Latest signal
last curated pass Sep 4, 2026 · 8 public entries · source-aware summaries from 6 legacy curated cards plus llm-wiki tweets
Recent curated signals
Measurement Johnson announces an Eight Sleep AI model that reads a “sleep fingerprint” and estimates biological age Johnson posted that he “just launched an AI model based on sleep data… and it accurately predicts your age,” built with Eight Sleep CEO Matteo Franceschetti and available on their platforms. The claims are unusually checkable because a preprint exists: BCG-FM (Kjaer et al., arXiv:2606.07692) pretrained on 2.04 million hours of bed-sensor ballistocardiography from 136,575 participants, and its reported figures match the post — biological age within 3.26 years (the lowest reported for any ambient, contactless modality), 92.5% Rank-1 identity retrieval from one night’s signal, and AUROCs of 0.852 for diabetes, 0.822 for heart failure (external cohort), 0.810 for hypertension, 0.792 for sleep apnea, 0.751 for snoring, and 0.734 for general heart conditions, with log-linear batch-size scaling (R²=0.982). The boundaries: this is a vendor collaboration (Eight Sleep’s membership “Biological Age” feature, developed with Johnson’s Immortals), the internal disease labels are self-reported, the external cohorts are small, the preprint is not peer-reviewed, and the post’s “diabetes better than Apple’s model” has no corresponding comparison in the preprint, whose Apple-Watch PpgAge benchmark is age accuracy, not disease detection. Johnson’s own post concedes it is “a research milestone, not a diagnostic device.” The dashboard records an attributed product-launch and research claim with verified preprint context—not a validated diagnostic, a reader health claim, or medical advice. 3 sourcesSleep biosignal foundation model / biological age Claim Johnson claims he is “indistinguishable from an 18-year-old” across fourteen measured systems Johnson posted that “in many ways, I am 18”—claiming an alien would struggle to distinguish him from his son—and listed fourteen systems where he says he is indistinguishable from an 18-year-old: sleep quality, erection function, fertility, resting heart rate, vascular health, cardiovascular health, blood pressure, metabolic health, blood glucose control, bone mineral density, muscle, fat, inflammation (“undetectable”), and, jokingly, shitposting. He concedes two measures remain age-typical—hearing and somatic mutations—and frames the result as “the first rep,” achievable by others in five years because “the most powerful things are free, and gated only by choice.” The post publishes no dataset, assay methods, reference cohort, or comparison standard for a single listed system; “indistinguishable from an 18-year-old” is Johnson’s self-characterization of his own test results against undisclosed norms, not a validated multi-system equivalence. The dashboard records an attributed N=1 outcome claim—not evidence of system-by-system youthful equivalence, a promise readers can replicate the result, or medical advice. 1 sourceMulti-system biological-age claim Algorithmic health A cafe’s decaf mistake becomes a measured investigation—anchored by an unattributed “70–130%” injury-risk figure After a cafe served him fully caffeinated coffee instead of decaf, Johnson reported wrecked sleep and cancelled his morning dunk training because “acute sleep disruption can increase injury risk by 70-130%.” Two days later he returned with a sound meter and published the investigation: ~75 dBA cafe background against his replicated 72 dBA order volume (a −3 dB signal-to-noise ratio in which “a decaf coffee” and “a cup of coffee” could sound nearly identical), a photographed order screen where adjacent “brewed” caff/decaf buttons invite a “capture error,” a cited ~1-in-8 cafe order error rate, and an estimated ~55% chance the barista recognized him. The forensic habit—quantifying a mundane mistake rather than dismissing it—is a textbook algorithmic-health move. But the 70–130% injury figure names no study, and the nearest literature does not match it: chronic insufficient sleep in adolescent athletes raises injury risk 1.7× (Milewski et al., 2014, J Pediatr Orthop; 112 athletes, <8 h/night, multivariate) and up to 2.25× when training load rises while sleep falls (von Rosen et al., 2017, Scand J Med Sci Sports; 496 athletes)—associations in young athletes over a season, not acute single-night disruption in a 49-year-old. Skipping one training session after a wrecked night is a defensible personal decision; the specific number is not. The dashboard records an attributed event narrative and self-measurement example with literature context—not a validated acute-injury statistic, sleep guidance, or medical advice. 4 sourcesSleep disruption / injury risk / self-measurement Female health A menstrual cycle becomes a logistics emergency: Johnson live-logs a six-hour blood-collection window Johnson live-logged the start of Kate Tolo’s period as an operational event: bleeding started 10:34 am, the cells “start dying immediately,” a laboratory must process the sample within 24 hours, and a roughly six-hour collection window opened—insertion cup, tubes, ice, packing, and a booked overnight courier activated by an “entire global team” to collect 7 mL of menstrual blood, with a missed window meaning a 35-day delay. He says the blood is “being used for frontier science…more details soon.” The post is the executed follow-through on the July 31 proposal to treat menstrual blood as a repeatable uterine sample, and it shows the program’s operational intensity: cycle-triggered sampling run like a time-critical logistics event. It reports collection timing and volume only—no assay, analysis, result, or validation of the sampling method. The dashboard records an attributed N=1 program-operations update—not a validated diagnostic workflow, evidence the sample is scientifically useful, or medical advice. 2 sourcesFemale health / biosampling operations Female health Johnson declares Kate Tolo’s current cycle “the most measured menstrual cycle in history” After “waiting for 34 days,” Johnson posted running totals for Kate Tolo’s current menstrual cycle: 408 minutes of Kernel brain data, ~10,000 blood-glucose readings, 48,960 core-temperature readings from an ingestible pill, 136 body photographs, and a stated 14 million data points overall—across a roughly thirty-item inventory spanning cortisol, DNA methylation, oral/stool/vaginal microbiomes, urine hormone metabolites, cervical mucus, sleep, wearables, resting metabolic rate, reaction time, skin biological age, and Kernel brain measurement—calling Tolo “the world’s most measured woman.” The totals match the scale of the 100-day, 14-million-datapoint baseline announced in July, so this reads as a mid-program tally of self-reported counts: no dataset, assay methods, results, or clinical interpretation accompany it. The dashboard records an attributed N=1 program readout—not completed female-health evidence, a representative protocol, or medical advice. 1 sourceFemale health / program readout Measurement Johnson publishes his Kernel brain-scan percentiles and reads them as a personality operating manual Johnson posted regional percentiles from his own Kernel brain measurements—amygdala 95th (emotion/threat), putamen 99th (learning/habits), caudate 97th (goals/decisions/working memory), globus pallidus 93rd (movement), and frontal gray matter at the 78th percentile “and climbing,” up 1.35% over two years—mapping each region onto his personality (processing hate as information rather than threat, fast habit formation, juggling founder/creator/“rejuvenation athlete” threads). He cites twin-study heritability of roughly 0.75–0.89 for the subcortical regions and gestures at real plasticity literature: the juggling result is Draganski et al., 2004 (Nature; 3 months learning to juggle produced a transient gray-matter expansion selectively in visual-motion cortex hMT/V5 that receded when practice stopped—not frontal, and over months rather than “within weeks”), and the exercise trial is Colcombe et al., 2006 (J Gerontol A; 6-month randomized trial in 59 sedentary adults aged 60–79: aerobic training increased frontal and temporal gray/white-matter volumes versus stretching controls, roughly matching his “people decades older than me” framing). But the post names neither study, the normative cohort behind the percentiles is undisclosed, and Johnson himself closes with “speculative for my n of 1 context.” The dashboard records an attributed self-report with literature context—not a validated brain-health metric, evidence that his protocol grew his frontal cortex, or medical advice. 3 sourcesBrain measurement / self-interpretation Biomarkers Johnson posts NTE-by-age decline tables that nearly match classic normative data—while naming none of it Two days after his NTE “personal best,” Johnson posted age-binned nightly-duration tables—Age 20: 190 min, Age 50: 103 min, Age 60: 81 min, Age 75+: 50 min—repeating that “ED predicts cardiovascular events yrs before symptoms” and the “2x risk of heart attack and stroke over 4 years” figure. The decline itself is among the better-documented observations in sleep/sexual medicine: Karacan et al., 1975 (Am J Psychiatry; 125 healthy males aged 3–79, EEG-verified) established age-normative NPT data, and Johnson’s age-20 anchor of 190 min sits almost exactly on published values—Horita & Kumamoto, 1989 (189 healthy males aged 3–84) measured 189.6 min of total nightly tumescence at age 20, declining to ~62 min at age 70—while Schiavi et al., 1988 (J Gerontol; 40 healthy men aged 23–73) showed NPT frequency and duration fall progressively with age independent of sleep changes. But the post names no study; its intermediate bins (103/81/50 min) match no single published cohort—older-age values differ between series (Horita: ~62 min at 70; Karacan’s series: ~102 min at 70)—and the repeated flat “2x in 4 years” remains stronger than meta-analytic averages (Vlachopoulos et al., 2011: RR 1.48 for CVD, 1.35 for stroke). Consumer NTE scores are still not a validated clinical risk tool. The dashboard records an attributed normative-data claim with literature context—not a validated risk prediction, a monitoring recommendation, or medical advice. 3 sourcesNighttime erections / aging Biomarkers Johnson publishes a nighttime-erection “personal best” and calls NTE a tier 1 longevity biomarker Johnson posted a self-reported record—4 hr 2 min of nighttime erections (NTE) at “93% strength,” ranked against a personal database of 55,000 measurements—calling NTE “a tier 1 longevity biomarker as critical to systemic health as VO2 max, resting heart rate, HRV, and blood pressure.” He claims low nighttime-erection scores mean “2x risk of heart attack and stroke in the next 4 years” and lists his protocol: high sleep quality, cardiovascular fitness, a calm nervous system, and daily tadalafil 5 mg (“I take this longevity medication daily”). The direction tracks real literature—erectile dysfunction precedes and predicts cardiovascular events (Vlachopoulos et al., 2011 JACC meta-analysis of 12 cohorts, 36,744 men: RR 1.48 for CVD, 1.35 for stroke; the Krimpen population cohort found HR 2.6 for MI/stroke/sudden death only in *severely reduced* rigidity over ~6 years)—but the post names no study, and its flat “2x in 4 years” is stronger than the meta-analytic averages. Consumer NTE scores are not a validated clinical risk tool, and daily tadalafil is a prescription medication requiring clinician oversight. The dashboard records an attributed N=1 biomarker claim with literature context—not a validated risk prediction, a monitoring recommendation, or medical advice. 3 sourcesNighttime erections / cardiovascular risk
// Reference paths
Move from feed to dossiers
Each route keeps source trails and medical caution visible. Protocol Health
Habits, Longterm, and Don’ts for measurement loops, sleep consistency, oral care, exercise, and experimental-intervention caution.
Protocol LongevityHabits, Longterm, and Don’ts for biomarkers, healthspan basics, critiques, and low-confidence immortality claims.
Protocol NutritionHabits, Longterm, and Don’ts for plant-led meals, Blueprint products, microplastics claims, and metabolic-drug caveats.
Protocol SleepHabits, Longterm, and Don’ts for bedtime discipline, front-loading examples, and metric caution from tweet-backed sources.
Reference Concepts39 source-aware glossary entries, including concepts without dedicated wiki pages yet.
entity BlueprintBlueprint is bryan-johnson's longevity protocol, product company, and public measurement system. It began as Project Blueprint, an N=1 effort to measure Johnson's organs and biological aging, and later became a consumer
concept Blueprint ProtocolThe Blueprint Protocol is bryan-johnson's closed-loop health system: measure body state, consult scientific evidence, implement interventions, measure again, and update. Johnson treats the body as a system whose organs a
concept Don't DieDon't Die is bryan-johnson's community slogan, longevity philosophy, and civilizational frame. The official Don't Die page describes a community “united in defeating all causes of human and planetary death and building a
concept Biomarker-driven longevity protocolsA biomarker-driven longevity protocol is a health system that treats the body as a measurable control system: collect biomarkers, choose interventions, re-measure, and iterate. project-blueprint is the most visible curre
Signal MetricsInspect source velocity by range, content type, and fallback API state.