Live Better Science
Apr 2026
Confidential — for Dr. Rimm
The Personal Scientific Evidence System

Your source of truth
on yourself.

01 · per question
Dossier
A compiled answer. Sources, personalization, math, caveats.
02 · per person
Codex
Your lifetime compiled evidence body. Grows dossier by dossier.
03 · per platform
Corpus
The aggregate. Consented, structured, unmatched in the world.
Karen Nguyen & Scott Hughes · founders
Deck v0.1
Expert consult · Harvard i-lab
The Gap

Two explosions that never met.

On one side, the world's peer-reviewed biomedicine keeps exploding. On the other, your own biological data is now cheap, rich, and yours. Between them sits no compiler — no system that takes today's Nature Medicine paper and tells you what to do about your body.

Published biomedicine
  • 2M+ new biomedical papers per year
  • Huge public cohorts · CRISPR screens · -omics panels
  • Foundation models that can read the whole thing
  • All trapped in PDFs and supplementary tables
missing compiler
the gap
Your biological data
  • $200 whole-genome sequencing
  • Function Health, Superpower, InsideTracker panels
  • CGMs, wearables, microbiome, epigenetic clocks
  • All scattered across apps that don't talk

"No system takes today's Nature Medicine paper and tells you what to do about your body. We are that system."

02 / 12 Live Better Science Confidential — Apr 2026
The Ontology

Three primitives. One engine.
Every question your biology can ask.

Palantir turned messy enterprise data into queryable ontologies. We do the same for personal biomedicine: three primitives, one compiler, one output unit that can be audited line by line.

Primitive 01

Study

Any peer-reviewed dataset. A 10,068-person cohort. An 18,000-gene CRISPR screen. A foundation model's published tables.

Primitive 02

Self

Your biological reality. Genome. Labs. Microbiome. Epigenetic age. CGM stream. Hormones. Skin imaging.

Primitive 03

Dossier

A compiled, traceable answer. Every line cites a published number and a value from your data. No black box.

The Engine
The Science Compiler
Study × Self → Dossier
Verticals — the beachhead, then the rest
Longevity Aesthetics Disease Risk Drug Response Nutrition Fertility Sleep Cognition Pediatric Elder Care
03 / 12 Live Better Science Confidential — Apr 2026
A Dossier

A real answer, shown side-by-side.

Same question, two systems. On the left, a typical AI health app. On the right, a Live Better Science Dossier. Only one can be audited.

Today · any AI health app

An opinion you have to trust

Should I eat more legumes, given my microbiome and lipids?
"Yes — legumes are a great source of fiber and plant protein. Aim for 3–4 servings per week!"
  • No source cited
  • Same answer for every user
  • Cannot be audited or reproduced
  • Evaporates on next model update
Live Better Science · a Dossier

An answer with its derivation attached

Should I eat more legumes, given my microbiome and lipids?
Yes — legumes score +37.31 under your plant-forward preset.
Derivation · every line is a line-item
Source
Segev et al., Nature Medicine 2026 · n=10,068 · 4-year follow-up
Pub #
WFPB strength = 22.66 · 518 microbes q<0.05
Your body
B. theta = 2.1% low · ApoB = 114 elevated
Compute
22.66 × 1.4 × 1.18 = 37.31
Caveats
Cohort skewed 45–65; medium confidence for your ancestry subgroup.
Dossier #LBS-2026-04-21-00184 Reproducible · Auditable · Yours
04 / 12 Live Better Science Confidential — Apr 2026
Why Now

The substrate arrived this quarter.

Personal biological data has been getting cheap for years. But the models capable of composing it against the world's science — for real, at citation-grade provenance — only shipped in the last six months. Every frontier lab released a bio stack in the window that opens with this company.

Your data is cheap.

The personal-biology substrate

  • $200 whole-genome sequencing — 1000× cost collapse in a decade
  • Function Health · Superpower · InsideTracker — 110+ marker panels
  • CGMs · wearables · microbiome kits — continuous streams
  • At-home epigenetic clocks — biological-age tracking
The models arrived.

The frontier-bio substrate

  • AlphaGenome · DeepMind · Jan 2026 — genome variant effect
  • Evo 2 · Arc + NVIDIA · Mar 2026 Nature — 100K+ species, 40B params
  • Claude Operon · Anthropic · Mar 2026 — CRISPR, scRNA, molecular biology
  • GPT-Rosalind · OpenAI · Apr 16 · five days ago
  • AlphaFold 3 · DeepMind · public server

"Every frontier lab released a dedicated bio stack in the last six months. The substrate is not coming. It arrived."

05 / 12 Live Better Science Confidential — Apr 2026
Traction

The methodology works — 22 times over.

We compiled the same methodology across six biomedical domains in one month. Every tool cold-starts in a clean environment. An independent quality audit ranked the portfolio #1. Two Harvard Medical School PhD candidates have committed to join pending funding.

01

22 peer-reviewed papers

All on clawRxiv (Stanford/Princeton competition track, Apr 2026). 20 compiled tools across 6 biomedical domains.

02

Ranked #1 in independent audit

CQI 68.1 / 100 — highest among prolific agents on the platform. Zero competitors could even install in a clean environment.

03

Published at AAAI 2026

AI research beyond biomedicine. Publication capacity at a top AI venue, in parallel with the biomedicine portfolio.

Domains covered
Longevity Cancer genomics Nutrition / microbiome Sleep medicine Protein engineering Gene annotation + 335 automated tests
Talent signal
Two PhD candidates from Gladyshev & Sinclair labs (Harvard Medical School) have committed to join — pending the close of this round.
06 / 12 Live Better Science Confidential — Apr 2026
The Business

Three layers. One ontology.
Compounding revenue.

Human-services beachhead generates proprietary structured data. The structured data becomes the product. The ontology and the revenue model are the same shape — Dossier becomes Codex becomes Corpus.

Layer 03 · Platform
Corpus
Consented, structured real-world evidence licensed to pharma, insurers, academic researchers. The compounding data moat.
YEAR 3+
platform licensing
Layer 02 · Product
Codex
Consumer subscription — bring your own labs. Ask questions. Your Codex accumulates a lifetime of Dossiers.
YEAR 2
$99 – $499 / mo
Layer 01 · Practice
Dossier
Concierge membership — MD-led. Quarterly panels, genome, microbiome, epigenetic age, aesthetics. Every consult produces Dossiers that seed the member's Codex.
NOW · YEAR 1
$10K – $50K / yr
Human-services beachhead generates proprietary structured data. Structured data becomes the product. Same shape that turned Palantir's human FDEs into the Foundry moat.
07 / 12 Live Better Science Confidential — Apr 2026
The Beachhead

Concierge as the front door.

Today's concierge longevity market is growing fast and has no evidence layer. We enter as a credentialed practice, deliver a category others can't, and seed every member's Codex from day one.

Membership structure

Three tiers, one shared compiler

Founding $15K / yr · cap 300
Quarterly panels, genome, microbiome, continuous Dossiers.
Standard $25K / yr
Founding + epigenetic age tracking, aesthetics, named MD.
Invitation $50K+ / yr · cap 50
Bespoke protocols, full-stack multi-omic, family member extension.
Comparables (what they lack)

We are the first with a Dossier layer

  • Fountain Life — $20K–$100K/yr · no compiled evidence layer
  • Next Health — $15K–$30K/yr · spa + diagnostics, no citations
  • Human Longevity Inc. — $25K+/yr · strong diagnostics, no personalized compilation
  • Sollis Health — $7K–$15K/yr · concierge medicine, not longevity-science
  • Function Health — $500/yr · panels without integration
The practice is the only way to train the compiler on real patients at scale. Every concierge consult produces Dossiers. Every Dossier grows the member's Codex. Every Codex, consented and structured, feeds the Corpus. The clinic is not a side business — it is the data-generation engine.
08 / 12 Live Better Science Confidential — Apr 2026
The Moat

A Codex that compounds,
on a stack that costs nothing to build.

The clinical moat is the Codex — a proprietary personal evidence body that cannot be replicated. The technical moat is what sits above it: every frontier lab shipped a bio stack in the last six months, and we compose them all through our own Dossier Verifier.

The Clinical Moat

The Codex · The Corpus

01 Grows with every memberHundreds of Dossiers per member per year.
02 Grows with every paperNew study → new Dossiers across every member.
03 Consented and structuredRegulatory-clean RWE nobody else holds.
04 Methodology IP22 peer-reviewed papers document the method.
Not replicable by any AI company without a clinical practice. Not replicable by any clinic without the compiler.
The Technical Moat · Composed, Not Trained

The substrate arrived in the last six months

Layer 01 · Self
Biology foundation
AlphaGenome · DeepMind Evo 2 · Arc + NVIDIA AlphaFold 3 · DeepMind Claude Operon · Anthropic
Layer 02 · Study
Frontier reasoning
GPT-Rosalind · OpenAI Claude Opus 4.7 (1M ctx) GPT-5.4
Layer 03 · Retrieval
Grounded RAG
PaperQA-class agentic RAG pgvector + citation graph
Layer 04 · Ours
The Dossier Verifier
Deterministic extraction Source-linked provenance 22-paper methodology
We compose, we don't train. DeepMind, OpenAI, Anthropic, and Arc shipped $10B+ of bio R&D into our stack in the last six months — all API or open-weight. Capital goes to the Verifier and the clinical practice that trains it.
09 / 12 Live Better Science Confidential — Apr 2026
Market

Longevity and aesthetics first.
Then everything.

The beachhead is a $30B+ US market growing >15%/yr. The expansion path — every other domain where personal biology meets published science — is category-defining.

The Beachhead

Longevity + Aesthetics

~$30B
US market · >15% CAGR · zero incumbents with a compiled-evidence layer
The Expansion

Same ontology, new verticals

Disease risk screening · drug response / PGx · nutrition · fertility · cognition · sleep · pediatric · elder care. Each vertical is a new application of the same Dossier → Codex → Corpus stack.
The Platform

Corpus as infrastructure

De-identified structured RWE, biomarker discovery services, insurance risk modeling. The eventual business that looks most like what Rimm underwrites at scale.
A platform-shaped business requires a practice to earn it. We do not pitch a pre-product platform. We pitch a concierge practice that generates the structured evidence that becomes the platform.
10 / 12 Live Better Science Confidential — Apr 2026
Team

Two founders. Twenty-two papers.
Two HMS PhDs pending close.

A scientist and an operator who ship. Plus two PhD candidates from HMS's most-cited longevity labs, committed pending funding.

Founder · Science

Karen Nguyen

BA, MA in Biology (Harvard) · hon345@g.harvard.edu

  • Co-author · 22 peer-reviewed clawRxiv papers (Apr 2026)
  • Co-author · AAAI 2026 · AI research
  • Scientific direction, domain coverage across the 6 biomedical verticals
Founder · Operations

Scott Hughes

BA, MA, JD · Harvard Certificate in Entrepreneurship & Finance

  • Co-author · 22 peer-reviewed clawRxiv papers (Apr 2026)
  • Co-author · AAAI 2026 · AI research
  • Architected the Science Compiler and the Dossier Verifier
Committed pending close
PhD candidate · Gladyshev Lab (HMS)
Aging biology, epigenetic clocks, methylome — one of the most-cited aging labs in the world.
PhD candidate · Sinclair Lab (HMS)
Sirtuins, NAD metabolism, rejuvenation — founding architect of the modern longevity field.
AI-native operators. We compose frontier models, we don't build them. Our R&D and product stack use LLMs and ML end-to-end with deterministic verification — which is how two people shipped 22 peer-reviewed papers across six biomedical domains in one month, and why two HMS PhDs are willing to bet their careers on what comes next.
11 / 12 Live Better Science Confidential — Apr 2026
The Ask

Pre-seed. And three things we're still debating.

The Round
$1M – $3M pre-seed
Specific target and structure discussed live.
Use of funds
  • 2 PhD hires (Gladyshev · Sinclair) — committed pending close
  • Medical Director (MD) · concierge practice lead
  • Senior research engineer · compiler + retrieval
  • First clinical operator
  • Concierge practice launch · first 50–100 members
  • Compiler productization atop the composed frontier stack
Three open strategic questions

We'd love your take.

01 · Sequencing
Bootstrap via concierge revenue and raise later (cleaner, slower) — or raise pre-seed now to move faster on the compiler?
02 · First wedge vertical
Longevity as #1 (biggest market, noisiest) — or aesthetics as #1 (clearer outcomes, faster case studies)?
03 · Concierge positioning
Invitation-only luxury ($50K/yr, 50 members) — or broader founding tier ($15K/yr, 300 members) with a separate invitation track?
Thank you, Dr. Rimm.
Karen Nguyen · Scott Hughes · hon345@g.harvard.edu
12 / 12 Live Better Science Confidential — Apr 2026