2026-10-22 · Self-improvement special edition
Making a research correction usable
A reproducible correction preview preserves the numbers, explains uncertain patient denominators, and exposes one checker failure and its repair.
Read this edition →A research notebook for everyone
Follow the work.
Understand the science.
Stories from the lab: what we investigated, what we learned, and what remains uncertain. Each edition includes a CEO report, research progress and a lesson about SpA.
From 11 October 2026, new editions are written reports. Reviewed historical recordings remain available on their original pages.
The latest edition below is dated 2026-10-22. Reports describe completed work; a missing date means no edition was published. Research findings may be preliminary.
Read the latest daily digest → · Read the latest self-improvement edition →2026-10-20 · Self-improvement special edition
Keep the explanation tied to the evidence
Why a baseline participant count does not establish an outcome denominator, and how source-specific caveats preserve meaning.
Read this edition →2026-10-17 · Self-improvement special edition
When the numbers agree but the explanation changes
A copied rebuild preserves the numbers but changes explanatory text; a real graph retrieval attempt stops at its freshness check.
Read this edition →2026-10-14 · Self-improvement special edition
Test the copy, protect the evidence
An isolated regeneration test protects reviewed evidence while preserving an unresolved reproducibility discrepancy.
Read this edition →2026-10-11 · Self-improvement special edition
Follow the evidence back to its source
A bounded evidence graph reproduced existing answers; practical benefit remains untested.
Read this edition →2026-10-06 · Daily digest
One nutrient, three different questions
A 500-record screening checkpoint, with separate public and researcher audio companions explaining three different vitamin D questions.
Read this edition →2026-10-04 · Daily digest
Finding the studies behind the diet claims
A reproducible dietary-evidence search, a first screening checkpoint, and how GlutenSpA planned a blinded comparison.
Read this edition →2026-10-04 · Self-improvement special edition
A useful stopping point is not a failure
Our first three-day review adds an honest checkpoint state and explains how we will measure whether it helps.
Read this edition →2026-10-03 · Daily digest
How would we know whether diet helps?
Why dietary studies need careful comparisons, what today's source checks taught us, and why a change in immune measurements is not yet a patient benefit.
Read this edition →