Chapter 36 — Key Takeaways

The governing idea

  • This is the chapter most likely to be wrong, and it says so first. A chapter about the future cannot rate claims against evidence, because the evidence does not exist — true by construction, not a limitation of anyone's research.
  • So it does two other things. It reports status — what is in development and at what stage, as of this writing, in 2026 — which is a present-tense fact you can check and the author can be caught getting wrong. And it teaches the reading skill, which is the part that lasts.
  • "Compound X is in phase 3" and "compound X will be approved" are different kinds of sentence. Only the first can be an error. The second can only be a failed guess.
  • A future-facing chapter that hypes retroactively discredits every chapter before it, by revealing that the evidence discipline was a style rather than a commitment.

The six questions (§36.2) — the part with no shelf life

Run these on any press release, news article, conference slide, clinic email, or excited summary. They take about ninety seconds once memorized.

  1. What phase, actually? Preclinical / 1 / 2 / 3 — and how far from an answer.
  2. What endpoint? A surrogate, or something a person would notice.
  3. What population? Who was enrolled, versus who will receive it.
  4. Against what comparator? Placebo, standard of care, or nothing at all.
  5. Who is claiming, and when? Press release, abstract, peer-reviewed paper, earnings call.
  6. What is the base rate? Most candidates fail. Start there and make the claim move you.

If you can only ask one, ask 6. If you can ask two, add 1. They are ordered by how much time each takes, not by how much each matters.

What the base rate does — and what cannot move it

  • Most compounds entering human trials never reach approval, and this holds even for compounds that reached phase 3. The chapter deliberately quotes no single percentage; direction and magnitude matter more than a figure that varies by area, by counting method, and by decade.
  • A mechanism story does not move the base rate. This is rating rule 3 — never upgrade with mechanism — applied to the future. Every failed drug had a mechanism story; that is why it entered trials at all.
  • What does move you: stage; a hard outcome already demonstrated for the same molecule in a related indication; precedent in the same mechanistic class; a large trial with an active comparator that has read out. What does not: elegance, investigator enthusiasm, preclinical effect size, volume of coverage, or your sense that it ought to work.
  • "It's already in clinical trials" should lower your confidence relative to the excitement, not raise it — and is often true of a different indication.

Where things actually stand, as of this writing, in 2026

  • Oral delivery works, at about 1 percent. Oral semaglutide is approved and in use, co-formulated with the absorption enhancer SNAC. Roughly ninety-nine percent of drug substance never reaches circulation, so an oral tablet must contain far more material than an injection — which worsens the synthesis bottleneck even as it relieves the fill-finish one.
  • The winning oral drugs may not be peptides. Orforglipron, a non-peptide small-molecule GLP-1 receptor agonist, needs no absorption enhancer, no cold chain, and no solid-phase synthesis. Not a defeat for peptide science — captopril's story again (Chapter 35), and arguably its highest achievement.
  • Longer duration is a trade-off, not a ladder. You lose the ability to stop, titration becomes coarse or impossible, and sustained receptor occupancy is exactly what drives desensitization and downregulation — the two termination mechanisms engineering cannot defeat.
  • The multi-agonist race is optimizing a surrogate. A drug producing more weight loss than tirzepatide has not thereby been shown to prevent more heart attacks. Outcome trials are longer, larger, and slower, which is why the weight number always arrives first.
  • "Disappointing" and "ineffective" are different words. A topline figure can miss investor guidance and still describe weight loss that would have been extraordinary a few years earlier. The number was not wrong; the frame was built for somebody else.
  • The muscle question is genuinely open, and the endpoint that would settle it is function — not a body-composition scan. Several leading candidates are antibodies rather than peptides.
  • Acting on the brain and crossing the blood-brain barrier are different claims. Circumventricular organs such as the area postrema have fenestrated capillaries, and vagal afferents carry the signal without the molecule.
  • Patent expiry changes price and access — and no evidence about whether the drugs work. Keep two ledgers; the error runs in both directions.

The frontier discipline

  • 🔬 is not a promise. It means a serious question is being asked seriously. It carries no timeline and no probability of success. Most 🔬 becomes ❌.
  • A 🔬 entry must name the readout that would move it. If you cannot name one, it is not a frontier entry — it is an enthusiasm.
  • A 🔬 entry must carry a date. An undated entry cannot age, and an entry that cannot age can never be found stale.
  • A 🔬 entry may never be upgraded by news. Not a press release, abstract, funding round, partnership, mechanism paper, or animal result. Only the named readout moves it. Write the "do not count" line now, while you have no attachment to the outcome.
  • Ratings drift upward and almost never downward, because the stream of news about anything you follow is filtered in one direction. Absence of bad news is close to uninformative, and it feels exactly like good news.

The last idea, which is the point of the book

  • A prediction says what will happen. A forecast says what will happen and what would prove it wrong. §36.10's three lists are falsifiable and date-stamped to 2026 so a reader can grade them.
  • Every claim this book has taught you to distrust lacks that property. "Peptides are the future of medicine." "The next generation will be far better." Not one of those sentences can be wrong. They are not claims; they are moods with the grammar of claims.
  • Ask the next futurist you encounter — including this one — what would make them wrong. If they answer specifically, listen. If they cannot, you have learned that whatever they are doing, it is not forecasting.

Next

Chapter 37 — The Peptide Evidence Table. Chapter 36 asked you to add 🔬 entries and to run every "coming soon" story in your dossier through the six questions. Chapter 37 does the parallel job for the settled material: it assembles every rating issued across the whole book into a single reference — each claim, its rating, and the evidence behind it, in one place to look things up rather than hunting through chapters.

Read it with this chapter's warning attached. A table is the most authoritative-looking object in a book, and the ratings in it attach to claims, not to molecules — one molecule can hold ✅ for one claim and 🔬 for another, as retatrutide does here and as PRRT does in §36.7. Chapter 37's usefulness depends entirely on your not collapsing those rows into a verdict on a compound.

Bring the frontier entries you dated today and the six questions. The table tells you where things stood in 2026; the six questions are what you will use on everything published after it.