Chapter 40 — Key Takeaways
The object
You built a personal reference, not a summary. A summary compresses someone else's conclusions and goes stale on a schedule you do not control. A dossier records what you established, about compounds you chose, in a form you can update — because you also recorded what would unmake each entry.
It is organized by question, not by compound. That is why it keeps working on molecules it does not contain, and why it works on things that are not molecules at all. The twelve fields are not a filing system; they are twelve places where a specific error is easy to make, arranged so that making it is visible.
Three fields are almost always the empty ones: 9, 10, and 12. Risks, status, and what would change my mind. They are the least satisfying to write and the most useful to have. Field 9 is what someone will actually ask you about, Field 10 is what dates your file fastest, and Field 12 is the only thing separating a dossier from a scrapbook of opinions.
The audit
The question is not "are my ratings right" — it is "did I apply one standard?" Correctness is measured against an external source and is arguable. Consistency is measured against your own other entries and is detectable from inside the document, which is what makes it a real audit.
Four checks, and they take about an hour.
- Same-evidence test — two entries with comparable evidence and different ratings. Justify the difference by design, population, or endpoint, or fix one of them.
- Population check — every Field 6 row names a population and an endpoint. An unpopulated rating rates a molecule, which rating rule 1 forbids, and it is the most common defect in reader dossiers.
- Date check — an undated rating is a claim about the eternal. Cheapest check, most often skipped, and the one everything in §40.4 and §40.5 depends on.
- Field 12 check — every entry falsifiable. An entry with an empty Field 12 is not a conclusion; it is a belief, and that is as true of a confident ❌ as of a hopeful ⚠️.
The test for a Field 12 entry: could a reasonable person disagree about whether the named readout had arrived? If yes, it is not specific enough.
Audit your confident material first. In the worked example, every defect sat on the entry the reader was surest about, and the only entry that passed all four checks was the one they cared least about. Difficulty is not where the errors are.
Your own error
Look at the column, not the rows. Scattered disagreement with the book's ratings is noise. Disagreement running consistently in one direction is a bias, it is measurable, and almost nobody has ever measured theirs.
Two directions, equal billing. The reader generous toward what they hoped would work, and the reader harsh toward anything that sounded like marketing. The second feels like rigor and is the same failure — rating rule 4, never downgrade with distaste, is the exact mirror of rule 3, never upgrade with mechanism.
The harsh reader gets no corrective signal. Nobody is ever embarrassed for having been too critical of a supplement, so those errors compound quietly — and they travel, because a ❌ looks identical whether it came from Field 5 or from taste.
Being right for the wrong reasons is invisible from the inside. It feels exactly like being right. The same-evidence test is what makes the difference visible.
Keeping it alive
Date every change and keep the old version. A dossier with a history tells you how you were wrong before and in which direction. A snapshot cannot, because it contains no trajectory.
The most valuable revision is one that records no change — news arrived, the rating held. A file where every revision is an upgrade has been tracking enthusiasm, not evidence.
Review on events, not on a schedule. A readout, an approval or withdrawal, a shortage, a safety signal, or a question you could not answer. Calendar review sends you looking for news at intervals, which manufactures upgrades, because there is always something.
Expect to remove entries. A working file, not an archive. A stale entry is not neutral — it will eventually tell you something that used to be true.
Nothing upgrades a rating except the readout named in Field 12. Not a press release, an abstract, a preprint, a mechanism paper, a funding round, a clinician's enthusiasm, or a friend's result. This is the rule you will be most tempted to break, because every item on that list genuinely is information — just not about the thing the rating tracks.
Drift
Ratings creep upward and almost never downward, and the reason is structural. Every piece of news reaching you about a compound you are watching is, by selection, news someone thought worth publicizing. Null results are published quietly, late, or not at all. Then your own attention applies the same filter again, because you search for what you care about.
This is Chapter 6's testimonial dynamic operating on your own attention — selection on the outcome, applied to publications instead of people.
Vigilance is not the defense; the written Field 12 is. Drift does not feel like drift, it feels like being appropriately responsive. So do not adjudicate whether the picture has changed — open the entry, read the line you wrote before you had a stake, and ask whether that happened.
What it is for
Three uses. Answering your own question, which usually decomposes rather than resolves. Having the Chapter 39 conversation, which requires converting a verdict into a question whose answer is in the room. And saying "I don't know" precisely.
Three different sentences, routinely collapsed into one shrug:
- "I don't know" — a fact about you. Go look.
- "Nobody knows" — a fact about the world. ❌, evidence absent.
- "It's been tested and it doesn't work" — a fact about the world, and a far stronger one. ❌, evidence present and negative.
A compound with a large, well-conducted null trial is better understood than one with no trials at all. Readers get this backwards because "nothing has disproved it" sounds favorable — but it is true of every claim nobody has ever examined.
A verdict that cannot be turned into an answerable question was probably a feeling.
The transferable skill
The peptides were the curriculum; the judgment is what you keep. Run the twelve fields at a diet, a supplement, a wearable, a consumer device, and the same four errors appear in the same four places: a name mistaken for a specification, an origin story doing evidentiary work, a mechanism standing in for an outcome, and a claimed use outrunning anything established.
A modality is not a product, exactly as a peptide is not a vial. Evidence attaches to specified exposures; something that does not specify its exposure has inherited nothing.
Some fields come out structurally empty rather than damningly empty — nobody approves eating patterns — and telling those apart is a real skill.
What should survive after you forget the content: for whom, measured how, compared to what, and how would we know if it were false.
The honest limits
A rating is an input to a decision, not the decision. Two people with identical dossiers can reasonably choose differently, weighting things the file does not contain. When that happens the dossier has still succeeded: it made the disagreement be about values rather than facts.
It cannot make the evidence exist. For most compounds readers arrive here about, the honest final position is not a verdict but a well-specified open question — and stating one precisely is a real result. "Does it work?" cannot be resolved even in principle. A map of a specific hole in human knowledge, drawn accurately enough that you would recognize the moment it got filled, can be.
And the handoff
This chapter closes the method arc. It does not close the book.
Chapters 41 through 44 are the field test: how a drug's brand name entered the language, the creator economy selling these compounds, enhancement outside the clinic, and what a society on metabolic drugs can honestly claim about itself.
They are harder for three specific reasons: no trial arms, no endpoints, no placebo. The twelve fields do not transfer cleanly and the honest thing is to say so — what transfers is the discipline underneath them.
Claims about molecules are made about compounds. Claims about society are made about you — more often, more confidently, and with almost nobody applying any standard at all. That is why Part VIII is the point of the book rather than an appendix to it.