Source Tiers

The four tells of a surface-tier source: a study named but never its effect size; a mechanism described but never its limitations; certainty that outruns the literature; and something for sale at the end that the certainty was manufacturing demand for.

Fight Aging! — researcher-adjacent

Mechanism-first tracking of the damage-repair paradigm, running for two decades, and unusually willing to publish things that cut against its own thesis. Its recent threads worth following: enhancing aquaporin-4 activity at astrocyte endfeet to improve glymphatic tau clearance (reframing Alzheimer's partly as a clearance failure rather than purely an overproduction problem); mitochondrial aging as an *adaptability* failure rather than an output failure; senescent-neutrophil clearance via EP2 blockade in tissue macrophages, distinct from classical senolytics; and TET2-mutant clonal haematopoiesis linked to epigenetic clock acceleration.

Lifespan.io / Longevity.Technology

Better on the funding, policy and regulatory layer than on mechanism. Useful for tracking what is moving structurally.

Optispan — Matt Kaeberlein

The standout. An actual aging-biology researcher with genuine mTOR and rapamycin pedigree, running patient-facing content that cites primary literature directly and cross-examines other researchers on mechanism rather than protocol. The rare case where the most credible person in a subfield is also the one communicating publicly.

Peter Attia

Not a bench researcher, and worth holding at the right tier for that reason — but he consistently brings on and genuinely cross-examines primary researchers, and his own taurine deep-dive engaging directly with the original Science paper is a good demonstration of the standard.

Full tier table in the mechanism ledger. In short: Chris Beardsley is the benchmark for mechanism, working from primary literature at the fibre level. Milo Wolf, Pak Androulakis-Korakakis, Brad Schoenfeld and Eric Helms are the publishing academics running the trials — expect them to disagree with Beardsley in places, which is what a healthy field looks like. Greg Nuckols brings the best statistical literacy in the space. Elijah Mundy, Yo Talks and TJR sit at operator tier: excellent execution instinct and real coaching reps, which is genuinely valuable and genuinely not a citation.

The distinction to hold: a researcher tells you why a set works. An operator tells you what the set should feel like. Both are load-bearing. Neither substitutes for the other, and the common failure is a popular channel presenting operator-tier intuition in researcher-tier packaging.

Rodney Sinclair (Professor of Dermatology, University of Melbourne) is the primary-research-tier reference for androgenetic alopecia — a large publication record across minoxidil and finasteride RCTs and the low-dose oral minoxidil consensus work.

Baremetrics 'Open Startups' / 'Open Benchmarks' — the strongest resource in this whole domain

Live, public revenue dashboards for real companies, plus a 600+ company benchmark dataset sourced directly from Stripe and billing integrations rather than from self-report. This is the closest thing to primary data that exists in business content — structurally analogous to working from study data rather than from anecdote. If one source in this section is worth building a habit around, it is this one.

Pieter Levels (levels.io) — operator, verified over a decade

Specific, dated revenue figures published continuously for over ten years, from $202,785/year in 2015 through a viral $420,000/month spike in 2024 to a steadier ~$105,000/month (~$80,000 profit) in 2026 for Photo AI — run on a single 40,870-line PHP file on one VPS handling roughly 4 billion requests a year for $244/month. No course, no mastermind, no funnel. The mechanism is stated plainly: extreme solo-founder efficiency, minimal infrastructure cost, building in public as distribution.

Andrew Youderian / EcommerceFuel — a real survey instrument

The annual Trends Report is built from roughly 300 real operators opening their actual books — margins, growth, ad spend by revenue tier — tracked longitudinally by the same organisation year over year. One genuinely humbling data point from it: ecommerce net margins reportedly fell from ~17.7% to ~10.6% over the past decade, which is the single best available rebuttal to 'anyone can start a seven-figure store'.

Patrick McKenzie (patio11) — 'Bits About Money'

The closest equivalent to Fight Aging!-grade rigour for understanding how money and business infrastructure actually work. A real operator — four software companies, pandemic-era state vaccine and unemployment infrastructure, Stripe — writing on pricing, negotiation and payments mechanics from disclosed real transaction data. One disclosure this section's own standard demands: much of it was written while he was employed by Stripe, so it is not incentive-free — it is simply unusually candid about the incentives it has.

Dan Shipper / Every

Runs an actual AI-native media and software company (five products, ~15 people, a disclosed seven-figure consulting arm) and writes candidly about restructuring a team around AI agents. Granular financials are paywalled, so treat the figure-level claims as unverified while the operating account itself is real.

The demotion worth making explicit: agency founders and brand operators who publish frameworks but never numbers sit at press tier, not operator tier — however good the frameworks are. The bar for 'operator' is disclosed, dated, verifiable figures over a period long enough to include a bad year. Almost nobody in the ecommerce content economy clears it, and the handful who do are listed above.

• Red flags for a fake operator: lifestyle-flexing without disclosed business evidence; revenue and profit used interchangeably; a promise that years of experience compress into under 90 days; denial that luck plays any role; expertise generalised from one product in one niche; and fear-based 'you've been doing it wrong' framing.
• Structural shift worth tracking: Brynjolfsson, Chandar & Chen's 'Canaries in the Coal Mine?' (Stanford Digital Economy Lab, 2025), working from ADP payroll microdata, found employment for software developers aged 22–25 down roughly 20% from its late-2022 peak, against flat-to-rising employment for older cohorts in the same occupations — characterised as structural rather than cyclical. The traditional career-capital route of doing unglamorous entry-level reps to build rare skill is exactly what is being automated. Read the working paper directly rather than the summaries of it.
• The index-versus-active evidence base remains among the most rigorously documented findings in finance: S&P's SPIVA scorecard, running since 2002, has consistently found no equity or fixed-income category in which a majority of active managers beat their benchmark net of fees over 15 years — and survivorship bias understates it, since roughly two-thirds of funds active 20 years ago no longer exist. If you want the methodology rather than the headline, Larry Swedroe (long-time director of research at Buckingham) is the best writer on it — but note he argues *for* the passive and factor-based conclusion, not against it. Read him as the rigorous version of the finding, not as the opposition.

• Relationship science: R. Chris Fraley (attachment measurement), Paul Eastwick and Eli Finkel (attraction, ideal preferences), Harry Reis (perceived partner responsiveness). Finkel, Simpson & Eastwick's 'Fourteen Core Principles' (Annual Review of Psychology, 2017) is the entry point.
• Philosophy: the Stanford Encyclopedia of Philosophy is free, written by specialists in each area, and better than almost any book-length secondary source. Use it to find the primary paper, then read the primary paper.
• Cognition: Examine.com for supplement evidence synthesis — it grades by evidence quality, states effect sizes, and sells no supplements, which puts it in a category of roughly one.