poor sleep quality or caffeine tolerance. She tried shifting her morning flat white by ninety minutes, added cold showers, and downloaded another tracking app that merely confirmed she felt exhausted at 14:30. Nothing resolved the slump.
- Oct 12: terrible focus after 2pm, bought oat latte
- Oct 14: brain fog again, slept 7.5 hrs though?
- Oct 18: try no caffeine past noon (failed by 2:30pm)
- Oct 22: foggy, maybe need magnesium?
Working state
Core Course, doing its job
In the second week of the Core Course, she built her Research dossier on postprandial somnolence and set up an eight-column Google Sheets ledger. She tracked four variables: sleep hours, morning caffeine timing, lunch composition, and 2pm mental acuity on a 1-to-5 scale. At day twenty-eight, she exported the raw entries into Claude, prompting it to test her sleep hypothesis against the actual numbers.
Prompt
Here are 28 days of my cognitive ledger (sleep, caffeine timing, lunch macronutrients, 2-4pm subjective focus score 1-5). Cross-reference sleep duration and caffeine timing against the 2pm slump scores. What pattern emerges?
Here are 28 days of my cognitive ledger (sleep, caffeine timing, lunch macronutrients, 2-4pm subjective focus score 1-5). Cross-reference sleep duration and caffeine timing against the 2pm slump scores. What pattern emerges?
AI
Sleep duration (p=0.41) and morning caffeine timing (p=0.38) show no statistically meaningful correlation with your 2-4pm focus scores. However, carbohydrate volume at lunch strongly predicts the slump severity (r=0.68). On all 11 days your lunch exceeded ~45g carbs, focus dropped to 1 or 2 by 14:15, regardless of sleep.Use case implemented
The finished system, running on its own
By week four, the routine ran quietly in the background. Her lunch remained under thirty grams of carbohydrates on working days, shifting starches to dinner. Midday fog ratings dropped without touching sleep schedules or morning espresso. When fatigue surfaced on a Thursday, she checked her spreadsheet rather than ordering cold brew. The ledger was no longer an aspirational audit; it was a calibrated operating manual for her working afternoons.
| Week 3 Protocol | Sub-30g carb lunches (Mon-Thu) |
| Slump severity (1-5 scale) | 1.2 average |
| Afternoon caffeine intake | 0 mg |
| Focus blocks completed | 4 of 4 |
What an outside observer would notice
11 down to 2
Weekly afternoon coffees
4 of 4 completed
Unbroken focus blocks (2-4pm)
The stack — build it yourself
Handles messy multi-week markdown exports and spots subtle multi-variable correlations calmly.
Fast zero-latency mobile data entry without proprietary app silos or subscription bloat.
These are the tools used in this story. Any can be swapped for an equivalent you already trust.
Go deeper
Do this yourself
Explore the Core Course
This story runs on Core Course. The tools and prompts above are the real build — swap any tool for your own equivalent and follow the same steps.