Biohacking nutrition

TL;DR. Biohacking is consumer-grade self-measurement: CGMs, expanded blood panels, wearables, microbiome kits, time-restricted eating, zone-2 cardio, sauna, cold. The basics (real food, less ultra-processed, varied plants, sleep, movement) carry most of the metabolic benefit at zero cost. These tools are optional. They have uses and failure modes. If you want a starter setup: one yearly expanded blood panel, a two-week CGM trial, a basic sleep wearable, a 14-hour overnight eating window if it fits your life. Past that, most of what you can buy is marketing.

What you'll learn

• Casey Means's "bio-observability" framing and what it does and does not buy.
• What CGMs measure, where they help non-diabetics, and the "every spike is bad" trap.
• Direct-to-consumer biomarker panels vs. a primary-care visit.
• The expanded markers worth asking a doctor for: fasting insulin, HOMA-IR, ApoB, hs-CRP, vitamin D, ferritin, ALT/AST/GGT, uric acid, triglyceride-to-HDL ratio.
• What wearables measure well and what they get wrong.
• Time-restricted eating: the Spector / Panda evidence and where it stops being magic.
• Zone-2 cardio and resistance training as metabolic levers adjacent to nutrition.
• Cold and sauna: where the data is strong and where it is hype.
• Microbiome testing: what one stool sample can and cannot tell you.
• The risks: obsession, false confidence, gameable proxies, cost, eating-disorder adjacency.
• A short starter stack if you want one.

1. What "biohacking" means

Casey Means uses the phrase "bio-observability" in Good Energy. For most of medical history, your metabolic data lived in a clinic chart and updated once a year. Now, for a few hundred dollars, you can see your own glucose, sleep, HRV, and an expanded blood panel without a referral. That shift is also a market.

Most of the benefit is in the basics: real food, less ultra-processed, varied plants, sleep, movement. The rest of this curriculum covers those, and you can do all of it without spending a dollar on a sensor.

This module is for the reader who wants to look under the hood. The tools are an optional second layer, not a hidden tier of "real" nutrition. Some produce useful data; some produce expensive anxiety. The job is to tell which is which.

2. Continuous glucose monitors for non-diabetics

A CGM is a small arm patch with a hair-thin filament under the skin. It samples interstitial glucose every 1 to 5 minutes for 10 to 14 days, ~35,000 readings per patch. Consumer brands: Levels, Lingo, Stelo, Nutrisense, Veri. Subscriptions: $200 to $400 per month.

What is real. Your own post-meal response. The Berry et al. PREDICT-1 trial (Spector's group, Nature Medicine 2020) showed identical meals producing glucose, lipid, and insulin responses varying by a factor of 5 to 10 between people. White rice spikes one person and barely moves another. You also see interactions: a walk flattens the curve, fiber first blunts a later dessert, poor sleep raises everything the next day.

What is marketing. The "every spike is bad" framing. Non-diabetics spike from normal food. A peak of 140 to 160 mg/dL after a banana is physiology, not pathology. The cardiovascular and mortality data sit on fasting glucose, HbA1c, and fasting insulin, not on the area under one banana curve. A CGM gives you a reading, not a verdict.

The honest use. Wear one for two weeks, run a few n-of-1 experiments on foods you eat often, then take it off. The marginal value of month 3 is near zero. Spector has CGM access and does not wear one continuously.

3. Direct-to-consumer biomarker panels

Function Health, InsideTracker, SiPhox, and Marek Health run 30 to 100+ markers from one draw and ship results to your phone with an LLM-flavored interpretation. Annual subscriptions: $300 to $1,000+.

What is useful. A modern panel covers ApoB, Lp(a), NMR LDL fractionation, fasting insulin, HOMA-IR, hs-CRP, homocysteine, ALT/AST/GGT, uric acid, vitamin D, ferritin, sex hormones, thyroid, full iron, and a basic micronutrient set. More than a standard insurer-covered annual.

What is expensive optimization. Most healthy adults under 50 see little actionable beyond what a careful primary-care annual would catch. More markers means more "yellow" results, mostly harmless variation or false signals that trigger months of follow-up. The labs themselves are the same labs your doctor uses; the narrative wrapper is what you are paying for.

The honest use. If your doctor will not order the expanded markers in the next section, a one-time DTC pull is reasonable. The annual subscription rarely is. One pull every one to two years is enough for most people.

4. The expanded panel to ask your doctor for

This is the cheap version. Standard insurance covers most of these markers if you give the ordering doctor a reason; few doctors run them by default.

• Fasting insulin and HOMA-IR. Lustig's argument in Metabolical Ch. 18: fasting insulin rises years before fasting glucose or HbA1c, making it the earliest readable signal of insulin resistance. Standard panels rarely include it; you have to ask.
• ApoB. A direct count of every atherogenic lipoprotein particle. More accurate than LDL-C for cardiovascular risk. Many lipid specialists now consider it the single best lipid number.
• NMR LDL fractionation. Splits LDL into small-dense and large-buoyant. Useful if your LDL-C is borderline.
• Lp(a). Genetic, run once in a lifetime. Elevated Lp(a) is an independent cardiovascular risk factor.
• hs-CRP. A trailing marker of systemic inflammation. Read in context.
• Homocysteine. Tracks B-vitamin status and cardiovascular risk.
• Uric acid. Rises with fructose load and metabolic syndrome.
• Triglyceride-to-HDL ratio. Free, sitting in any standard lipid panel. Under about 1.5 (mg/dL units) tracks with good insulin sensitivity. Means flags this one repeatedly.
• ALT, AST, GGT. Elevated ALT and GGT in a non-drinker is one of the cleanest signals of fatty liver, the metabolic-syndrome organ most annual physicals skip.
• Vitamin D 25(OH)D. Most adults at northern latitudes test low. Supplementation evidence is mixed; the baseline is worth knowing.
• Ferritin and full iron panel. Iron deficiency in menstruating women and iron overload in older men are both common and both missed.
• Fasting glucose and HbA1c. Read together with fasting insulin.

Most insurers cover the standard subset and balk at fasting insulin, ApoB, Lp(a), and NMR fractionation without a stated indication. A direct request, framed around family history or cardiovascular risk, usually gets them ordered.

5. Wearables and the HRV question

Whoop, Oura, Apple Watch, Garmin, and Fitbit all sell "recovery" or "readiness" scores. Underneath the marketing, they measure four things at varying accuracy.

• Steps and movement. Reliable.
• Sleep. Total sleep time is good. Stage breakdowns are approximate; wrong about specific minutes but right about trends.
• Heart rate and HRV. Resting heart rate is reliable. HRV (beat-to-beat variation, a rough proxy for autonomic balance) is reliable on trends, noisy night-to-night. Read it as a moving average.
• Calorie burn. Poor. Often off by 30% or more. Do not feed it into a deficit plan.

Poor sleep flattens glucose tolerance the next day. A heavy late dinner or some drinks the night before show up as suppressed HRV and an elevated resting heart rate. You see the same effects on a CGM, faster and more directly. None of this replaces a doctor, a blood draw, or your own judgment; a high recovery score does not mean your ApoB is fine.

6. Time-restricted eating

TRE compresses eating into a daily window. Common variants: 14:10, 16:8, 18:6. Spector adopted a 14-hour window in Food for Life Ch. 11 and stayed there. Satchin Panda's lab at Salk has the deepest mechanistic literature, in mice and humans.

The evidence. Modest weight loss in most trials, mostly via calorie reduction (people eat less when they have less time). Improvement in fasting insulin and triglycerides in several trials. Easier to sustain than calorie counting. Front-loaded TRE (finishing dinner by 6 or 7 pm) outperforms skipping breakfast in most metabolic-marker trials, because circadian insulin secretion and glucose tolerance are better in the morning.

The honest limits. TRE does not replace food quality. It is hard for shift workers, hard during travel, and can collide with family dinner. The 18:6 frame is where TRE often turns into a restrictive identity; a 14-hour overnight window (dinner by 8, breakfast at 10) is enough for most people, and anything stricter shows diminishing returns.

7. Zone-2 cardio and resistance training

These are not nutrition. They sit next to nutrition because they change how the food you eat is handled.

Zone-2 cardio. San-Millán and Brooks, Cell Metabolism (2018), describe zone 2 as the pace where fat oxidation dominates and lactate clearance keeps up with production. A pace you can converse through but not sing through, around 60-70% of max heart rate. Means makes the Good Energy point that this intensity drives mitochondrial biogenesis, the growth of new mitochondria in muscle. More mitochondria, better fat oxidation, better glucose disposal, better insulin sensitivity. Dose: three to four hours per week, 45-60 minute blocks.

Resistance training. Skeletal muscle is the largest insulin-sensitive tissue you have. Building and keeping it changes glucose disposal at rest. Compound lifts (squat, hinge, push, pull, carry) two to three times per week is the minimum that shows up in body composition and metabolic markers.

Neither one replaces food quality, but stacked with TRE and decent sleep they compound.

8. Cold exposure and sauna

Also not nutrition, but biohacker culture lumps them in and readers ask.

Sauna. The Finnish KIHD cohort (Laukkanen et al., JAMA Internal Medicine 2015) followed about 2,300 middle-aged men for 20+ years. The 4-7 sessions per week group had roughly half the cardiovascular mortality of the one-session group. Observational, with the usual healthy-user-bias caveats, but the dose-response and the biological plausibility (raised core temperature, heart rate, heat-shock proteins) make this one of the better-supported lifestyle interventions in the consumer stack.

Cold immersion. The data is thinner. Most of the literature sits on surrogate markers (brown-fat activation, norepinephrine, short-term insulin sensitivity), and randomized trials on mood and recovery are mixed. Cold is fine in moderation if you enjoy it; the "magic metabolic intervention" framing is ahead of the evidence.

9. Microbiome testing services

Viome, Thryve, Sun Genomics, BiomeSight, and others sequence one stool sample and return a report on the species and pathways you carry. Pricing: $100 to $400+ per test.

What they measure. 16S rRNA sequencing (cheaper, bacteria to genus) or shotgun metagenomics (species and functional pathways). The technology is real.

What they can tell you. Which species and pathways were in your stool at one point in time.

What they cannot tell you. Causation. What to eat to change it. Whether anything is "missing" in any clinically meaningful sense. Whether their personalized food rules outperform "eat more fiber and varied plants." Spector, who built ZOE and runs the largest microbiome research database, is the most careful voice. In Spoon-Fed he repeats the same point: the products are moving faster than the science, and one stool sample run through a proprietary algorithm is not personalized medicine.

The honest use. Run one if you are curious, read the report, eat more varied plants, and skip the subscription.

10. The risks

This is the section enthusiast literature underplays.

Obsession. A CGM on the arm of an anxious eater produces a constant low-grade panic, every meal a test, every reading a verdict. Wearables do the same thing with sleep scores.

False confidence. Two weeks of CGM data is not your metabolic future, and a single blood panel is a snapshot. People build whole identities on one set of numbers, then are surprised when their biology moves next year.

Gameable proxies. If your only goal is a flat glucose curve, you can flatten it while wrecking your diet. Fat plus protein plus fiber lowers spikes; so does cheese and salami at every meal. A CGM tells you nothing about ApoB, fiber intake, or the microbiome.

Cost. A full biohacker stack (CGM, Function Health, Whoop, microbiome kit, sauna and cold plunge access, supplements) clears $5,000 to $10,000 a year. The dollar that buys 5% at the margin would buy 50% if a household that is not yet eating real food spent it on groceries.

Eating-disorder adjacency. This one matters most. A CGM in the hand of a restrictive eater, or someone in recovery from a clinical eating disorder, is fuel. So is obsessive macro tracking, or any tool that turns eating into a daily score. If you have a history of disordered eating, or if any of these tools is raising your anxiety around food, stop and step back to the basics. The relevant referral is the eating-disorders deep dive (D7).

11. The starter stack, if you want one

A short list for the reader who wants the data without spending five figures.

• One yearly expanded blood panel. Ask your primary-care doctor for the markers in section 4. Insurance covers most. If yours will not, a one-time direct-to-consumer pull covers the rest.
• One two-week CGM trial. Levels, Stelo, or Lingo. Run a few n-of-1 experiments on foods you eat often, then take the patch off and do not subscribe.
• A basic wearable for sleep tracking. Oura, a basic Garmin, or whatever you already own. Read trends, not nightly scores.
• A 14-hour overnight eating window if it fits your life. Finish dinner earlier and skip the late snack. Do not push to 16:8 unless it stays easy.
• Three to four hours per week of zone-2 cardio. Walk-jog, stationary bike, swim. Conversational pace.
• Two to three resistance sessions per week. Compound lifts, 20 to 40 minutes each.

The basics in the rest of the curriculum are the thousand-dollar dial; this stack is the hundred-dollar dial on top. The five-figure biohacker stack with CGM-for-life and four subscriptions and a cold plunge is mostly the ten-dollar dial dressed up. Spend in that order.

Frequently Asked Questions

Are CGMs for non-diabetics worth it?

For two weeks, yes, if you are curious and can afford it. You will learn your own post-meal patterns, which population averages cannot tell you. The long-term subscription is harder to justify. Most non-diabetics get the useful information in the first ten days. Continuous wear past a month tips into anxiety more than insight.

Should I get a Function Health subscription?

Probably not the recurring subscription. The annual draw is valuable; the year-round dashboard is not. If your primary-care doctor will run the expanded markers in section 4, you do not need Function at all. If not, a single one-time pull is reasonable. Re-evaluate every 1 to 2 years.

Is fasting insulin worth asking my doctor for?

Yes. It is the cheapest, earliest, most overlooked signal of insulin resistance. By the time fasting glucose or HbA1c is elevated, you have spent years on the curve. A fasting insulin under about 7 µIU/mL with normal fasting glucose is reassuring; higher numbers, especially with a triglyceride-to-HDL ratio over 2, are an early warning. Most doctors will run it if you ask.

Does Whoop or Oura affect what I should eat?

Not directly. A wearable shows the cost of a late heavy dinner, a few drinks, or a short night, which is useful feedback. It does not change the underlying rule. A high recovery score does not earn you dessert; a low one does not require fasting. Read the patterns over weeks, not meal-by-meal.

What is the cheapest version of biohacking?

Close to free. Ask your doctor for the expanded panel. Walk an hour a day, lift something heavy twice a week, eat your last meal earlier, and sleep seven to eight hours. That captures most of the metabolic benefit of the full stack with zero sensors and one annual draw.

Will continuous tracking give me an eating disorder?

For some people, yes. The risk is highest with a history of disordered eating, an anxious or restrictive relationship with food, or a tendency to score-keep every input. A CGM, a calorie tracker, or a macro app can each act as fuel. If a tool is making your relationship with food worse, stop and see the eating-disorders deep dive (D7). Self-tracking is optional; eating well is the part that is not.

Does cold exposure really do anything?

The biology checks out (norepinephrine release, brown-fat activation, short-term insulin sensitivity). The clinical evidence for cardiovascular or mortality benefit, at the level the sauna data sits at, is not there yet. Cold is fine if you enjoy it; the "major metabolic intervention" claim is ahead of the data. Sauna has stronger cohort evidence. Neither one substitutes for sleep, food quality, or movement.

How long should I stay on TRE?

As long as it stays easy. A 14-hour overnight window (dinner by 8, breakfast at 10) is sustainable indefinitely for most people and captures most of the metabolic benefit. The 18:6 and 20:4 windows are harder to sustain, harder to socialize around, and produce diminishing returns. If TRE starts to feel restrictive, or you find yourself anxious about meal timing, stop. None of this should add pressure to your eating.

Sources

1. Means, C. Good Energy: The Surprising Connection Between Metabolism and Limitless Health (Avery, 2024). Ch. 1-3 on bio-observability and the shift from clinic-only to consumer-grade metabolic data.
2. Lustig, R. Metabolical: The Lure and the Lies of Processed Food, Nutrition, and Modern Medicine (Harper Wave, 2021). Ch. 18 on fasting insulin and HOMA-IR as early markers of insulin resistance.
3. Spector, T. Food for Life: The New Science of Eating Well (Jonathan Cape, 2022). Ch. 1 on the ZOE PREDICT study; Ch. 11 on Spector's own 14-hour TRE adoption and the limits of one-size advice.
4. Spector, T. Spoon-Fed: Why Almost Everything We've Been Told About Food Is Wrong (Jonathan Cape, 2020). On microbiome testing companies running ahead of the science.
5. Berry, S. E., Valdes, A. M., Drew, D. A., et al. "Human postprandial responses to food and potential for precision nutrition." Nature Medicine, 2020;26(6):964–973. The PREDICT-1 paper documenting wide between-person variance in glucose, lipid, and insulin responses to identical meals.
6. Hall, K. D., Ayuketah, A., Brychta, R., et al. "Ultra-processed diets cause excess calorie intake and weight gain: an inpatient randomized controlled trial of ad libitum food intake." Cell Metabolism, 2019;30(1):67–77.
7. Laukkanen, T., Khan, H., Zaccardi, F., Laukkanen, J. A. "Association between sauna bathing and fatal cardiovascular and all-cause mortality events." JAMA Internal Medicine, 2015;175(4):542–548.
8. San-Millán, I., Brooks, G. A. "Assessment of metabolic flexibility by means of measuring blood lactate, fat, and carbohydrate oxidation responses to exercise in professional endurance athletes and less-fit individuals." Sports Medicine, 2018;48(2):467–479.
9. Panda, S. The Circadian Code: Lose Weight, Supercharge Your Energy, and Transform Your Health from Morning to Midnight (Rodale Books, 2018). And primary papers from the Panda lab on time-restricted feeding in mice and humans.
10. Ross, A. C., Caballero, B., Cousins, R. J., Tucker, K. L., Ziegler, T. R., eds. Modern Nutrition in Health and Disease, 12th ed. (Wolters Kluwer, 2024). Ch. 123 on precision nutrition: where it has evidence and where it is exploratory.

Related glossary terms

• CGM
• Fasting insulin
• HOMA-IR
• ApoB
• HRV
• Time-restricted eating
• Zone-2 cardio
• Bio-observability