bio age

· Zi Wang · 1 min read

Z / Runner 🏃🏻‍♂️

🏃🏻‍♂️ <3 the "that's our strava, oura, reju oly'".

  • We know what to measure (brain, muscle, microbiome, epigenetic, aging-pace), but we can't measure it. So the hard question is how do we build a system that gives everyday people access to the signals that matter, when the signals are locked behind invasive procedures and expensive sequencing? What is the AI-native, accessible proxy for these "mission impossible" biomarkers?

🏃🏻‍♂️ product hypothesis: build ai-native proxies for inaccessible biomarkers. Core problem: the highest-signal biomarkers of aging cannot be measured by normal people (csf proteomics, skeletal muscle transcriptomics, microbiome sequencing, methylation patterns…), BUT, day-to-day behaviors are accessible, movement, sleep, diet. MAYBE, ai can reconstrucut the trajectory of "impossible biomarkers" by stitching together the everyday signals.

  • voice + cognition patterns ≈ brain age slope. Csf [cerebrospinal fluid] clock track changes in brain-specific proteins, hardest to measure, but ai agent can track and approximate speech speed, memory recall, reaction time, focus drift… to create longitudinal curve to approx' csf aging.
  • strength curve + recovery curve ≈ muscular transcriptomic aging. No biopsy, use fitness outputs such as rep, loaded weights, gait variance, mobility degradation, stability under load, force output, soreness patterns map out muscle age.
  • diet behavior + postprandial response ≈ microbiome aging. Capture daily food intake, use CGM to infer microbial diversity.
  • sleep + stress + exercise ≈ epigenetic aging slope. Enough wearables to correlate methylation drift, use sleep scores, HRV, resting heart rate, weight, BMI, training load as a strong proxy.

🏃🏻‍♂️ my product hunch: a daily ai "peer", collect routine behavior patterns, give a "loose" proxy score for aging velocity. This maybe our Whoop recovery score, a strava for longevity.

  • 🏃🏻‍♂️ PDF files (drive link).

Stephen / Basketball 🏀

🏀 open source of calculating biological age (methyl groups' placement on dna)? can the metric of life expectancy be more exciting instead? brain / neurological age / duneid aging pace / skeletal muscle / microbiome / cardiorespiratory (vo2max) / healthspan escape velocity / ..? that's our strava (vs rejuvenation olympics or oura circles)!

  • 43.6 - 34.9 = 8.7 or 20.0% for zi, 46.8 - 46.4 = 0.40 or 0.86% for stephen, 35.5 - 33.1 = 2.40 or 6.77% for aaron, 42.6 - 28.8 = 13.8 or 32.4%
  • the science of super longevity (morgan levine at altos labs); an epigenetic clock analysis.

🏀 muscle gain hugely depends on the optimal amount of protein. how to experiment on your individual amount more accurately and scientifically? what's the _% difference in the effectiveness of natural food, protein bars (EPG + sucralose), protein powders?

  • z, export the full notes from attia? intake distribution, don layman. (already posted in full? optimal intake, 299 luc van loon.).
  • aaron's workout log vs aaron's research on protein, best powder, best bar. KING of all protein bar is David.