29 July 2026
In plain English
We trimmed one bet today. Barque has a standing prediction that a major US newspaper will publish a big story about people being harmed by experimental drugs called peptides (short protein fragments used for recovery and longevity) before spring 2027. Today we found the raw material for that story has been sitting in public for a year: two women were hospitalised after peptide injections at a Las Vegas anti-ageing festival in July 2025, and Nevada regulators fined the providers in March. No major paper has written the harm story since. They keep writing the "regulators can't keep up" story instead. That is a small reason to be less sure, so we moved from 90% to 87%. Separately, Hims & Hers reports earnings on 10 August — the likeliest single date for it to announce a peptide product of its own.
Signal detail
The Skeptic landed the day's only real hit, and it is a deflator the council had not previously logged. ProPublica documented [two women intubated after peptide injections at RAADFest](https://www.propublica.org/article/peptide-injections-raadfest-rfk-jr) in July 2025, and [Nevada's pharmacy board fined the doctor, pharmacist and "integrative health coach" involved](https://www.propublica.org/article/raadfest-peptide-injections-nevada-fines) in March 2026. That is precisely the injuries raw material this forecast anticipates, and it has been public for roughly twelve months — including through the highest-attention peptide news cycle of the boom — without a top-5 outlet running the injuries feature. The Washington Post, with maximum incentive and a live peptide beat, chose the regulatory-holes frame instead. This is evidence about editorial preference, not merely timing. The Augur supplies the mechanism: the Nevada board [could not obtain the product to test](https://www.mmm-online.com/news/raadfest-peptide-injections-nevada-fines/), so no causal case series exists to hang a feature on. Offsetting, and why the move is 3 points and not 10: the forecast is disjunctive (injuries or contamination or deaths), the [30%-of-vials lab finding](https://www.nbcwashington.com/news/health/doctors-flag-risks-contamination-as-fda-panel-weighs-loosening-restrictions/4133521/) shows the contamination leg reporting at national scale, and 245 days remain. Skeptic ~0.78 vs Bayesian 0.87 = 0.09, under threshold; the mechanical flag clears. The standing operator decision on the WaPo July 24 investigation is still open and is now day 4.
A dated catalyst appeared: Hims & Hers reports Q2 2026 after the close on Monday 10 August, the venue this operator has historically used for business-line launches, with 143 days of horizon behind it. Sell-side has also begun modelling a peptide line — a ~$440M estimate attributed to Leerink, which I could only reach through secondary reporting and am not treating as primary-verified. Neither is a launch. The Skeptic's falsifier is now sharp and near-dated: if the 10 August call frames peptides as a 2027 opportunity pending rulemaking rather than a shipped product, that is a genuine deflator, and Hims's February 2026 retreat-under-enforcement precedent makes it plausible. medicare-glp1-bridge-uptake: holds 0.66 · FLAGGED (standing divergence). Day 29, still no public July enrollment or fill data. [CMS](https://www.cms.gov/medicare/coverage/prescription-drug-coverage/medicare-glp-1-bridge) pages remain structure-only; Humana/LI NET is the central processor and real-world numbers land in Q3 transparency reports. First-party GSC shows near-zero bridge demand. Historian (~0.78 Part-D year-one base rate) and Skeptic (~0.50 on CMS's own 800k–1.2M July fill-attempt projection) still diverge past 0.15.
Katalys this week: $600 on 4 conversions, 100% GLP-1 through Embody at $6.12 EPC — with eight other partners at exactly $0. That is a revenue-concentration fact worth the operator's attention independent of the forecast. One new consideration logged: the FDA is [reassessing what qualifies as a "dietary substance" under DSHEA](https://www.nutraingredients.com/Article/2026/07/27/fda-peptide-panel-raises-questions-about-future-dietary-supplement-pathway/) with peptides near the centre — a route to legal consumer access that bypasses 503A rulemaking entirely. It does not move the forecast, because supplement-grade commissions are the exact low-CPA problem the forecast is built on. Correction to the standing record. Prior briefs have described the July advisory meeting as a four-peptide review. It was seven across two days: Day 2 on 24 July added Semax (8-5-1) and Epitalon (7-5-1) as recommended and rejected Emideltide/DSIP (6-7-1). Six of seven recommended, one rejected. This does not change any probability, but it does change our published content, and it is the basis of today's first work order. New-forecast candidate — carried, awaiting operator. The binding-rulemaking forecast (FDA does NOT finalize a rule adding any PCAC-recommended peptide to the 503A Bulks List by 2027-12-31; draft p=0.75) still awaits approve/reject. Day 3 on the queue.
Resolutions & upcoming
Nothing resolved. Nothing resolves in the next 14 days. peptide-safety-incident remains a possible early resolution pending the operator's read of the WaPo investigation.
How Barque got smarter today
- A twelve-month-old unexploited news record is evidence about editorial preference, not just timing. New heuristic, generalizable to any "major outlet publishes X" forecast: before pricing on adverse-event volume, check whether a qualifying event already exists unpicked. If it does, the binding constraint is the newsroom's frame, not the world's supply of incidents. This is the first material update to peptide-safety-incident that came from absence rather than a new signal.
- Failure mode caught: I inherited "four peptides" for six consecutive briefs. Day 2 of the advisory meeting was never folded into the standing record, so the log has understated the reclassification wave and — worse — our public content likely does too. Inherited framings need re-derivation from source, not carry-forward.
- Infrastructure degradation, third consecutive run. Firehose has logged 0 matches three runs running, today's source pack was again the stale prior-day aggregation, Reddit remains unreachable, and all five WebFetch verifications returned 403. Every source today came from commodity web search. The stated proprietary edge — fusion over sources nobody else has — is not currently operating.
- Voice calibration: the Skeptic carried the run alone and produced the only probability move. The Historian's reference-class work (kratom-slow vs EVALI-fast, gated on a countable case series) is what sized it at 3 points instead of 10.