“The rules we treat as given usually turn out to have an author, and the author was often just guessing.”

Eight glasses of water a day began as a guess.

The Food and Nutrition Board recommended 2.5 liters a day for adults, then added the part everyone forgets: most of that water already comes from the food you eat.

Fredrick Stare founded Harvard's nutrition department. He wrote that the average adult needs around six to eight glasses in twenty-four hours, then added the other forgotten part: coffee, tea, milk, soft drinks, and beer all count. His obituary described him as an early champion of drinking at least six glasses a day, so the number was a position he held for a long time, not a finding he nailed down.

When a researcher named Heinz Valtin went looking for the source in 2002, those two statements were all he found. Nobody has traced the specific hop from either one to the rule everyone now recites. It spread through health columns and physicians repeating each other.

Then bottled water became an industry in the 1980s, and a guess with no author got a marketing budget and a mascot: the gallon jug on a man’s desk. Now men carry gallon jugs to their desks and feel behind if they miss a number that was never really a number.

Ten thousand steps a day began as a product name.

A Japanese clock company built the first consumer pedometer after the 1964 Tokyo Olympics and called it the Manpo-kei, the ten-thousand step meter. The number was never a health finding. They picked the number the way a product name is picked: it sounded auspicious, and its written character resembles a person walking.

A researcher attached to the story noticed the average Japanese person walked three to five thousand steps a day and figured ten thousand was a worthy target. Japanese walking clubs adopted it and kept it alive for decades. Then Fitbit shipped ten thousand as the default goal on every device, and every rival followed, because a default is easier to copy than to defend.

A man now checks the number on his wrist before he checks how his body feels. He is chasing a target named by a clock company.

The saturated fat rule began with a fish scientist.

Ancel Keys studied the physiology of fish before he turned his attention to human diets. The seed of the saturated fat rule had nothing to do with what a family ate for dinner. A man named Nikolai Anitschkow fed cholesterol to rabbits and produced fatty deposits in their arteries. From that seed, Keys traveled to over 20 countries. He cherry-picked seven that supported his hypothesis, built an idea from them, and took it to the World Health Organization.

He got a seat on the American Heart Association’s nutrition committee. The next year, the AHA adopted his position as official policy and told American men to cut saturated fat. Keys landed on the cover of Time magazine the same year. Then the New England Journal of Medicine ran a two-part review naming fat and cholesterol the culprits. The Sugar Research Foundation paid for it. A Senate committee repeated it. The first Dietary Guidelines repeated it.

The rabbits stayed in the lab, and the policy walked out into every kitchen in the country.

None of these began as what they became. A guess, a product name, a seed funded by someone with something to sell. Institutions repeated them until they felt like the laws of nature.

By the time you met them, they were no longer claims. They were just the way things are.

Does that make the rules worthless? No. It makes your attention worth something.

It means the number you are chasing has an author, and authors have motives.

Two psychologists rigged a wheel of fortune in the 1970s to stop on either 10 or 65. They spun it in front of students, then asked a question unrelated to the wheel: What percentage of African nations are in the United Nations?

The students who landed on 10 guessed a median of 25 percent. The students who landed on 65 guessed a median of 45 percent. Twenty points of movement, produced by a number the students knew was random.

Daniel Kahneman, who wrote Thinking Fast and Slow, named it anchoring. The first number a mind encounters becomes the reference point, and every later judgment starts from there.

Here is the part most people miss, and it is the part that should worry you.

The effect survives knowing about it.

Kahneman describes a study where professional real estate agents were given a house’s asking price and asked to appraise the property themselves. The asking price pulled their estimates toward it, and when the agents insisted the number had not influenced them, their own numbers said otherwise.

These were professional real estate agents. This was their entire expertise. The anchor still tied them down.

Knowing a number is arbitrary doesn’t make it stop working. The mind does not delete a reference point because you learned to doubt it. It adjusts. It just adjusts too little.

The anchor survives because thinking hard is expensive. The fast part of the mind snaps up the first number it sees and treats it as a reasonable starting point. The slow part can override it, but overriding takes effort. Most days there is no reason to spend the effort. So the number stays half-questioned, and every downstream decision inherits a small error.

So the number stays half-questioned, and every downstream decision inherits a small error. Twenty points on a trivia question. A decade on the wrong plan.

This is also why the first number named in a negotiation wins the room. The side that names a salary first is not being rude. They’re planting an anchor.

Setting a number someone else picks locks you into a reference point you never agreed to. Name your own instead.

Now make it yours. Put down the drinking glass and the step counter. This is an essay about the numbers already installed in your head, so write down the first number that comes to mind and ask whose it is.

The salary figure you’ve been chasing since a recruiter said it first, back when you had no reason to doubt it. The size of the house. The school district. The net worth number you would be ashamed to say out loud. Financial freedom by forty-five. The title, the watch, the car.

Ask whose number each one is. Not whether the number is wrong, but whose is it.

That question alone does more than most goal-setting systems, because it’s the one question the system is built to skip.

I haven't met a man over thirty-five who measures himself against someone worse off. We measure against the visible surface of whoever is just ahead of us. That person is an anchor too, and the surface is curated.

You’re comparing your unedited life to their highlight reel and then calling yourself behind. The number isn’t yours, and neither is the view.

And the feed makes it daily. Every time you open it, you get a fresh reference point: what normal looks like, what a real house looks like, what a real body looks like, and what a man your age should have achieved by now.

Each one is a new anchor, and the mind stores them without asking for a source. This is why the filter matters.

“You do not get to decide whether the numbers come in. You only get to decide what you let in.”

The feeling of not enough usually has a trigger, and the trigger is usually a number. You were fine until you saw his. That’s anchoring, and it’s diagnosable, and the diagnosis is half the fix.

So run the audit once. Write down the three numbers you feel the most pressure around. Money, body, career, marriage; it doesn’t matter which.

For each one, ask three questions.

• Who said it first? Trace it back as far as you can. A parent, a boss, a guru, an algorithm. Most numbers have a first mention, and most have been repeated so many times since that the repetition has replaced the source.

• Did they set it for their good or for mine? This is not paranoia. It’s due diligence. A commission is a reason. An audience is a reason. A product to sell is a reason. None of those make the number false. They just make it someone’s, and the difference matters.

• Would I choose it again, starting from zero, knowing what I know now? This question separates a standard from a leftover. If the answer is yes, keep it, protect it, and stop resenting it. If the answer is no, you’ve just found an anchor, not a value, and an anchor can be pulled up.

The most common mistake, once a man notices the anchor, is to trade it for a bigger one. He trades the job for a better one, the suburb for a nicer one, and the watch for a more expensive one. Same anchor, more zeros. He just moved the reference point up and called it growth.

That’s the tourist disappointed by Paris who now books a ticket to London. The city wasn’t the problem. The image they had of the city, before visiting, was the problem.

The number in your head isn’t the problem. The fact that someone else set it is.

The gentle version of this is eight glasses and ten thousand steps. The harder version is the number a guru handed you, the one he sells a plan to reach. He’s guessing too, steered by someone above him, and the whole arrangement works because he never has to show you the author.

“Authority is the perfect hiding place for a guess.”

An anchor is just the first number someone planted where your eyes were. It takes hold because it was first. Truth never entered into it.

I want to be clear about what this is. It’s not a case against standards. Some standards protect your health, your family, and your ability to sleep at night. Keep those without apology. The test is never whether a number exists. The test is whether you chose it or absorbed it, and whether the person who set it meant it for your good.

Some numbers hold up under questioning. Those are the ones a man should keep with both hands. What you are looking for is not fewer anchors. It’s ownership of the ones you keep.

Here’s the part that should bring relief. You get to change any of this, any day you choose, and the change doesn’t require a big announcement.

The wind and the water are circumstances. You can’t control them. You can control the sail, and you can turn it today.

Noticing the anchor is the first adjustment. Most men never learn they were handed the wrong coordinates. They sail toward a port someone else picked and call the fatigue discipline.

You don’t have to be one of them. Pull up the anchor that was never yours, and the boat moves with less wind than you expected.

Own the number, or the number owns you.

The eight glasses. The ten thousand steps. The saturated fat hypothesis.

It all began with a person guessing in the dark. So did the salary, the house, and the net worth figure that has been steering you. They all have authors.

You get to be the author now, and the first page is the number you refuse to inherit.

I'm not saying AI is useless. I use it every day. But here's the problem.

When I ask ChatGPT about nutrition, it gives me answers that sound right. It cites studies. It uses confident language. It tells me things like "a balanced diet includes whole grains and lean protein" because that's what the internet says.

I've been self-studying nutrition for over 10 years. I've reversed pre-diabetes and cured 14 years of panic attacks through diet alone. So I can spot when it's wrong.

Most people can't.

AI Doesn't Know Anything

After working with AI for a few years as an SRE (Site Reliability Engineer) at a major software company, here's the thing you have to understand about how AI actually works.

It doesn't know anything. It doesn't have beliefs. It doesn't reason about what's true.

What it does is predict the next word.

You give it a prompt, and it asks itself: given everything I've been trained on, what word is most likely to come next?

Then it picks that word. Then it does it again. And again.

That's basically it. That's the whole thing.

So when the result sounds intelligent, it's not because the machine understands nutrition, it's because the machine has been trained on an enormous amount of text written by humans, and it has gotten very good at predicting what a human would probably say next.

This is what I call the 'Averaging Problem.'

AI doesn't give you the truth. It gives you the average.

It takes everything the internet has ever said about a topic, finds the statistical center of all that text, and hands it back to you in polished, confident-sounding sentences.

If the internet's consensus on a topic is correct, the average will be useful.

If the internet's consensus is wrong, the average will be confidently, convincingly wrong.

And you will have no way to tell the difference unless you already know the subject well enough to catch the errors.

The Internet's Nutrition Consensus Is Broken

When it comes to health and nutrition, the internet's consensus is not just occasionally wrong. It's systematically wrong.

The training data AI pulls from includes decades of industry-funded studies designed to shift blame away from sugar and toward fat.

It includes dietary guidelines written by committees that were, in many cases, advised by people with financial ties to processed food companies.

It includes the same "balanced diet" and "everything in moderation" messaging that has coincided with the worst metabolic health crisis in human history.

And don't think this is a conspiracy theory. That would be lazy of me and unfair to you (which I talk about in an up and coming newsletter).

It's well-documented.

The sugar industry paid Harvard scientists in the 1960s to downplay sugar's role in heart disease and point the finger at saturated fat instead.

That single piece of corrupted science shaped five decades of dietary advice.

It's still embedded in the training data. It's still in the average.

So when you ask ChatGPT what to eat, it doesn't consult biochemistry.

It doesn't look at comparative anatomy. It doesn't check the latest rigorous metabolic research.

It averages everything it's ever read about nutrition, and the average says: whole grains, lean protein, everything in moderation, calories in versus calories out.

The average got us here.

What the Average Misses

Here's what the average will never tell you.

It won't tell you that the same dietary advice that helped you get lean might be destroying your mental health.

It won't tell you that some people reverse their panic attacks by eliminating certain foods, because the average doesn't account for individual response.

It won't tell you that the studies it's citing might have been funded by the same companies that profit from the advice those studies support.

It also won't tell you that all human health studies are based on inferential statistics, which CAN NOT establish cause and effect.

The average doesn't know what worked for you. It doesn't know what failed.

It can't ask you to audit yourself honestly and check whether what you're doing is actually producing results.

That's the fundamental limit of AI in health. It can give you the consensus. It cannot give you the truth. And the distance between those two things, in nutrition, is enormous.

What Actually Works

If you can't trust AI, and you can't trust the internet's consensus, what do you trust?

Here's the framework I've settled on after 10 years of self-study, after reversing pre-diabetes, after curing panic attacks that no doctor connected to diet.

First, do your own due diligence.

Do not outsource your health decisions to a language model.

AI can help you articulate what you already know. It can help you organize your thinking. It should not give you the answers.

If you don't know the subject before you ask the question, you have no way to evaluate what comes back.

You'll take confident-sounding nonsense as truth because it feels right.

Second, follow the hard science.

Not the epidemiology. Not the food-frequency questionnaires where people try to remember what they ate six months ago.

I mean biochemistry. Comparative anatomy. Mechanisms.

The kind of science that asks: what actually happens in the body when you eat this? What are humans biologically adapted to eat? What does the metabolic pathway look like?

Third, audit yourself honestly.

A lot of health advice out there will help you get lean. It will not necessarily help you feel good. It will not necessarily improve your mental health.

You have to be honest with yourself about whether what you're doing is actually working, across all the dimensions that matter. Not just the mirror.

Your energy, your mood, your sleep, your anxiety, your ability to think clearly.

If the advice says you should feel great and you don't, the advice might be wrong.

For you, specifically.

Fourth, reject social proof and authority as substitutes for evidence.

The fact that a million people follow a certain diet does not make it correct. The fact that a credentialed expert endorses it does not make it correct.

The fact that it's the official recommendation does not make it correct. Popularity and authority are not evidence. They are shortcuts, and shortcuts in health are expensive.

The Filter Is the Value

There's a bigger point here, and it goes beyond nutrition.

Everyone has access to the same AI now. The same ChatGPT. The same information. The same averages.

If information were the thing that mattered, everyone would be equally capable of making good decisions about their health, their work, their life.

But they're not. And the reason they're not is that the value was never in the information. It was always in the filter.

The filter is your years of study. Your lived experience. Your mistakes. The things you learned the hard way.

The patterns you can recognize because you've seen them before. The bad advice you can spot because you followed it once and it made everything worse.

AI can't replicate any of that. It can't live in your body. It can't make your mistakes.

It can't spend a decade following a thread and arriving at a conclusion that contradicts the consensus.

It can only average what the consensus has already said.

The people who use AI best are the ones who already deeply understand their craft.

AI sharpens their thinking. It doesn't replace it. It helps them articulate what they already know. It doesn't give them the answers.

Finding the Truth Is Hard

I'm not saying you should ignore AI. I use it every day. I'm saying you should treat it like a research assistant who has read every book in the library but has never actually lived in a human body.

It can summarize. It can organize. It can help you see connections. It cannot tell you what's true.

Finding the truth about your own health requires more than a prompt.

It requires study. Self-audit. The willingness to be wrong. The willingness to go against the consensus when the evidence and your own experience point somewhere else.

It requires getting away from the default setting: going with the flow, following the crowd, doing what's easy.

Those instincts are exactly what the Averaging Problem exploits.

The average is comfortable. The average is popular. The average is what everyone else is doing.

The average is also what got us here.

Work on your critical thinking. Build your own filter.

Learn to trust your own mind and body, not because they're infallible, but because they're the only filter you have that AI can't replicate.

Finding the truth is hard. But it's worth it.

If you want more on cutting through the noise and learning to trust your own mind and body, subscribe below. This is just the beginning.

Subscribe