Nine Things
The whole protocol, stripped to only the levers whose removal you would actually notice. About an hour a day.
There's a failure mode built into ambitious protocols that the protocols themselves cannot see: almost every intervention in health is dose-over-time rather than dose-per-session. Which means a maximal protocol executed at 60% adherence loses to a minimal one executed at 95%.
So this is the brut version. Brut as in champagne — nothing added. Everything optional, marginal or merely pleasant removed, leaving only the levers with the largest effect per unit of time and money.
The selection rule isn't comfort. Several survivors are unpleasant, and several casualties are enjoyable.
The nine
1. Fix the wake time. Seven days a week. Same time daily, daylight in the eyes within 30 minutes, cold dark room, no alcohol. Regularity predicts mortality better than duration does, and this costs nothing. Everything else on this list works better when it's fixed.
2. Lift three times a week, 45 minutes. Full body, every other day. Two sets per movement, 5–8 reps, within a rep of failure. Six movements total: squat pattern, hinge, horizontal push, horizontal pull, vertical push, vertical pull. Load the stretched position. Add weight or a rep whenever you can.
3. One hard interval session a week. 4 × 4 minutes at an effort you couldn't hold for six. Thirty minutes including warm-up. Each 1-MET of fitness gained associates with roughly 13% lower all-cause mortality, with no observed ceiling.
4. Walk, outdoors, a lot. 8–10k steps. Prolonged sitting suppresses lipoprotein lipase in the legs and impairs glucose disposal independently of whether you trained that morning — a single workout doesn't buy it back. A ten-minute walk after the largest meal is the highest-yield placement.
5. Hit protein. Ignore the rest of nutrition. 1.6–2.2 g/kg daily across 3–4 feedings of 30–40g. Leucine binds Sestrin2, releasing GATOR2 inhibition so mTORC1 gets recruited to the lysosome — a threshold effect at roughly 2.5–3g leucine per meal, not a gradient. Protein also has the highest thermic effect of any macronutrient and the strongest satiety signal, so it does most of a diet's work without requiring one.
6. Four supplements. Creatine 5g. Vitamin D3 titrated to a measured level. EPA/DHA 1–2g. That's it. Everything else in the longevity aisle is unproven at the outcome level, refuted, or too small to notice.
7. Sunscreen daily, retinoid at night. UV activates AP-1, which simultaneously upregulates collagen-degrading MMPs and suppresses TGF-β-driven collagen synthesis. Sunscreen blocks the input; retinoids bind nuclear RAR/RXR and directly inhibit AP-1 while raising procollagen transcription. This is the best-evidenced anti-aging intervention that exists, and it's about £15.
8. Floss. Periodontal pathogens sustain a chronic systemic inflammatory burden with a direct bacteraemic route into the circulation. Two minutes a day. Nobody will ever make a video about it.
9. One blood panel a year. ApoB, Lp(a) once ever, fasting insulin, HbA1c, hs-CRP, vitamin D. Measurement is what separates a protocol from a set of habits you feel good about.
What got cut, and why
| Cut | Does it work? | Why it didn't survive |
|---|---|---|
| Zone 2 volume | Genuinely, yes | Large time cost for an adaptation the weekly interval largely covers. First thing to add back |
| Cold exposure | Real catecholamine response | Blunts hypertrophy near lifting; no unique long-term outcome |
| Sauna | Good observational data | Cut for access cost only. If you have one, it'd rank about fifth |
| Daily meditation | Real attentional effects | High adherence cost, slow accrual |
| The supplement stack | Mostly unproven | Cost and attention, no measurable return |
| Macro tracking | Works redundantly | Protein target plus whole food captures nearly all of it |
| Wearables and daily HRV | Real data | Measurement without an attached decision is a hobby |
| Anything prescription | Varies | Requires a physician, carries real risk, and isn't where the first 80% lives |
The uncomfortable observation: almost everything cut here is more interesting than almost everything kept. Interest and effect size are close to uncorrelated in this field, and the content economy pays for interest.
When to run this
Always, if you're not currently running anything — nine things done properly beats forty done partially, and the ramp is what kills most attempts. During any period where adherence is under pressure. And as a diagnostic: run it clean for eight weeks, and whatever hasn't improved is where the additional effort should actually go, rather than where you assumed.
Call it an hour a day and about £15 a month. The full protocol adds perhaps two more hours daily and an order of magnitude of cost for the remaining fraction.
Both are defensible. Only one gets done for a decade.
This is one lever out of a larger system. The full protocol — every mechanism, every dose, every source tier, and an append-only record of everything it has gotten wrong — is at /protocol.
More writing
One Kinase, Two Answers
Endurance training and hypertrophy don't merely compete for time. They send opposing instructions to the same signalling node, and the conflict has a name: AMPK. Here's what it actually does, how big the effect is, and how to schedule around it.
Measurement Without a Decision Is a Hobby
More data makes the signal-to-noise problem worse, not better, unless the analysis accounts for it. A counterweight to my own tracking article, and the rule that decides whether a metric earns its place.